energies

Article
Crop Yield Estimation Using Deep Learning Based on Climate
Big Data and Irrigation Scheduling
Khadijeh Alibabaei 1,2, * , Pedro D. Gaspar 1,2 and Tânia M. Lima 1,2

1 C-MAST Center for Mechanical and Aerospace Science and Technologies, University of Beira Interior,
6201-001 Covilhã, Portugal; [email protected] (P.D.G.); [email protected] (T.M.L.)
2 Deparment of Electromechanical Engineering, University of Beira Interior, Rua Marquês d’Ávila e Bolama,
6201-001 Covilhã, Portugal
* Correspondence: [email protected]

Abstract: Deep learning has already been successfully used in the development of decision support
systems in various domains. Therefore, there is an incentive to apply it in other important domains
such as agriculture. Fertilizers, electricity, chemicals, human labor, and water are the components
of total energy consumption in agriculture. Yield estimates are critical for food security, crop
management, irrigation scheduling, and estimating labor requirements for harvesting and storage.
Therefore, estimating product yield can reduce energy consumption. Two deep learning models,
Long Short-Term Memory and Gated Recurrent Units, have been developed for the analysis of
time-series data such as agricultural datasets. In this paper, the capabilities of these models and their
extensions, called Bidirectional Long Short-Term Memory and Bidirectional Gated Recurrent Units,
to predict end-of-season yields are investigated. The models use historical data, including climate



data, irrigation scheduling, and soil water content, to estimate end-of-season yield. The application


of this technique was tested for tomato and potato yields at a site in Portugal. The Bidirectional
Citation: Alibabaei, K.; Gaspar, P.D.; Long Short-Term memory outperformed the Gated Recurrent Units network, the Long Short-Term
Lima, T.M. Crop Yield Estimation Memory, and the Bidirectional Gated Recurrent Units network on the validation dataset. The model
Using Deep Learning Based on was able to capture the nonlinear relationship between irrigation amount, climate data, and soil
Climate Big Data and Irrigation water content and predict yield with an MSE of 0.017 to 0.039. The performance of the Bidirectional
Scheduling. Energies 2021, 14, 3004.
Long Short-Term Memory in the test was compared with the most commonly used deep learning
https://doi.org/10.3390/en14113004
method, the Convolutional Neural Network, and machine learning methods including a Multi-Layer
Perceptrons model and Random Forest Regression. The Bidirectional Long Short-Term Memory
Academic Editor: Amparo López
outperformed the other models with an R2 score between 0.97 and 0.99. The results show that
Jiménez
analyzing agricultural data with the Long Short-Term Memory model improves the performance of
Received: 20 April 2021 the model in terms of accuracy. The Convolutional Neural Network model achieved the second-best
Accepted: 20 May 2021 performance. Therefore, the deep learning model has a remarkable ability to predict the yield at
Published: 22 May 2021 the end of the season.

Publisher’s Note: MDPI stays neutral Keywords: agriculture; deep learning; LSTM; support decision-making algorithms; yield estimation;
with regard to jurisdictional claims in irrigation management
published maps and institutional affil-
iations.

1. Introduction
Agriculture is in a state of flux, and obstacles are emerging, such as climate change,
Copyright: © 2021 by the authors. environmental impacts, and lack of labor, resources, and land. Annual population growth
Licensee MDPI, Basel, Switzerland. and increasing demands on agricultural society to produce more from the same amount of
This article is an open access article agricultural land while protecting the environment are the significant challenges of this
distributed under the terms and century [1]. This scenario reinforces the constant need to seek alternatives in the face of
conditions of the Creative Commons
challenges to ensure higher productivity and better quality. Sustainable production of
Attribution (CC BY) license (https://
sufficient, safe, and high-quality agricultural products will be achievable if new technolo-
creativecommons.org/licenses/by/
gies and innovations are adopted. Smart farms rely on data and information generated
4.0/).

Energies 2021, 14, 3004. https://doi.org/10.3390/en14113004 https://www.mdpi.com/journal/energies

Energies 2021, 14, 3004 2 of 21

by agricultural technology, bringing the producer closer to digital technology [1]. This
includes the use of sensors and drones and the collection of accurate data such as weather
data, soil mapping, and others. Extracting knowledge from these data and creating de-
cision support systems is becoming increasingly important to optimize farms and add
value to meet the food needs of the population and ensure the sustainable use of natural
resources [1].
Deep learning (DL) is a subfield of machine learning. DL algorithms can be used
throughout the cultivation and harvesting cycle in agriculture and are receiving consid-
erable attention in developing such decision-making systems. The idea is to feed large
artificial neural networks with increasingly large amounts of data, extract features from
them automatically, and make decisions based on these data [2]. Deep here refers to
the number of hidden layers of the neural network. The performance of the model im-
proves as the network becomes deeper [2].
Crop yields and crop yield forecasts directly affect the annual national and inter-
national economy and play a major role in the food economy. Crop yields are highly
dependent on irrigation and climate data. More irrigation does not necessarily increase
yield [3], and therefore, optimization of irrigation and more efficient irrigation systems
are critical. Predicting yield based on different types of irrigation is one way to optimize
the process.
Machine learning algorithms are already used to estimate yield from images. Bar-
goti and Underwood [4] used Multi-Layer Perceptrons (MLP) and Convolutional Neural
Network (CNN) models to extract features from input images, and then two image pro-
cessing algorithms, Watershed Segmentation (WS) and Circular Hough Transform (CHT),
were used to detect and count individual fruits in these features. They added metadata
such as pixel position, row number, and sun azimuth to their algorithms and improved
the detection performance. The best performance was obtained for fruit detection by CNN
and WS with R2 = 0.826. Habaragamuwa et al. [5] developed a Region-based CNN
(R-CNN) model with AlexNet as the backbone and detected ripe and unripe strawber-
ries in greenhouse images. The model achieved an average precision of 82.61%. Kang
and Chen [6] implemented a clustering CNN model (C-RCNN) and a deep learning model,
LedNet, to detect apples on trees. The C-RCNN module was used to generate a label
for the training dataset, and LedNet was trained to detect apples on trees. A lightweight
network (LW-Net), ResNet110, ResNet50, and Darknet-53 were used as the backbone.
LedNet with the ResNet110 backbone, with 86% accuracy, and LedNet with LW-Net, with
weight size and computation time of 7.4 M and 28 ms, respectively, outperformed the other
models in terms of detection performance and computational efficiency. Koirala et al. [7]
developed a DL model, named Mango-YOLO, based on YOLO-v3 and YOLO-v2 (tiny)
for counting mangoes on trees. Mango-YOLO achieved the best performance in terms
of memory consumption, speed, and accuracy compared to the Faster R-CNN, Single
Shot multi-box Detector (SSD), and You Only Look Once (YOLO). Liang et al. [8] applied
the SSD network to detect mango and almond on tree fruits. The SSD model with the data
augmentation techniques and the smaller standard box was more accurate than the original
SSD network in detecting mango on trees. Stein et al. [9] developed an FR-RCNN using
VGG16 as a backbone for fruit detection and localization in a mango orchard. They used
three datasets for training. The first contained the image of apple trees from one side of
the trees, the second contained the image from both sides of the trees, and the third con-
tained images from multiple views of the trees. Training the model with images from two
and multiple views showed excellent performance (R2 ≥ 0.90). Tian et al. [10] developed
YOLO-V3 with DenseNet as the backbone to detect apples on trees. They used two datasets
for training. The first contained images of apples at one growth stage, and the second
contained images taken at different growth stages. Their results showed that the F1 score
of the model trained with the first dataset was higher than that of the model trained with
the second dataset. Apolo-Apolo et al. [11] used a Faster R-CNN model and a Long
Short-Term Memory (LSTM) model to estimate fruit number and fruit size. An average
Energies 2021, 14, 3004 3 of 21

