Scribd
Upload
63 views

Forest Fire Detection Using Image Processing

The document describes a method for detecting forest fires using image processing. It begins with an introduction explaining that forest fires cause environmental and economic damage, and that remote sensing techniques can help overcome limitations of traditional detection methods. It then discusses using image thresholding to detect fires from pictures taken by drones by partitioning images based on heat levels. The objectives are to automatically detect forest fires worldwide using infrared images from satellites. The methodology applies hot object segmentation and binary thresholding to identify fire pixels above a certain heat and area threshold. The design analyzes input, edge detection, pixel intensity detection and connected component analysis to label bright regions as potential fires. Testing on over 100 images achieved around 90% efficiency in detection.

Uploaded by

Chilupuri Sreeya
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
63 views

Forest Fire Detection Using Image Processing

The document describes a method for detecting forest fires using image processing. It begins with an introduction explaining that forest fires cause environmental and economic damage, and that remote sensing techniques can help overcome limitations of traditional detection methods. It then discusses using image thresholding to detect fires from pictures taken by drones by partitioning images based on heat levels. The objectives are to automatically detect forest fires worldwide using infrared images from satellites. The methodology applies hot object segmentation and binary thresholding to identify fire pixels above a certain heat and area threshold. The design analyzes input, edge detection, pixel intensity detection and connected component analysis to label bright regions as potential fires. Testing on over 100 images achieved around 90% efficiency in detection.

Uploaded by

Chilupuri Sreeya
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 16

Forest Fire Detection Using

Image Processing
Submitted by:
Harsh Agarwal (18IT117)
Adarsh Naidu (18IT102)
Amith Bhat Nekkare (181IT105)
Introduction

A forest fire is a phenomenon which can be defined as an uncontrolled fire in an area of


combustible vegetation. They can result in a great number of environmental disasters,
causing vast economical and ecological losses. In order to prevent this, forest fire
monitoring and detection have become a significant solution, which attract an increasing
interest around the world.
Conventional forest fire detection techniques make use of watchtowers and human
observers to search and observe fires. It consumes tremendous labour, threatens
observers’ safety and costs a great deal of time.
Introduction

Owing to the development of modern technologies, more advanced forest fire detection
approaches include integrating remote sensing techniques with various platforms (such
as satellites, ground-based equipments, and aircrafts), which are designed to overcome
drawbacks of traditional methods.In our project, we propose a method to detect fires
from a picture taken by UAVs, drones, etc. by using image thresholding.Thresholding is a
method used to partition images on the basis of some input parameters.
Literature Survey
Sl No. Paper Title and Author Work Done

1. C.Yuan,Z.Liu and Y.Zhang,”Fire detection


A paper which deals with fire detection using IR
using infrared images for UAV-based forest
fire surveillance”, 2017 International images for UAV-based devices. In this paper, an
Conference on Unmanned Aircraft Systems
algorithm which makes use of brightness and
(ICUAS), IEEE Xplore.
motion clues along with histogram-based
segmentation and optical flow approach for fire
pixels detection is used for detecting fire in IR
video sequences.

2. M.Srutek and T.Andrysiak, Flame Detection


A method for the detection and localization of
based on Infrared Images,Problemy
Exploatacji – Maintenance Problems,2-2013. flames with the use of infrared images, was
proposed.
Literature Survey

Sl No. Paper Title and Author Work Done

3. J.Seebamrungsat,S.Praising,P.Riyamongkol, Fire
An algorithm to detect different types of
detection in the buildings using image processing,
2014 Third ICT International Student Project burning objects, by analyzing their burning
Conference (ICT-ISPC), IEEE Xplore.
temperature, is proposed.
Objectives

To automatically detect forest fires around the world by using infrared(IR) images sourced
from satellites and other sources using different Image Processing Techniques.
Requirement Analysis

● Hardware Requirements
○ Processor: Intel Pentium IV or above
○ RAM: 2 GB or above
● Software Requirements
○ Operating System: Windows 7 / Ubuntu 14.04/ Debian or above
○ Python 3.6 or above
○ OpenCV, Imutils Libraries
Methodology
Methodology

The input is an infrared image which may or may not have fire pixels in it.

We then perform the hot object segmentation, which divides the image
parts on the basis of heat emitted, on the input image. Upon analysis of
the identified IR bands, we can detect the fire pixels, if it exists.
Methodology

If no fire pixels are found, the next input image is called for and the
process starts again. Else, binary thresholding is performed on the fire
pixels to identify those fires whose heat is above a particular heat and area
value to remove human settlements,etc. from consideration. Finally, a blob
counter is used and an output image is returned with the identified fire
spots.
Design Analysis

A. Input of Dataset
a. The images for processing are stored in a separate folder.
B. Contour/Edge Detection
a. The points at which image brightness changes sharply are typically organized into a set of
curved line segments termed edges.
b. The result of applying an edge detector to an image may lead to a set of
connected curves that indicate the boundaries of objects.
C. Pixel Intensity Detection
a. The pixels are converted into grayscale values by taking average of all the three
RGB values and the intensity is compared.
Design Analysis

D. Detection of Fire

a. To reveal the brightest regions in the blurred image we need to apply thresholding: This
operation takes any pixel value p >= 200 and sets it to 255 (white). Pixel values <
200 are set to 0 (black).
b. We perform connected component analysis on the resulting image to identify blobs of
bright regions and label them with a number. All these blobs are iterated through and
any blob with less number of pixels are discarded.
c. These blobs are highlighted with a red circle and overlapped on the original image.
Timeline/Gantt Chart
Conclusion

● The program is successfully able to detect the forest fire.An image


of a forest is taken as input and the output image will highlight the
areas of fire according to the heat intensity.
● The required fire regions will be encircled and each region will be
mapped to its count.The program also works for regular coloured
images in addition to grayscale Infrared Images.
● The program was tested on a set of over 100 different sized images
and by analysis, the overall efficiency is coming out to be around 90% .
Conclusion
Thank You!

You might also like