A comparative analysis of techniques for extracting features from the object in image processing

Abstract

Visual informations is very rich to control and pursue mobile robots, like wheeled mobile robots, underwater robots and aerospace mobile robots, etc. These are considered as mobile robots that move in different spaces, their main problem in this case is navigation especially in unknown environments. In effect, this navigation is possible only by the localisation and orientation of the robot by using different embedded sensors. We have the camera which is a necessary sensor in this work, it is an embedded instrument that gives very rich visual information as a sensor complementing the other sensors. In this context, the recognition of objects from visual informations is a main function among the functions very useful in image processing tasks due to its varied applications in the field of robotics. Based on the analysis of this informations and the determination of image features like color or shape or object primitives (points, lines, edges, etc.) or some other features. What interest us in this paper, various feature extraction techniques and classification of point and edge detection are discussed which are required for object recognition showing advantages and disadvantages of the selected algorithms. So, points and edge detection refers to the process of identifying and detecting sharp discontinuities in an image. In this work, we try to develop a novel algorithm using the work on the existing to create a novel detector.