standard error (SE) of 6.59% between visual fruit count and fruit detection by the model
was determined. An LSTM model was trained for per-tree yield estimation and total yield
estimation. Actual and estimated yields per tree were compared, yielding an approximate
error of SE = 4.53% and a standard deviation of SD = 0.97 kg. Maimaitijiang et al. [12]
used Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), Support
Vector Regression (SVR), DNN (DNN-F1) based on input-level feature fusion, and DNN
(DNN-F2) based on mid-level feature fusion to estimate soybean yeast. The results showed
that multimodal data fusion improved the accuracy of yield prediction. DNN-F2 achieved
the highest accuracy with an R2 score of 0.720 and a relative root mean square error (RMSE)
of 15.9%. Yang et al. [13] proposed a CNN architecture for predicting rice grain yield from
low-altitude remote sensing images at the maturity stage. The proposed model consisted
of two separate branches for processing RGB and multispectral images. In a large rice-
growing region of Southern China, a 160-hectare area with over 800 cultivation units was
selected to investigate the ability of the model to estimate rice grain yield. The network
was trained with different datasets and compared with the traditional vegetation index-
based method. The results showed that the CNNs trained with RGB and multispectral
datasets performed much better than the VI-based regression model in estimating rice
grain yield at the maturity stage. Chen et al. [14] proposed a faster Region-based Con-
volutional Neural Network (R-CNN) for detecting and counting the number of flowers,
mature strawberries, and immature strawberries. The model achieved a mean average
accuracy of 0.83 for all detected objects at 2 m height and 0.72 for all detected objects at
3 m height. Zhou et al. [15] implemented an SSD model with two lightweight backbones,
MobileNetV2 and InceptionV3, to develop an Android app called KiwiDetector to detect
kiwis in the field. The results showed that MobileNetV2, quantized MobileNetV2, Incep-
tionV3, and quantized InceptionV3 achieved true detection rates of 90.8%, 89.7%, 87.6%,
and 72.8%, respectively.
The disadvantages of estimating the yield from images are:
• Pictures of the entire field must be collected each year to identify the crop in the pic-
tures and then estimate the yield.
• To train the model, a large number of labeled images is needed, which is very time-
consuming.
• Illumination variance, foliage cover, overlapping fruits, shaded fruits, and scale
variations affect the images [16].
Ma et al. [17] used climate, remote sensing data, and rice information to estimate rice
yield. A Stacked Sparse Auto-Encoder (SSAE) was trained and achieved a percent mean
square error of 33.09 kg (10a)−1 . Han et al. [18] implicated machine learning methods
including Support Vector Machine (SVM), Gaussian Process Regression (GPR), Neural
Network (NN), K-Nearest Neighbor Regression, Decision Tree (DT), and Random For-
est (RF) to integrate climate data, remote sensing data, and soil data to predict winter
wheat yield based on the Google Earth Engine platform (GEE). SVM, RF, and GPR with
an R2 > 0.75 were the three best yield prediction methods, among others. They also found
that different agricultural zones and temporal training settings affected the prediction
accuracy. Kim et al. [19] developed an optimized deep neural network for crop yield pre-
diction using optimized input variables from satellite products and meteorological datasets.
The input data were extracted from satellite-based vegetation indices and meteorological
and hydrological data, and a matchup database was created on the Cropland Data Layer
(CDL), a high-resolution map for classifying plant types. Using the optimized input dataset,
they implemented six major machine learning models, including multivariate adaptive
regression splines (MARS), SVM, RF, highly randomized trees (ERT), ANN, and DNN.
The DNN model outperformed the other models in predicting corn and soybean yields,
with a mean absolute error of 21–33% and 17–22%, respectively. Abbas et al. [20] used four
machine learning algorithms, namely Linear Regression (LR), Elastic Net (EN), K-Nearest
Neighbor (k-NN), and Support Vector Regression (SVR), to predict tuber yield of potato
(Solanum tuberosum) from soil and plant trait data acquired by proximal sensing. Four
Energies 2021, 14, 3004 4 of 21

datasets were used to train the models. The SVR models outperformed all other models
in each dataset with RMSE ranging from 4.62 to 6.60 t/ha. The performance of k-NN
remained poor in three out of four datasets.
In these papers, however, the amount of irrigation was not considered as an in-
put to the model. Yield is highly dependent on the amount of irrigation, and a change
in the amount of water can make a big difference in the yield. Considering irrigation
scheduling as an input to the model can help to create an intelligent system that selects
the best irrigation schedule to save water consumption without affecting production. To
optimize irrigation based on productivity, the irrigation amount must be considered as
an input to the model.
In this work, Recurrent Neural Networks (RNN), including the LSTM model and Gated
Recurrent Units (GRU) model and their extensions, Bidirectional LSTM (BLSTM) and Bidi-
rectional GRU (BGRU), were implemented to estimate tomato yield based on climate
data, irrigation amount, and water content in the soil profile. Agricultural datasets are
time-series, and agricultural forecasting relies heavily on historical datasets. The advantage
of RNN is its ability to process time-series data and make decisions for the future based
on historical data. The proposed models predict the yield at the end of the season given
historical data from the field such as temperature, wind speed, solar radiation, ETo, the wa-
ter content in the soil profile, and irrigation scheduling during a season (the codes are
available at the following links: https://github.com/falibabaei/tomato-yieldestimation/
blob/main/main (accessed date 22 May 2021), https://github.com/falibabaei/potato-
yield-estimation/tree/main (accessed date 22 May 2021)). The performance of the model
was evaluated using the mean square error and R2 score. In addition, the performance of
these models was compared with a CNN, an MLP model, and a Random Forest Regression
(RF). The advantages of the yield estimation model are:
• Using RNN models to extract features from past observations in the field and predict
yield at the end of the season.
• Using climatic data collected in the field as input to the model, which is easier than
using collected images from the field.
• Irrigation amount was used as input in the model, and it is shown that the model can
capture the relationship between irrigation amount and yield at the end of the season.
• It is shown that the model can be used as part of an end-to-end irrigation decision-
making system. This system can be trained to decide when and how much water to
irrigate and maximize net return without wasting water.

2. Materials and Methods

2.1. Deep Learning Algorithms
Machine learning (ML) is a subfield of artificial intelligence that uses computer algo-
rithms to transform raw data from the real world into valuable models. ML techniques
include Support Vector Machines (SVM), Decision Trees, Bayesian learning, K-Means
clustering, association rule learning, regression, neural networks, and many others [2].
Deep learning is a subfield of ML. The word “deep” refers to the number of hidden
layers in DL algorithms, making them more complex than ML algorithms. Deep neural
networks can learn the features from data with multiple hidden layers and solve more
complex problems. Unlike ML methods, DL models automatically extract useful features
from the raw data through training and do not require feature engineering. The training
time of DL models is longer than that of ML methods, and they require a large amount of
data to train, but when trained, they are more accurate and faster [21]. For these reasons,
they have been widely used in recent years.
The most widely used algorithm of DL consists of the CNN model and Recur-
rent Neural Network. The CNN models are already used for classification, recognition,
and localization. In 2012, AlexNet [22] won the LSVRC competition for classification.
Sermanet et al. [23] showed that DL algorithms could be used for classification, recognition,
Energies 2021, 14, 3004 5 of 21

and localization and achieved excellent results. However, CNN models make predictions
based on current input data and do not use past observations to make future decisions.
Unlike CNN, information in recurrent neural networks goes through a loop that
allows the network to remember the previous outputs [24]. It enables the analysis of
sequences and time series. RNN is commonly used for natural language processing
and other sequences. A recurrent network can be thought of as multiple copies of the same
network, each passing information to the next (see Figure 1).

Figure 1. RNN looped and unfolded sequentially.

The output of an RNN unit is calculated by Equation (1).

ht = tanh(Wx xt + Wh ht−1 + bt ), (1)

where ht−1 is the recurrent output from the previous step, xt is the input at the current
time, and Wx , Wh , bt are the weights and bias of the network to be trained when training.
The problem with RNN is that if the sequential input is large, the gradient of the loss
function can end at zero, effectively preventing the weights of the model from being
updated [24].

