Learning based Obstacle Detection for Textural Path

Abstract

Obstacle Detection is a quite important problem in the field of image processing. Various algorithms have been introduced to solve this problem with different environment. This paper proposes procedure to detect any obstacle in the textural path on the basis of some learning techniques. These learning techniques makes easy to detect any obstacle in Textural path. In this paper we have divided an image into some part, describing as left, right and ground or front part. These separate part matches with the environment by using learning. LEP (Local Edge Pattern) and Colour histogram will be used for efficient matching and front part will be segmented to find the obstacle, if exists. Learning is used to take any decision about the obstacle appearance.