Phd.

Progressive Report-3
Basaweshwar Engineering College, Bagalkot
Department of Computer Science and Engineering

Name of the Research Scholar: Pavan N.Kunchur

USN – 2BA15PEJ06
Title : Design of frame work-Creating photographic mosaics with Hadoop in
big
data image processing

Work Carried out – A detailed Survey is done on Photographic image

processing
1. Introduction
Image processing is always an interesting field as it gives enhanced visual data for human
simplification and processing of image data for transmission and illustration for machine perception.
Digital images are processed to give better solution using image processing. Techniques such as Gray
scale conversion, Image segmentation, Edge detection, Feature Extraction, Classification are used in
image processing. In this survey studies of different image processing techniques and its methods has
been conducted. Image segmentation is the initial step in many image processing functions like
Pattern recognition and image analysis which convert an image into binary form and divide it into
different regions. The technique used for segmentation is Otsu’s method, K-means Clustering etc. For
feature extraction feature vector in visual image is texture, shape and colour. Edge detector with
morphological operator enhances the clarity of image and noise free images. This survey also gives
information about algorithm like Artificial Neural Network and Support Vector Mechanism used for
image classification. The image is categorized into the receptive class by an ANN and SVM is used to
compile all the categorized result. Overall the survey gives detail knowledge about the techniques
used for image processing and identification.
Image processing is a technique to translate an image into digital form and execute some operation on
it, in order to get an improved image or to retrieve useful data from image. It is a procedure of signal
distribution. The process takes input as an image and then apply efficient algorithms, and the results
may be image, data or features associated with that image [15]. The processing stages start with image
segmentation. There is some desire from image segmentation algorithms. first of them is speed. While
processing for segmentations of an image, it does not want to spend much time. The second is good
shape integration of the object. This will enhance results in picture acknowledgment. If the result of
shape is incomplete, it need to take many properties to record the edge of the over-section results [2].
In computer vision, picture division is the way toward parcelling an advanced picture into various
sections. The objective of division is to disentangle or potentially change the portrayal of a picture
into something that is more important and less demanding to examine. Picture division is regularly
used to find articles and limits in pictures. All the more absolutely, picture division is the way toward
allotting a mark to each pixel in a picture to such an extent that pixels with a similar name share
certain attributes [1]. Division is generally the essential stage in any undertaking to analyse or
interpret an image consequently [3]. Division conquers any hindrance between low-level picture
preparing and abnormal state picture handling. A few sorts of division procedure will be found in any
application including the discovery, acknowledgment, and estimation of items in pictures. Otsu's
division strategy, in light of histogram examination, is extensively applied as a part of different
applications [2]. The approach sections a picture by enhancing the change amongst fragments and, all
the while, limits the difference inside the portions. Proposes an Otsu-strategy adjustment for dividing
hand compositions from an uproarious foundation. In, the Otsu-technique is utilized to extract
different focuses in the data pictures, proposes to extend Otsu 1D-histogrambased technique into a
"2D-Otsu" for division. The first single-edge Otsu strategy expressions for one ideal limit for dividing
the information image into "forefront" and "foundation".
Proposes to apply this unique Otsu technique for acquiring a first layer of content versus foundation;
at that point, a re-thresholding is focused on the foundation pixels to get additional layer of frontal
area content [2]. The frequent layers of content are linked to create the last content division. A
comparative recursive methodology was at that point proposed in. The recursive division thought is
likewise utilized as a part of, again for dim scale pictures as it were. Dependent upon the sort of data
that is the matrix, the photos are separated into pictures of power scale and recorded (each fragment
being a novel number, a scalar) and vector pictures (each portion being a vector, vector number which
in this manner parts into a couple of areas) [3]. Scalar picture constrain is where each pixel regard
(real or normal numbers) is seen as a measure of sparkling force. Scalar recorded picture is a photo in
which the estimation of a pixel is where information can be connected with the shade of the pixel
being mentioned to. Edge recognition is a procedure of finding an edge of a picture. Recognition of