2.1.1. LSTM and GRU Structures

The LSTM [25] and Gated Recurrent Units (GRU) [26] were developed to address the
problems of the RNN.
The LSTM contains a forget gate that can be used to train individual neurons on what
is important and how long it remains important. An ordinary LSTM unit consists of a block
input zt , an input gate it , a forget gate f t , an output gate ot , and a memory cell ct (see Figure 2).
The forget gate f t is used to remove information that is no longer useful in the cell state
using Equation (2). The input at a given time xt and the previous cell output ht−1 are fed to
the gate and multiplied by weight matrices, followed by the addition of the bias. The result is
passed through a sigmoid function that returns a number between 0 and 1. If the output is 0,
the information is forgotten for a given cell state; if the output is 1, the information is retained
for future use. Adding useful information to the cell state is performed by the input gate it
using Equation (3). First, the information is controlled by the sigmoid function, which filters
the values to be stored, similar to the forget gate. Then, a vector of new candidate values of
ht−1 and xt is generated with the block gate zt using Equation (4), which outputs from −1 to
+1. The vector values and the controlled values are multiplied to obtain useful information
using Equation (5). The output gate ot decides which information in the cell is used to calculate
the output of the LSTM unit using Equation (6).

f t = σ (W f x x t + W f h h t − 1 + b f ) , (2)
it = σ (Wix xt + Wih ht−1 + bi ), (3)
zt = tanh(Wzx xt + Wzh ht−1 + bz ), (4)
c t = f t ∗ c t −1 + i t ∗ z t . (5)
ot = σ (Wox xt + Woh ht−1 + bo ) (6)

First, a vector is created by applying the tanh function to the cell. Then, the information
is regularized using the sigmoid function, which filters the values to be stored based
Energies 2021, 14, 3004 6 of 21

on the inputs ht−1 and xt . The vector values and the regulated values are multiplied by
Equation (7) to be sent as output and input to the next cell.

ht = ot ∗ tanh(ct ). (7)

GRUs discard the cell state and use the hidden state to transmit information. This ar-
chitecture contains only two gates: the update gate zt and the reset gate rt . Like LSTM gates,
GRU gates are trained to selectively filter out all irrelevant information while preserving
the useful information and can be calibrated using Equations (8)–(11):

zt = σ(Wux xt + Wuh ht−1 + bu ), (8)

rt = σ(Wrx xt + Wrh ht−1 + br ), (9)
ot = tanh(Wox xt + Woh (rt ∗ ht−1 ) + bo ). (10)
h t = (1 − z t ) ∗ o t + z t ∗ h t −1 . (11)

Figure 2. LSTM and GRU structures. Left side shows LSTM cell, right side shows GRU cell.

The functions tanh and σ add nonlinearity to the network. These functions allow
the model to capture the nonlinear relationships between the inputs and outputs of
the model. At the beginning of training, the weights and biases in Equations (2)–(4),
(6) and (8)–(10) are set randomly. During training, the model tries to set the weights and bi-
ases in such a way that the loss function is minimized. Therefore, the algorithm of an RNN
model is an optimization problem.
The LSTM and GRU models have similar units, but they also differ. For example,
in the LSTM unit, the amount of memory content seen or used by other units in the network
is controlled by the output gate. In contrast, the GRU releases all of its content without
any control [27]. From these similarities and differences alone, it is difficult to conclude
which model performs better on one problem than another. In this paper, both models
were implemented to see which model performs better on the yield estimation problem.

2.1.2. Bidirectional LSTM Structure

Bidirectional LSTM (BLSTM) is an extension of the LSTM model that can improve
the results [28]. Its architecture consists of a stack of two separate intermediate LSTM layers
that send a sequence forward and backward to the same output layer, using contextual
information from both sides of the sequence (see Figure 3).
Energies 2021, 14, 3004 7 of 21

Figure 3. BLSTM layer.

2.1.3. Dropout and Early Stopping

Overfitting occurs when the model learns noise in the training data, and the generaliza-
tion of the model to the unseen data is unreliable and cannot make accurate estimates. It is
proposed to use regularization techniques to prevent overfitting. One of them is “dropout”,
which consists of randomly selecting some neurons of the hidden layer and blocking their
output so that they are not evaluated in the learning algorithms, and then, after a while,
releasing the outputs of the blocked neurons and blocking other neurons [29]. This leads
to the neural network becoming more general and not depending only on one group of
neurons to make certain decisions. In LSTM models, the dropout can be added to the inputs
of the layers, the outputs of the layers, and the recurrent outputs [29]. The dropout size is
a hyperparameter that should be set during training (see Figure 4). Table 1 shows the effect
of the dropout on the validation loss.

Figure 4. Left side shows the normal neural network and the right side shows the network with dropout.

Table 1. Validation lost under different dropout sizes.

Dropout Size
Dropout Size
0 0.1 0.2 0.3 0.4
Tomato 0.00021 0.00019 0.00033 0.00050 0.00068
Potato 0.00140 0.00109 0.00101 0.00092 0.00185

Another solution is early stopping, which consists of splitting the dataset into three
sets, one for training, one for validation, and one test set [30]. In this method, the validation
Energies 2021, 14, 3004 8 of 21

loss is constantly evaluated in each episode, and if the validation loss does not improve
for a certain number of episodes, the training is stopped. This technique does not allow
the network to be very specific about the training set.
In this work, dropout and early stopping were used to prevent overfitting.

2.2. Data Collection

The climate big data were collected by an agricultural weather station for a site
in Portugal. They were retrieved from the government agency of the Ministries of Agri-
culture and the Sea, the Direção Regional de Agricultura e Pescas do Centro, Portugal
(www.drapc.gov.pt (accessed date 22 May 2021)). The soil type in Fadagosa is either sandy
or sandy loam, permeable, with low to medium organic matter content, with low acid to
neutral reaction, rich in phosphorus and potassium, and without salt effects. The climate
type of Fadagosa is the Mediterranean hot summer climate (Csa). These are areas with
mild winters, with precipitation falling mainly in autumn and winter and occasionally
in spring. Summers are hot and dry, with maximum temperatures above 40 ◦ C. Figure 5
shows the Fadagosa region from Google Earth and Table 2 shows location details.

Figure 5. Map of Fadagosa region.

Table 2. Details of the location.

Location Longitude Latitude Start Date End Date Temporal Resolution

Fadagosa 40◦ 10 46.5500 N 7◦ 260 36.2700 W 1 January 2010 23 March 2020 15-min

The big data included minimum, maximum, and average temperature; minimum,
maximum, and average relative humidity; average solar radiation; minimum and average
wind speed; and precipitation. They were recorded every 15 mins and converted to
the daily variable in the data preprocessing section. Table 3 shows the details of these
variables. Figure 6 shows the daily variables from 2010 to 2019 (10 years of data).

Table 3. Dataset details.

Variables Unit Data Source Max Min Mean SD

TMin ◦C DRAP-Centro 27 −4.7 9.76 5.63
TMax ◦C DRAP-Centro 42.7 1.8 21.84 8.42
TAvg ◦C DRAP-Centro 34.84 −0.12 15.68 6.90
HR Min % DRAP-Centro 95 0 38.72 20.20
HR Max % DRAP-Centro 97 24 81.29 15.05
HR Avg % DRAP-Centro 95.89 27.75 60.38 18.78
SR Avg wm−2 DRAP-Centro 346.66 6.35 172.02 89.25
WS Max ms−1 DRAP-Centro 86.5 3.5 24.67 10.61
WS Avg ms−1 DRAP-Centro 28.85 0.031 4.62 3.80
Prec mm DRAP-Centro 101.6 0 2.28 7.20
ETo mmd−1 Penman-Monteith 9.8 0.2 3.68 2.088
equation (AquaCrop)
Energies 2021, 14, 3004 9 of 21

Table 3. Cont.

Variables Unit Data Source Max Min Mean SD

WCTot (Tomato) mm Aquacrop 365.2 145.9 247.96 36.18
WCTot (Potato) mm Aquacrop 432.4 165.6 311.62 51.95
Potato Yield ton (ha)−1 Aquacrop 12.706 5.539 10.8 2.053
Tomato Yield ton (ha)−1 Aquacrop 8.482 3.434 7 1.39
The abbreviations stand for the following: Max: maximum, Min: minimum, SD: standard deviation, Avg: average,
T: temperature, HR: relative humidity, SR: Solar Radiation, WS: Wind Speed, Prec: precipitation, ETo: Reference
Evaporation, WCTot: water content in the total soil profile.

35
25 40
30
20 35
25 30
15
20 25

Tmax
TAvg
Tmin

10 15 20
5 10 15
10
0 5
5
5 0
0
0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500
Time step Time step Time step

100
90 90
80
80 80
60 70 70
60

HRmax
HRmin

HRAvg

60
40 50
50
40
20 40
30
30
0 20
0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500
Time step Time step Time step

350 30
100
300 25
80
250 20
60 200
15
WSAvg
SRAvg
Prec

40 150
10
100
20 5
50
0 0 0
0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500 0 500 1000 1500 2000 2500 3000 3500
Time step Time step Time step

Figure 6. Fadagosa dataset.

