A Method for Detecting Skewed Symmetry Axis of Objects

Abstract

This paper presents a method for detecting skewed symmetry axed of objects. The adaptive Hough transform, an iterative approach that requires small memory size, is used to determine threee parameters of the axis feature. Location of the memory address in three dimensional space is determined from the position of each paired-pixel in the edge of the image. Parameters are determined by finding the local peak location in space. Paremeter range is confined by the gradient of each pixel, which results in a clear calculation of the local peak and computing cost saving. Experimental results for both computer generated and natural objects show that this method is applicable to both curved and linear objects and is suitable for a small computer system.