Detection Of Human In Flames Using HOG & SVM

Detection Of Human In Flames Using HOG & SVM

Manish Rajput
Dept. of Information Technology SVKM Iot Dhule
Dhule, India
manishrajpu2442@gmail.com

Akshay Badgujar
Dept. of Information Technology SVKM Iot Dhule
Dhule, India
akshaybadgujar740@gmail.com

Madhuri Patil
Dept. of Information Technology SVKM Iot Dhule
Dhule, India
madhuridesale11120@gmail.com

Yogesh Gaikwad
Dept. of Information Technology SVKM Iot Dhule
Dhule, India
imyogeshgaikwad@gmail.com

Jyotiraditya Samudre
Dept. of Information Technology SVKM Iot Dhule
Dhule, India
jyotiradityasamudre1551@gmail.com

Abstract— The project helps in finding people who are trapped in the fire. The project can be divided into two overlapping segments: the first is identifying the fire, and the second is detecting the human. The module operates using the colour standard YCbCr, and the fire detection algorithm identifies fire and flames in the environment if they are there. To find a human in the fire, it employs the Histogram of Oriented Gradients (HOG) and the Support Vector Machine (SVM). It details needed motion-based selected features for human activity recognition as videos. To make the two modules function together, we integrated them. For such detection of a fire, we used a learning method that makes use of a trained model with a wide range of human feature sets. Color median filter and background differencing are used in four separate rules to detect moving objects. The requirement that if a Fire occurs, then humans must be discovered using the method unites the two algorithms. The major goal of this technique is to locate the trapped people in the flames so they may be immediately rescued. This can aid firefighters in making important plans and locating dangerous zones.

Keywords— OpenCV, Machine learning Algorithms like Histogram of Oriented Gradients (HOG) and the Support Vector Machine (SVM)., Image processing, YCbCr component ,  Motion detection.