To train a deep learning model on a dataset, a sufficient amount of data is needed.

However, in reality, recording crop yields at the end of the season based on different
irrigation schedules is extremely slow and sometimes impossible (e.g., for a season without
irrigation). The simulation model package Aquacrop was used to overcome this problem.
AquaCrop is a crop growth model developed by the Food and Agriculture Organization
(FAO) to ensure food security and assess the impact of environmental and management
influences on crop production [31]. The structure of the model was designed to be appli-
cable across different locations, climates, and seasons. To achieve this goal, AquaCrop
distinguishes between conservative (fixed) and non-conservative (case-specific) param-
eters. The conservative parameters do not change with geographic location, crop type,
management practices, or time and are intended to be determined with data from favorable
and non-stressful conditions but remain applicable under stressful conditions by mod-
ulating their stress response functions [31]. The Aquacrop model was calibrated with
climate data collected over the past ten years at Fadagosa. The crops selected for model
calibration were tomato and potato. In the crop menu of Aguacrop, a planting/sowing date
is generated by automatically evaluating the rainfall or temperature data prior to seeding.
The temperature criterion was selected to determine the planting/sowing date. April 7
was generated for tomato and February 17 for potato. The experiments were conducted
on sandy loam soil typical of Fadagosa with no salinity. The irrigation method chosen
was sprinkler irrigation, which is adjustable in Aquacrop. A fixed interval of four days
and a fixed value between 0 and 60 mm were used as time and depth criteria for calculating
the irrigation regimes. Different irrigation treatments were applied in each experimental
Energies 2021, 14, 3004 10 of 21

year, including no irrigation and a fixed irrigation depth of 5, 10, 15, 20, 25, 30, 35, 40, 45,
50, 60 mm every four days or when the allowable depletion reached the thresholds of
90%, 80%, and 70%, respectively. Other parameters were kept unchanged. Figure 7 shows
tomato yield under the fixed irrigation depth of 20 mm and water content throughout
the soil profile from 2010 to 2018. As can be seen in the figure, the yield varies under
different climatic conditions.

tomato yield under 20 mm irrigation WcTot under 20 mm irrigation

7 2010 300 2010
2011 2011
6 2012 280 2012
2013 2013
5 2014 260 2014
2015 2015
tomato yield

4 240
2016 2016

WcTot
2017 220 2017
3 2018 2018
200
2
180
1
160
0 140
0 20 40 60 80 100 0 20 40 60 80 100
Time step Time step

Figure 7. Evolution of tomato yield and total water content profile under fixed irrigation depth of
20 mm. Left side shows tomato yield and right side shows WCTot.

Evapotranspiration is the amount of water that evaporates from the Earth, and soil
water content is the volume of water per unit volume of soil. Evapotranspiration (ETo)
and water content in the total soil profile (WCTot) were also simulated during the simula-
tion and used as inputs to the models. AguaCrop estimated ETo from meteorological data
using the FAO Penman–Monteith Equation (12) [32].

( es − e a )
∆( Rn − G ) + ρ a c p rs
ETo = rs (12)
∆ + γ (1 + ra )

where Rn is the net radiation, G is the soil heat flux, (es − ea ) represents the vapor pressure
deficit of the air, ρ is the mean air density at constant pressure, c p is the specific heat of
the air, ∆ represents the slope of the saturation vapor pressure temperature relationship,
γ is the psychrometric constant, and rs and r a are the (bulk) surface and aerodynamic
resistances. Figure 8 shows the ETo from 2010 to 2019.
The advantage of using deep learning models is that they do not require manual
adjustments once the model is trained and can be used automatically. These models can be
used to create an end-to-end decision support system for irrigation scheduling.

14
12
10
8
ETo

6
4
2
0
0 500 1000 1500 2000 2500 3000 3500
Time step
Figure 8. ETo calculated by Aquacrop.
Energies 2021, 14, 3004 11 of 21

2.3. Data Preprocessing

The quality of the features extracted from the data by a deep learning model is deter-
mined by the quality of the dataset provided as input. The actions performed in the prepro-
cessing phase aim to prepare the data so that the feature extraction phase is more effective.
In this work, the missing big data were filled using the moving average method [33].
The other preprocessing in this paper involves the normalization and removal of
multicollinear parameters. The multicollinear parameters contain the same information
about the variables and can lead to redundancy in the calculations. One way to measure
multicollinearity is the Variance Inflation Factor (VIF), which assesses how much the vari-
ance of an estimated regression coefficient increases when its predictors are correlated.
If the VIF is equal to 1, there is no multicollinearity between factors, but the predictors may
be moderately correlated if the VIF is greater than 1. A VIF between 5 and 10 indicates
a high correlation, which may be problematic [34]. If the VIF is greater than 10, it can
be assumed that the regression coefficients are underestimated due to multicollinearity.
The VIF was used to determine the multicollinear variables, and the variables with a VIF
greater than five were removed. Finally, average temperature (TAvg ), average humidity
(HR Avg ), average wind speed (WS Avg ), reference evapotranspiration (ETo), water content
in total soil profile (WCTot), irrigation, and precipitation (Prec) were used as inputs.
The normalization aims to bring the values of the dataset into a normal form without
distorting the differences in the ranges of values. In this work, min–max normalization
was used. For each feature in the dataset, the minimum value of that feature is converted
to 0, the maximum value is converted to 1, and every other value is converted to a decimal
number between 0 and 1 using Equation (13).

xold − xmin
xnew = (13)
xmax − xmin

where xmin and xmax are the minimum and maximum of each variable in the dataset.

2.4. Metric Evaluation

In regression problems, the evaluation of the model can be calculated by the distance
between the actual value and the value predicted by the model. In this work, the mean
square error (MSE) and R2 score were used to evaluate the model’s performance. The MSE
calculates the variance explained by the model, and the R2 score is a statistical measure
that calculates the variance explained by the model over the total variance [35]. The higher
the R2 score, the smaller the differences between the observed data and the fitted values [36].
MSE and R2 score can be calculated using Equations (14) and (15):
n
1
MSE =
n ∑ (yi − ŷi )2 , (14)
i =1

∑in=1 (yi − ŷi )2

R2 = 1 − , (15)
∑in=1 (yi − ȳi )2
where yi , ŷi , and ȳi are, respectively, the true value, the value determined by the model,
and the mean of the true values.

3. Results and Discussion

The predictions were made using a computing system consisting of an Intel Core
i7-8550 CPU, an NVIDIA GEFORCE MX130 graphics card, and 8.0 GB of RAM. Anaconda
is a distribution of the Python and R programming languages for scientific computing
that aims to simplify package management and deployment. TensorFlow is an open-
source software library for numerical applications with computational graphs. It was
developed directly by Google Brain Team for machine learning and deep neural networks.
Keras is a deep neural network library written in Python and running as a front-end
Energies 2021, 14, 3004 12 of 21

in TensorFlow or Theano. It was developed to enable rapid experimentation [37,38]. The