edges in a picture is a critical advance towards comprehension picture highlights. Edges comprise of
significant highlights and contain remarkable data. It essentially reduces the picture size and channels
out data that might be viewed as less important, in this manner preserving the imperative auxiliary
properties of a picture [5]. Most pictures contain some measure of redundancies that can here and
there be displaced when edges are recognized and succeeded amid remaking. This is the place edge
discovery becomes an integral factor. to automate these photointerpretation tasks. It particularly
reiterates on the purpose of the most suitable input data to manage with these two arrangement
problems [6]. Two kinds of optical images have been used: Rapid Eye data and 50cm ground
resolution aerialortho-images.
 2. IMAGE PROCESSING TECHNIQUES
Image processing by digital means has many branches including image recognition, image
segmentation, image compression, etc. It is likewise the fundamental square in numerous applications
like pattern recognition, object identification etc. Image processing normally states digital picture
processing, yet process like optical and analog are additionally being possible. This survey is all about
general techniques that applied to them. The recovery of pictures (delivering the input info in any
case) is referred to as imaging. Image processing techniques isolate the discrete colour planes of an
image and then apply standard signal-processing approaches to them. Images are also regards as
three-dimensional signals. There are few papers which describe about image processing techniques.
2.1 A Study and Comparison of Different Image Segmentation Algorithms [2] Image segmentation is
a procedure, which split a picture, which are comparative in some viewpoint and change over it into
paired frame for preparing. Segmentation process is the primary step in many image processing.
Procedure incorporates object characteristic and portrayal and detail estimation. Higher request errand
takes after the grouping of object. Hence, classification, imagining of region of interest in any image,
description plays a substantial role in image segmentation. There are numerous segmentation
algorithms available in the literature, which split an image into number of regions based on some
picture attributes like pixel quality esteem, shading, colour, shape etc. These all calculations are
described based on the segmentation strategy utilized. Segmentation method split the region using
different method such as single or multiple thresh holding, segmentation on parallel region,
segmentation using clustering, edge detection, and also segmentation on fuzzy logic technique etc.
The chosen methodology is Otsu's calculation, K-means, quad tree, Delta E, Region developing and
ft. calculations. To check the execution of the calculation, they applied 6 straightforward and complex
pictures accessible in the literature. The obtained result demonstrates the viability of the division. The
paper provides the best approach for segmentations. Advantages: It can segment the image by simply
finding edges in the image. Higher order task follows the classification of object. Disadvantage: The
methods are difficult to identify multiple objects. 2.2 Generalization of Otsu’s Binarization into
Recursive Color Image Segmentation [7] Otsu’s segmentation method, based on thresholding and
histogram study. The method segments an image by maximizing the variance between segments and,
simultaneously, minimizes the variance within the segments. proposes an Otsu-method adaptation for
segmenting hand writings from a noisy background.
Otsu-technique is utilized to remove various focuses in the input pictures. proposes to extend Otsu
1Dhistogrambased method into a “2D-Otsu” for segmentation. The original single-threshold Otsu
method searches for one optimum threshold for segmenting the input image into “foreground” and
“background”. proposes to apply this original Otsu method for obtaining a first layer of text versus
background; then, a re-thresholding is conducted on the background pixels to obtain another layer of
foreground text. The multiple layers of text are combined to generate the final text segmentation
results. A similar recursive strategy was already proposed in Image segmentation can detect regions of
objects, defined by the artist when painting at different shades or colors. These segments can reveal
the contrast of shadows in paintings or some conceptual base patterns in the paintings. The recursive
division thought is additionally utilized as a part of, again for gray scale pictures only. The paper
proposes a recursive thresholding algorithm for Colour images, which can also be generalized to any
multichannel image. Advantages: The method segments an image by maximizing the variance
between segments and, simultaneously, minimizes the variance within the segments. Otsu-strategy is
utilized to remove numerous objectives in the inputs. Disadvantages: Since working of the histogram,
a resulting segmented image has in general more than just n + 1 segments.

