rkrish97/Human_following_robot

A meticulous and robust system for object detection and tracking is still a milestone in the field of computer vision. The various factors that hinder the accuracy of the model includes illumination, noise in the scene, occlusion effect and pose variations of which the source of illumination and its orientation with respect to the object plays a pivotal role. An illumination variation may result in the tracking algorithm to lose the object in the scene. This paper proposes a simulation of a robot which follows a human under varying illumination. The camera on the robot is aided by a histogram of oriented gradients (HOG) approach with an SVM classifier for detection of the human in the scene.

MATLAB