Anaconda environment was used to implement the model using Python version 3.7.9
and the Tensorflow GPU 2.1.0 framework, and Keras library version 2.2.4-tf.
The model obtained daily historical data for a season, including average temperature
(TAvg ), average humidity (HR Avg ), average wind speed (WS Avg ), reference evapotran-
spiration (ETo), total soil profile (WCTot), irrigation amount, and precipitation (Prec),
and estimated the yield at the end of the season. The tomato season was a string of length
110 days and the potato season was a string of length 121 days.
The data were divided into a training set, a validation set, and a test set. Training
and test datasets are used to train the model and evaluate the performance of the trained
network, respectively. The test set is ignored during training and hyperparameter selection,
and its errors determine how well the model generalizes to the unseen data. Validation
data are used to compare different models, select hyperparameters, and prevent overfitting
during the training phase. Here, 30% of the data were selected as a test set, and the rest
were split into 60–10% as training and validation sets, respectively. The tomato prediction
model was trained on 77 samples (season) and tested on 34 samples, and the potato crop
prediction model was trained on 45 samples and tested on 20 samples.
The error of a model is defined as the difference between the actual value and the value
predicted by the model and must be minimized in the training step. The function used
to calculate this error is called the loss function. Training of DL models is based on
backpropagation [39]. Backpropagation uses the chain rule and computes the gradient
of the loss function with respect to the weights of the network for a single input–output
example, the adjusted weights of the network, and the biases. The optimization algorithm
must be chosen to minimize the loss function. The choice of optimization algorithm and loss
function is crucial. In this paper, the mean square error and Adam optimizer were used [2].
When training a neural network, some specific values must be initialized, which
are called hyperparameters. Hyperparameters such as the number of hidden layers,
activation function, etc., determine the structure of a model, and hyperparameters such
as learning rate, decay time, etc., determine how the model is trained [24]. Results from
multiple simulations of the same algorithm with different hyperparameters will vary.
The hyperparameters considered in this work are the number of layers, the number of
hidden units, batch size, the dropout size, the learning rate, and the learning rate decay.
The learning rate is a hyperparameter that controls how much the weights of the model
are adjusted with respect to the gradient of the loss function. It has a significant impact
on the training process of the deep learning models. A very low learning rate makes
the learning of the network very slow, while a very high learning rate causes fluctuations
in training and prevents the convergence of the learning process [24]. Learning rate decay
is another hyperparameter where the learning rate is calculated by decay at each update.
In this paper, the hyperparameters were selected manually. Unlike neural networks,
the recurrent network does not require a deep network to obtain accurate results [40].
Changing the number of nodes per layer has been shown to have a more significant impact
on results than changing the number of layers [40]. Since choosing LSTM and GRU with
four layers and BLTM with three layers significantly increases the number of trainable
parameters, leading to overfitting of the models, the number of LSTM and GRU layers
was kept below four and the number of BLSTM and BGRU layers below three. Moreover,
the LSTM and GRU with one layer showed poor performance and were excluded from
the experimental results. Finally, the network architectures with two and three LSTM
and GRU layers and one and two BLSTM and BGRU layers were tested.
Due to the hardware-based reasoning, the number of nodes per layer is usually chosen
as a power of two, which can speed up the training of the model [41]. Since the time step
in the input of the model is more than 110 days, the number of nodes per layer was kept
less than or equal to 512 to avoid overfitting due to the number of trainable parameters
and to make the training faster and more efficient [24]. The models with 16 and 32 nodes
Energies 2021, 14, 3004 13 of 21

also performed very poorly and were excluded from the experimental results. In the end,
the network architectures were tested with 64, 128, 256, and 512 nodes per layer.
The MSE was calculated for the validation dataset with all combinations of the number
of layers and the number of nodes. The results for the different crops are shown in Table 4.
As can be seen in Table 4, the BLSTM model improves the validation loss, but the GRU
model has worse performance compared to the LSTM model. Therefore, removing the cell
state from the LSTM reduces the performance of the model in the yield prediction problem.

Table 4. Validation MSE for different numbers of LSTM layers and LSTM nodes per layer.

Number of Nodes per Layer

Model Architecture Crop
64 128 256 512
Potato 0.00447 0.00561 0.00627 0.00132
LSTM-2 layers
Tomato 0.00389 0.00248 0.00119 0.00023
Potato 0.00497 0.00276 0.00140 0.00259
LSTM-3 layers
Tomato 0.00275 0.00333 0.00280 0.00226
Potato 0.00962 0.00213 0.00943 0.01067
GRU-2 layers
Tomato 0.00456 0.00150 0.00048 0.00022
Potato 0.01170 0.01360 0.00968 0.00974
GRU-3 layers
Tomato 0.00258 0.00118 0.00075 0.00030
Potato 0.00357 0.00461 0.00808 0.00110
BLSTM-1 layer
Tomato 0.00814 0.00046 0.00028 0.00017
Potato 0.01122 0.01058 0.0103 0.01250
BGRU-1 layer
Tomato 0.00793 0.00084 0.00020 0.00029
Potato 0.00914 0.00092 0.00167 0.00257
BLSTM-2 layers
Tomato 0.00329 0.00035 0.00012 0.00008
Potato 0.01610 0.00121 0.01183 0.01108
BGRU-2 layers
Tomato 0.00249 0.00050 0.00029 0.00022

Finally, a two-layer BLSTM with 128 nodes and a two-layer BLSTM with 512 nodes
were selected to predict potato and tomato yield, respectively. Tables 5 and 6 show the ar-
chitecture of the models selected for crop yield prediction. Adding an additional dense
layer after the BLSTM layers improved the results. The tanh function was used as an ac-
tivation function after each BLSTM layer to capture the nonlinear relationship between
input and output [24].

Table 5. BLSTM architecture used to predict tomato yield.

Layer (Type) Output Shape Param

Input (Batch Size, 110, 7) 0
BLSTM (Batch Size, 110, 1024) 2,134,016
Dropout (Batch Size, 110, 1024) 0
BLSTM (Batch Size, 1024) 6,295,552
Dropout (Batch Size, 1024) 0
Dense (Batch Size, 512) 524,800
Dense (Batch Size, 1) 513

The batch size is also chosen as a power of two due to hardware reasons. The number
of samples in the training datasets is less than 77, so the batch size is selected from
{16, 32, 64}. For simplicity, the learning rate and decay time were chosen as negative
powers of ten. With a learning rate and decay of 10−5 , the model trains very slowly,
and even after 500 epochs, the validation loss is very high, and the learning rate and decay
of 10−2 causes fluctuations in training. Therefore, the learning rate and decay are kept
below 10−6 and above 10−2 . Table 7 shows the loss on the validation set for different crops
when the hyperparameters are chosen differently. As can be seen, the same model with
Energies 2021, 14, 3004 14 of 21

different hyperparameters achieves different results. The models with a learning rate of
10−4 and 10−3 (potato and tomato, respectively), batch size of 64, and decay of 10−5 had
the best performance.

Table 6. BLSTM architecture used to predict potato yield.

Layer (Type) Output Shape Param

Input (Batch Size, 121, 7) 0
BLSTM (Batch Size, 121, 256) 139,264
Dropout (Batch Size, 121, 256) 0
BLSTM (Batch Size, 512) 394,240
Dropout (Batch Size, 512) 0
Dense (Batch Size, 512) 131,584
Dense (Batch Size, 1) 513

Dropout size is the percentage of nodes that randomly drop out during training.
The dropout size of 0.4 did not improve the validation loss, so it was chosen to be less than
or equal to 0.4 and from the set {0.1, 0.2, 0.3, 0.4}. In LSTM models, dropout can be added
to the input layer, outputs, and recurrent outputs [29]. Adding dropout on the recurrent
outputs with dropout size 0.1 for the potato yield estimation model and on the outputs of
the layers with dropout size 0.3 for the tomato yield estimation model improved the vali-
dation loss and was therefore selected for each model (see Table 1).

Table 7. Validation lost under different hyperparameters.

Batch Size
Batch Size 16 32 64
Tomato 0.00012 0.00020 0.00008
Potato 0.00164 0.00393 0.00092
Learning Rate
Learning Rate 10−3 10−4 10−5
Tomato 0.00005 0.00008 0.00571
Potato 0.00193 0.00092 0.02338
Decay
Decay 10−3 10−4 10−5
Tomato 0.00006 0.00007 0.00005
Potato 0.00140 0.00107 0.00092

Each model was trained for a maximum of 500 epochs, and each epoch lasted two
seconds. As mentioned earlier, early stopping was used to prevent overfitting. In this
method, training is stopped when the validation loss does not improve for a certain
number of epochs. Patience is a hyperparameter in the early stopping method that controls
the number of epochs in which the validation loss no longer improves [30]. The exact
amount of patience varies by model and problem. Examining plots of model performance
measures can be used to determine patience. In this work, by examining the plot, patience
was determined to be 30 and 50 for the tomato and potato models, respectively. As shown
in Figure 9, training of the tomato and potato yield prediction models was completed after
360 and 250 epochs, respectively. The tomato yield prediction model was trained with
more samples in the training set due to larger availability of experimental data, which
may result in the training of this model being more stable than that of the potato yield
prediction model.
Energies 2021, 14, 3004 15 of 21

Model loss Model loss

5 Train 0.6 Train
validation validation
4 0.5
0.4
3

Loss

Loss
0.3
2
0.2
1 0.1
0 0.0
0 50 100 150 200 250 300 350 0 50 100 150 200 250
Epoch Epoch

Figure 9. Model loss during training. The left side shows the loss of tomato yield prediction model
and the right side shows potato yield prediction model.

