HSI Color Conversion and Hog Feature Extraction in Traffic Light Detection using AdaBSF

Volume 5 Issue 3 July - September 2017

Research Paper

HSI Color Conversion and Hog Feature Extraction in Traffic Light Detection using AdaBSF

J. Jebisha*, V. Monisha**, B. Femila Jemi***

*-*** Department of Electronics and Communication Engineering, Arunachala College of Engineering for Women, TamilNadu, India.

Jebisha, J., Monisha, V., and Jemi, F, B. (2017). HSI Color Conversion and Hog Feature Extraction in Traffic Light Detection using AdaBSF. i-manager's Journal on Digital Signal Processing, 5(3), 20-26. https://doi.org/10.26634/jdp.5.3.13930

Abstract

Real time traffic light detection system is an important application of intelligent transportation system. In different driving environment, traffic light is regarded as a difficult one for various illuminations due to the distance measurements. Adaptive Background Suppresion Filter (AdaBSF) is used for background suppression of the images. In this proposed work, the authors include HSI image conversion and HOG feature extraction methods along with the conventional AdaBSF algorithm. Linear cascaded SVM is used for the classification of SVM based image. The enhanced method has been finally considered to be better for finding the traffic light even under strong sunlight, cloudy seasons, and during night times. The authors have performed traffic light detection in real time and they consider that this technique is efficient than other conventional traffic light detection systems.