3. ANALYSIS TABLE
The following table gives the analysis of techniques and methods used in research papers on image
processing and identification.

Sr. Paper Title Techniques Addressed Issue

No
1 A Smart Phone Image Plant disease recognition Image processing that analyses the
Processing Application technique; Matrix bgw2 is color features of the spots in plant
for Plant Disease Diagnosis constructed parts.
[1]

2 Scaffolding Progress BIM model This method, combining object

Monitoring of LNG Plant recognition techniques, can rapidly
Maintenance Project using estimate the total number of
BIM and Image Processing scaffolding components from images.
Technologies [9]
3 Detection of unhealthy Image acquisition; Pre- The optimum output were obtained
region of plant leaves using processing of input image; with less computation efforts. The
Image Processing and Segment the components framework demonstrates the
Genetic Algorithm [5] using genetic algorithm. effectiveness of proposed calculation in
acknowledgment and order of the leaf
infections.
4 Plant Diseases Detection Agrobot; K-means; HSV; To increase production in agricultural
Using Image Processing ANN; BPNN; CCM; Neural sector it is necessary to
Techniques [8] Network; SURF; RBF; detect diseases on plants and take
SIFT; RDI; GLCM; PCA; accurate measures.
SGDM
5 Color Image Segmentation ISKMO Algorithm The process shows proper
using Morphological edge Combination of K-means segmentation process and segmentation
Detector Algorithm [14] and edge detection operator for noise content image. The
combination of algorithm has reduced
the detection of false edge in
segmentation result.
 6 Image classification using ANN algorithm ANN classify the result based one by
support vector mechanism SVM algorithm one image feature vector.
and artificial neural network SVM integrate all result of ANN.
[12]
7 GPU based parallel Graphic processing Give best thresholding algorithm to get
processing for plant growth unit(GPU) partition of object and environment.
analysis [] Thresholding algorithm The parallelism processing gives more
efficient time in execution result.
8 A study on image k-mean clustering method The paper gives different method or
segmentation using different formula to find k-mean value to get
type of k-means clustering better results.
[13]
9 Content based image Feature extraction using Extraction of data from image using its
retrieval using feature color, shape, texture color and texture
extraction [11]
10 Automated identification of Active shape method Used pattern recognition method to get
two growth stage for -point distribution data and process for entire geometry of
rapeseed plant: Three leaf -local grey texture method plant
and four leaf stage [4]
11 A Study and Comparison of Otsu's algorithm, K-means, Image segmentation process, and
Different Image quad tree, Delta E, Region algorithm for the method based on
Segmentation Algorithms growing and fth algorithms. thresholding, parallel processing
[2] clustering, edge detection, histogram
analysis.
12 An Effective Algorithm for RGB Color, Edge detection, By examining the sudden change in the
Edges and Veins Detection Vein Detection Algorithm intensity values in leaf images the leaf
in Leaf Images [3] edge can be easily detected.
13 Contribution Of Texture Edge Detection Automate these photo-interpretation
And Red-Edge Band For tasks. It especially accentuates on the
Vegetated Areas Detection assurance of the most appropriate
And Identification [6] information to adapt to these two order
issues

References

[1] N. Petrellis, "A smart phone image processing application for plant disease diagnosis." In Modern Circuits and Systems
Technologies (MOCAST), 2017 6th International Conference on, IEEE 2017, pp. 1-4.

[2] V. Kumar, T. Lal, P. Dhuliya, and Diwaker Pant. "A study and comparison of different image segmentation algorithms."
In Advances in Computing, Communication, & Automation (ICACCA)(Fall), International Conference on, IEEE 2016, pp. 1-6.

[3] R. Radha, and S. Jeyalakshmi. "An effective algorithm for edges and veins detection in leaf images." In Computing and
Communication Technologies (WCCCT), 2014 World Congress on, IEEE 2014, pp. 128-131.

[4] Y. Fang, X. Wang, P. Shi, C. Lin, and R. Zhai. "Automatic identification of two growth stages for rapeseed plant: Three leaf and
four leaf stage." In Agro-Geoinformatics (Agro-geoinformatics), 2015 Fourth International Conference on,. IEEE 2015, pp. 148-
153.

[5]V. Singh and A. K. Misra. "Detection of unhealthy region of plant leaves using Image Processing and Genetic Algorithm."
In Computer Engineering and Applications (ICACEA), 2015 International Conference on Advances in, IEEE 2015, pp. 1028-1032.

[6] A. Le Bris, T. Francois, and C. Nesrine, "Contribution of texture and red-edge band for vegetated areas detection and
identification." In Geoscience and Remote Sensing Symposium (IGARSS), 2013 IEEE International, IEEE, 2013 pp. 4102-4105.

[7] Acuña, R. G. Gonzalez, Junli Tao, and Reinhard Klette. "Generalization of Otsu's binarization into recursive colour image
segmentation." In Image and Vision Computing New Zealand (IVCNZ), 2015 International Conference on, IEEE, 2015 pp 1-6.

[8] S. K. Tichkule and D. H. Gawali. "Plant diseases detection using image processing techniques." In Green Engineering and
Technologies (IC-GET), 2016 Online International Conference on, pp. 1-6. IEEE, 2016.
[9] H. Chi, C. Jian, C. Wu, J. Zhu, X. Wang, and C. Liu. "Scaffolding progress monitoring of LNG plant maintenance project using
BIM and image processing technologies." In Research and Innovation in Information Systems (ICRIIS), 2017 International
Conference on, pp. 1-6. IEEE, 2017.

[10] N. Senthilkumaran, and R. Rajesh. "Edge detection techniques for image segmentation–a survey of soft computing
approaches." International journal of recent trends in engineering 1, no. 2 (2009): 250-254.

[11] A. Devbrat, and J. Jha. “A Review on Content Based Image Retrieval Using Feature Extraction ” International Journal of
Advanced Research in Computer Science and Software Engineering Volume3, March 2016.

[12] L. H. Thai, T. S. Hai, Nguyen Thanh Thuy . “Image Classification using Support Vector Machine and Artificial Neural Network”
International Journal on Information Technology and Computer Science,2012, 5, 32-38 .

[13] S. Tharani and L. Sankari. “A Study on Image Segmentation Using Different Types of K-Means Clustering”International Journal
of Advanced Research in Computer Science and Electronics Engineering (IJARCSEE) Volume 33,December 2015