Since the test dataset was randomly selected, there were different seasons with differ-
ent amounts of irrigation in each test dataset. Table 8 shows the performance of the models
on the test dataset, and Figure 10 shows the actual value of yield compared to the values
predicted by the models. The model predicting tomato yield with an MSE of 0.017 per-
formed better on the test dataset than the model predicting potato yield with an MSE of
0.039. This result could be due to the fact that the standard deviation of tomato yield is
smaller than the standard deviation of potato yield (see Table 3) and also, as mentioned
earlier, the model used to estimate tomato yield was trained with a larger training set.
As shown in Figure 10, the tomato test dataset included the 2010 season under four
different irrigation levels. The irrigation amounts were 0, 10, 20, and 60, and the model
was able to achieve an MSE of 0.02 in this season. The same result was true for the potato
crop estimation model, and the model achieved an MSE of 0.09 to predict the 2012 crop
under four irrigation amounts. These results show that the model not only captures
the relationship between climate data and yield but also can accurately predict yield under
different irrigation amounts in a season.

Tomato yield estimation Potato yield estimation

8 12 2011
2012
t 11 2013
7 2010 2014
2011 2016
Predictions
Predictions

6 2012 10 2017
2013 2018
2014 9 2019
5 2015
2016 8
4 2017
2018
2019 7
3
4 5 6 7 8 7 8 9 10 11 12
True Values True Values

Figure 10. Predicted values vs. true values.

The ability to hold water depends on the soil type. As mentioned in the Data Collection
section, the soil type of Fadagosa is sandy loam, and these models are designed for a sandy
loam soil type. Therefore, these models work well on this soil type. Moreover, the models
were trained with simulated data. However, the combination of simulated data and real
data may give better results.
The performance of BLSTM models was compared with CNN, MLP, and the traditional
machine learning algorithm Random Forest (RF).
MLP is a computational model inspired by the human nervous system. It is able to
detect patterns in a mass of data in order to categorize it or regress a value. An MLP model
consists of an input layer, a stack of fully connected layers (hidden layers), and an output
layer. The fully connected layers connect every neuron in one layer to every neuron in the next
layer. Mathematically, these layers are a linear function where each layer takes the output
of the previous layer and adds weights and biases to it. To add nonlinearity to the model,
Energies 2021, 14, 3004 16 of 21

the activation function (e.g., ReLU, tanh) is used after each fully connected layer [42]. Again,
to avoid overfitting, the number of nodes in each layer of the MLP model was kept below
513 nodes, and the model with 512 neurons achieved the best performance in each dataset.
An MLP with 512 neurons and a different number of layers was implemented. The performance
improvement of the MLP model with five layers stopped on both datasets, so the number of
layers was kept below 5. Table 9 shows the performance of these models. The models with
three layers and four layers achieved the best performance in predicting tomato and potato
yields, with R2 scores of 0.89 and 0.71, respectively.
CNN is a deep learning algorithm. A CNN model consists of a stack of convolutional
layers, nonlinear activation functions, pooling layers (e.g., maximum pooling, average
pooling), Flatten layers, and fully connected layers [24]. Similar to the fully connected
layers, a convolutional layer is a linear function, and it contains a set of kernels. A kernel
is a matrix used for a matrix multiplication operation. This operation is applied multiple
times in different input regions, and extracts feature maps from the input. Pooling is
a reduction process. It is a simple process to reduce the dimension of feature maps
and hence the number of parameters trained by the network. A Flatten layer is usually
used in splitting convolutional layers and the fully connected layers. It basically performs
a transformation in the output of the convolutional layer and changes its format to an array.
The fully connected layer takes the feature maps from the Flatten layer and applies weights
and biases to predict the correct label or regress a value [24]. The number of kernels
in each convolutional layer and the number of convolutional layers was chosen manually.
For the same hardware reason, the number of kernels in each convolutional layer is kept
as a power of two. The models with a number of four convolutional layers or more than
512 kernels in each layer start to overfit. Therefore, the number of kernels and layers is kept
below 513 and four, respectively. All combinations of the number of layers and kernels were
implemented. The CNN model with 512 kernels and a number of layers of 2 and 3 achieved
better validation loss. The batch size, learning rate, and decay time from the BLSTM model
were used for the CNN models. Padding was used in each convolutional layer to ensure
that the output had the same shape as the input. The kernel size is usually chosen as an odd
number. The model with a kernel size greater than 11 starts with overfitting, and below
three, the model performance was poor, so the kernel size was chosen from {5, 7, 11}.
Table 9 shows the architecture of the CNN model.

Table 8. Performance of the models on the test set. L shows the number of layers, and K shows
the kernel size.

Model Architecture Crops MSE R2 Score Trainable Parameters

Potato 0.039 0.988 665,601
BLSTM
Tomato 0.017 0.99 6,853,633
Potato 0.791 0.792 17,061,889
CNN (2L-5k)
Tomato 0.099 0.910 15,751,169
Potato 0.499 0.882 17,594,369
CNN (2L-7k)
Tomato 0.087 0.925 16,283,649
Potato 0.443 0.915 18,659,329
CNN (2L-11k)
Tomato 0.102 0.912 17,348,609
Potato 0.615 0.792 18,373,121
CNN (3L-5k)
Tomato 0.075 0.934 17,062,401
Potato 0.965 0.600 19,429,889
CNN (3L-7k)
Tomato 0.169 0.854 18,119,169
Potato 0.515 0.860 21,543,425
CNN (3L-11k)
Tomato 0.141 0.878 20,232,705
Potato 0.870 0.601 1,021,953
MLP (3L)
Tomato 0.126 0.891 976,897
Potato 0.643 0.711 1,284,609
MLP (4L)
Tomato 0.133 0.884 1,239,553
Potato 1.270 0.428 1,547,265
MLP (5L)
Tomato 0.143 0.8765 1,502,209
Potato 0.5 0.872 -
RF
Tomato 0.107 0.901 -
Energies 2021, 14, 3004 17 of 21

Table 9. CNN architecture.

Layer (Type) Output Shape

Input (Batch size, time-steps, number of features)
(Conv1D) (kernel-size) (Batch size, time-steps, 512)
Dropout (Batch size, time-steps, 512)
Average-pooling (pool-size = 2) (Batch size, (time-steps/pool-size), 512)
Flatten (Batch size, (time-steps/pool-size) × 512 )
Dense (Batch size, 512)
Dropout (Batch size, 512)
Dense (Batch size, 1)

The CNN with two layers and three layers and a kernel size of 5 and 11 (tomato
and potato, respectively), with an R2 score of 0.96 and 0.933, performed better in predicting
yield than models with other combinations of the number of layers and kernel sizes.
RF is one of the most powerful machine learning methods. The Random Forest
consists of several Decision Trees. Each individual tree is a very simple model that has
branches, nodes where a condition is verified, and if it is satisfied, the flow goes through
one branch, otherwise through the other, always to the next node until the tree is finished.
As Table 8 shows, the RF model outperformed the MLP model in predicting yield, with
an R2 score of 0.87 to 0.90.
Computation time in the training process for each epoch lasted two, one, and less
than one second for BLSTM, CNN, and MLP, respectively. Although the computation time
in training BLSTM was higher than the other models, BLSTM achieved the best accuracy
in the test dataset.
As mentioned earlier, one of the applications of this model is to create a decision-
making system that decides when and how much to irrigate to avoid wasting water
without affecting productivity. The yield prediction model is used to calculate the net yield
at the end of the season. The net return in agriculture is calculated using Equation (16).

R = Y ∗ Py − W ∗ Pw (16)

where Y is the yield at the end of the season, Py is the price of the yield, W is the total
amount of water used for irrigation, and Pw is the price of the water.
To show an application of the model, the net return for tomato yield in 2018 and 2019
was calculated under random irrigation every five days. An RF model was implied to
predict WcTot after each irrigation. The model receives five days of climate data and irriga-
tion rate and predicts WcTot for the next day. The RF model predicts WcTot with an R2
score of 0.80. Algorithm 1 was used to calculate the net return at the end of the season
under random irrigation. To calculate the net return, the cost of irrigation per 1 ha-mm/ha
was assumed to be nearly 0.5 USD [43], and tomato prices were assumed to be nearly
728.20 USD/ton (www.tridge.com (accessed date 22 May 2021)).
Energies 2021, 14, 3004 18 of 21

Algorithm 1: Algorithm
Env step(current_state, action):
season_state = List ();
next_WcTot = RF_WcTot.predict (current_state, action);
season_state.append (current_state, action);
if time_passed = end_o f _season then
done = True;
Y = BLSTM _yield.predict (season_state);
calculate reward from Equation (16);
else
done = False;
reward = 0;
time_passed = time_passed+1;
return next_WcTot, reward, done
steps = 5;
n_steps = 2;
env = Env ();
for i = 1 to n_steps do
state = state [0];
action = 0;
done = False;
while done = False do
for k = steps to steps + 5 do
if k = steps then
action = random_action (0, 60);
else
action = 0 ;
new_WcTot, reward, done = env.step (states, action, k);
state = new_WcTot + climate_data

Tables 10 and 11 show the net returns and the parameters used in Algorithm 1.

Table 10. The net return of random irrigation.

Irrigation Yield Total Irrigation Net Return

(ton/ha) (ha-mm/ha) (USD/ha)
(2018) 5.04 660 3344
(2019) 3.84 770 2415

Table 11. The parameters used in Algorithm 1.

Param Explanation
action volume of water
Done a Boolean value represent whether a season is complete
n_steps number of season
steps time step between irrigation
state climate big data and WcTot
new_WcTot WcTot after irrigation
time_passed time elapsed after the start of the season
state [0] the data from the first 5 days of the season

In this example, the irrigation amount was randomly selected, but in future work, a re-
inforcement learning agent is trained to select the best irrigation amount, and the random
Energies 2021, 14, 3004 19 of 21

action (water amount) is replaced by the model. In Deep Reinforcement Learning algo-
rithms, there is an environment that interacts with an agent. During the training, the agent
chooses an action based on the current state of the environment, and the environment
returns the reward and the next state to the agent. The agent tries to choose the action that
maximizes the reward [44]. In the agricultural domain, the state of the environment can
be defined as the climate data and the water content of the soil; the action is the amount
of irrigation, and the reward is the net return. The function Env() in Algorithm 1 is used
as the environment. An agent can be trained to select the amount of irrigation based
on the condition of the field. Therefore, the yield estimation model can be used as part of
the environment of Deep Reinforcement Learning to calculate the reward of the agent. This
system can be trained end-to-end.

4. Conclusions
Irrigation and storage are closely related to the use of energy. Efficient irrigation
minimizes unnecessary water use, which contributes to energy conservation. Yield esti-
mation models help to reduce energy consumption, increase productivity, and estimate
labor requirements for harvesting and storage requirements. RNN models offer several
advantages over other deep learning models and traditional machine learning approaches.
The most important aspect is their ability to process time-series data such as agricultural
datasets. In this work, the ability of RNN models to predict tomato and potato yields based
on climate data and irrigation amount was investigated. The LSTM, GRU, and their exten-
sion BLSTM and LSTM models were trained on sandy loam soil for crop yield prediction.
The results show that the use of BLSTM models outperformed the simple LSTM, GRU,
and BGRU models on the validation set. In addition, the LSTM model performed better
than the GRU model in the validation set. Therefore, removing the cell state from the LSTM
nodes could be problematic in our context.
The BLSTMs achieved an R2 score of 0.97 to 0.99 on the test set. The results show
that BLSTM models can automatically extract features from raw agricultural data, capture
the relationship between climate data and irrigation amount, and convert it into a valuable
model for predicting future field yields. One drawback of these models is that a sufficient
amount of clean data is needed to train the model. With more data, the model can make better
predictions. In this work, the simulated yield dataset was used to overcome this disadvantage.
The performance of the BLSTM was compared with the CNN model, MLP, and RF,
and it was found that the BLSTM outperformed the MLP networks and CNN and RF
in yield prediction. The CNN model achieved the second-best performance and MLP
the worst performance. The results show that past observations of a season are important
in yield prediction. The BLSTM model can capture the relationship between past observa-
tions and the new observations and predict the yield more accurately. One disadvantage of
the BLSTM model was that the training time of the BLSTM model was higher than other
implemented models.
One of the applications of the yield prediction model is to develop an end-to-end
decision support system that automatically decides when and how much to irrigate. Deep
Reinforcement Learning models are used to build such a system. An agent can be trained
to select the amount of irrigation based on the condition of the field. To train such a model,
a reward function must be developed. In the agricultural domain, the net reward is used as
the reward for the agent. Since it is difficult and time-consuming to work with a simulation
system such as Aquacrop to train a deep learning model, the yield estimation model is
used to determine the net return of the agent at the end of each season, and the agent can
decide based on this reward. This system can help farmers to decide when and how much
to irrigate and reduces water consumption without affecting productivity.

Author Contributions: K.A.: investigation, methodology, writing—original draft, and writing—review.

P.D.G.: supervision, writing—review, project administration, and funding acquisition. T.M.L.: writing—
review and editing. All authors have read and agreed to the published version of the manuscript.
Energies 2021, 14, 3004 20 of 21

Funding: This work is supported by the project Centro-01-0145-FEDER000017-EMaDeS-Energy,

Materials, and Sustainable Development, co-funded by the Portugal 2020 Program (PT 2020), within
the Regional Operational Program of the Center (CENTRO 2020) and the EU through the European
Regional Development Fund (ERDF). Fundação para a Ciência e a Tecnologia (FCT—MCTES) also
provided financial support via project UIDB/00151/2020 (C-MAST).
Acknowledgments: We would like to express our sincere gratitude for the support provided by
AppiZêzere and DRAP-Centro with the data from the meteorological stations near Fadagosa. P.D.G.
and T.M.L. acknowledge Fundação para a Ciência e a Tecnologia (FCT—MCTES) for its financial
support via the project UIDB/00151/2020 (C-MAST).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Sundmaeker, H.; Verdouw, C.N.; Wolfert, J.; Freire, L.P. Internet of Food and Farm 2020. In Digitising the Industry; Vermesan, O.,
Friess, P., Eds.; River Publishers: Gistrup, Denmark, 2016; pp. 129–150.
2. Nguyen, G.; Dlugolinsky, S.; Bobak, M.; Tran, V.; Garcia, A.L.; Heredia, I.; Malik, P.; Hluchy, L. Machine Learning and Deep Learning
frameworks and libraries for large-scale data mining: A survey. Artif. Intell. Rev. 2019, 52, 77–124, [CrossRef]
3. Hayes, M.J.; Decker, W.L. Using NOAA AVHRR data to estimate maize production in the United States Corn Belt. Int. J.
Remote Sens. 1996, 17, 3189–3200. [CrossRef]
4. Bargoti, S.; Underwood, J.P. Image Segmentation for Fruit Detection and Yield Estimation in Apple Orchards. J. Field Robot. 2017,
34, 1039–1060. [CrossRef]
5. Habaragamuwa, H.; Ogawa, Y.; Suzuki, T.; Shiigi, T.; Ono, M.; Kondo, N. Detecting greenhouse strawberries (mature and imma-
ture), using deep convolutional neural network. Eng. Agric. Environ. Food 2018, 11, 127–138. [CrossRef]
6. Kang, H.; Chen, C. Fast implementation of real-time fruit detection in apple orchards using deep learning. Comput. Electron. Agric.
2020, 168, 105108. [CrossRef]
7. Koirala, A.; Walsh, K.B.; Wang, Z.; McCarthy, C. Deep learning for real-time fruit detection and orchard fruit load estimation:
Benchmarking of ‘MangoYOLO’. Precis. Agric. 2019, 20, 1107–1135. [CrossRef]
8. Liang, Q.; Zhu, W.; Long, J.; Wang, Y.; Sun, W.; Wu, W. A Real-Time Detection Framework for On-Tree Mango Based on SSD Net-
work. In Intelligent Robotics and Applications, Proceedings of the 11th International Conference, ICIRA 2018, Newcastle, NSW, Australia,
9–11 August 2018; Springer International Publishing: Cham, Switzerland, 2018; pp. 423–436.
9. Stein, M.; Bargoti, S.; Underwood, J. Image Based Mango Fruit Detection, Localisation and Yield Estimation Using Multiple View
Geometry. Sensors 2016, 16, 1915. [CrossRef]
10. Tian, Y.; Yang, G.; Wang, Z.; Wang, H.; Li, E.; Liang, Z. Apple detection during different growth stages in orchards using
the improved YOLO-V3 model. Comput. Electron. Agric. 2019, 157, 417–426. [CrossRef]
11. Apolo-Apolo, O.; Martínez-Guanter, J.; Egea, G.; Raja, P.; Pérez-Ruiz, M. Deep learning techniques for estimation of the yield
and size of citrus fruits using a UAV. Eur. J. Agron. 2020, 115, 126030. [CrossRef]
12. Maimaitijiang, M.; Sagan, V.; Sidike, P.; Hartling, S.; Esposito, F.; Fritschi, F.B. Soybean yield prediction from UAV using
multimodal data fusion and deep learning. Remote Sens. Environ. 2020, 237, 111599. [CrossRef]
13. Yang, Q.; Shi, L.; Han, J.; Zha, Y.; Zhu, P. Deep convolutional neural networks for rice grain yield estimation at the ripening stage
using UAV-based remotely sensed images. Field Crop. Res. 2019, 235, 142–153. [CrossRef]
14. Chen, Y.; Lee, W.S.; Gan, H.; Peres, N.; Fraisse, C.; Zhang, Y.; He, Y. Strawberry Yield Prediction Based on a Deep Neural Network
Using High-Resolution Aerial Orthoimages. Remote Sens. 2019, 11, 1584. [CrossRef]
15. Zhou, Z.; Song, Z.; Fu, L.; Gao, F.; Li, R.; Cui, Y. Real-time kiwifruit detection in orchard using deep learning on AndroidTM
smartphones for yield estimation. Comput. Electron. Agric. 2020, 179, 105856. [CrossRef]
16. Rahnemoonfar, M.; Sheppard, C. Deep Count: Fruit Counting Based on Deep Simulated Learning. Sensors 2017, 17, 905.
[CrossRef] [PubMed]
17. Ma, J.W.; Nguyen, C.H.; Lee, K.; Heo, J. Regional-scale rice-yield estimation using stacked auto-encoder with climatic and MODIS
data: A case study of South Korea. Int. J. Remote Sens. 2019, 40, 51–71. [CrossRef]
18. Han, J.; Zhang, Z.; Cao, J.; Luo, Y.; Zhang, L.; Li, Z.; Zhang, J. Prediction of Winter Wheat Yield Based on Multi-Source Data
and Machine Learning in China. Remote Sens. 2020, 12, 236. [CrossRef]
19. Kim, N.; Ha, K.J.; Park, N.W.; Cho, J.; Hong, S.; Lee, Y.W. A Comparison Between Major Artificial Intelligence Models for Crop
Yield Prediction: Case Study of the Midwestern United States, 2006–2015. ISPRS Int. J. Geo-Inf. 2019, 8, 240. [CrossRef]
20. Abbas, F.; Afzaal, H.; Farooque, A.A.; Tang, S. Crop Yield Prediction through Proximal Sensing and Machine Learning Algorithms.
Agronomy 2020, 10, 1046. [CrossRef]
21. Kamilaris, A.; Prenafeta-Boldú, F.X. Deep learning in agriculture: A survey. Comput. Electron. Agric. 2018, 147, 70–90. [CrossRef]
22. Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet Classification with Deep Convolutional Neural Networks. In Advances
in Neural Information Processing Systems 25; Pereira, F., Burges, C.J.C., Bottou, L., Weinberger, K.Q., Eds.; Curran Associates, Inc.:
Red Hook, NY, USA, 2012; pp. 1097–1105.
Energies 2021, 14, 3004 21 of 21

23. Sermanet, P.; Eigen, D.; Zhang, X.; Mathieu, M.; Fergus, R.; LeCun, Y. OverFeat: Integrated Recognition, Localization and Detection
using Convolutional Networks. arXiv 2013, arXiv:1312.6229.
24. Patterson, J.; Gibson, A. Deep Learning: A Practitioner’s Approach; O’Reilly: Beijing, China, 2017.
25. Hochreiter, S.; Schmidhuber, J. Long Short-Term Memory. Neural Comput. 1997, 9, 1735–1780. [CrossRef]
26. Cho, K.; van Merriënboer, B.; Gulcehre, C.; Bahdanau, D.; Bougares, F.; Schwenk, H.; Bengio, Y. Learning Phrase Representations
using RNN Encoder–Decoder for Statistical Machine Translation. In Proceedings of the 2014 Conference on Empirical Methods
in Natural Language Processing (EMNLP), Doha, Qatar, 25–29 October 2014; Association for Computational Linguistics: Doha,
Qatar, 2014; pp. 1724–1734.
27. Chung, J.; Gulcehre, C.; Cho, K.; Bengio, Y. Empirical evaluation of gated recurrent neural networks on sequence modeling.
In Proceedings of the NIPS 2014 Workshop on Deep Learning, Montreal, QC, Canada, 8–13 December 2014.
28. Schuster, M.; Paliwal, K.K. Bidirectional recurrent neural networks. IEEE Trans. Signal Process. 1997, 45, 2673–2681. [CrossRef]
29. Gal, Y.; Ghahramani, Z. A Theoretically Grounded Application of Dropout in Recurrent Neural Networks. In Proceedings of
the 30th International Conference on Neural Information Processing Systems, NIPS’16; Curran Associates Inc.: Red Hook, NY, USA,
2016; pp. 1027–1035.
30. Prechelt, L. Early Stopping—But When? In Neural Networks: Tricks of the Trade: Second Edition; Montavon, G., Orr, G.B., Müller, K.R.,
Eds.; Springer: Berlin/Heidelberg, Germany, 2012; pp. 53–67.
31. Raes, D.; Steduto, P.; Hsiao, T.C.; Fereres, E. AquaCrop—The FAO Crop Model to Simulate Yield Response to Water: II. Main
Algorithms and Software Description. Agron. J. 2009, 101, 438–447. [CrossRef]
32. Allen, R.G.; Pereira, L.S.; Raes, M.S.D. Crop eVapotranspiration—Guidelines for Computing Crop Water Requirements FAO Irrigation
and Drainage Paper 56; FAO—Food and Agriculture Organization of the United Nations: Rome, Italy, 1998.
33. Montgomery, D.C.; Jennings, C.L.; Kulahci, M. Introduction to Time Series Analysis and Forecasting; Wiley Series in Probability
and Statistics; Wiley: Hoboken, NJ, USA, 2011.
34. James, G.; Witten, D.; Hastie, T.; Tibshirani, R. An Introduction to Statistical Learning: With Applications in R; Springer Publishing
Company, Incorporated: Berlin/Heidelberg, Germany, 2014.
35. Willmott, C.J.; Ackleson, S.G.; Davis, R.E.; Feddema, J.J.; Klink, K.M.; Legates, D.R.; O’Donnell, J.; Rowe, C.M. Statistics for
the evaluation and comparison of models. J. Geophys. Res. Ocean. 1985, 90, 8995–9005. [CrossRef]
36. Zhang, J.; Zhu, Y.; Zhang, X.; Ye, M.; Yang, J. Developing a Long Short-Term Memory (LSTM) based model for predicting water
table depth in agricultural areas. J. Hydrol. 2018, 561, 918–929. [CrossRef]
37. Abadi, M.; Agarwal, A.; Barham, P.; Brevdo, E.; Chen, Z.; Citro, C.; Corrado, G.S.; Davis, A.; Dean, J.; Devin, M.; et al. TensorFlow:
Large-Scale Machine Learning on Heterogeneous Systems, 2015. Software. Available online: tensorflow.org (accessed on 3 March
2020).
38. Chollet, F. Keras. 2015. Available online: https://keras.io (accessed on 3 March 2020).
39. Dreyfus, S. The numerical solution of variational problems. J. Math. Anal. Appl. 1962, 5, 30–45. [CrossRef]
40. Schmidhuber, J. Deep learning in neural networks: An overview. Neural Netw. 2015, 61, 85–117. [CrossRef]
41. Vanhoucke, V.; Senior, A.; Mao, M.Z. Improving the speed of neural networks on CPUs. In Proceedings of the Deep Learning
and Unsupervised Feature Learning Workshop, NIPS 2011, Granada, Spain, 12–17 December 2011.
42. Murtagh, F. Multilayer perceptrons for classification and regression. Neurocomputing 1991, 2, 183–197. [CrossRef]
43. Rodrigues, L.C. Water Resources Fee in Portugali, 2016. Led by the Institute for European Environmental Policy. Available
online: https://ieep.eu/ (accessed on 3 March 2021).
44. Sutton, R.S.; Barto, A.G. Reinforcement Learning: An Introduction, 2nd ed.; The MIT Press: Cambridge, MA, USA, 2018.