Multi-targets attitude recognition method based on Hough and improved Radon transform

Abstract

Traditional Hough transform can not locate the accurate position of line segments. It is too complex in practical liner target detection based on improved Radon. Both transforms is used associated in this paper in order to recognize the attitude of multi-targets as line segments accurately with an acceptable complexity in calculation. A new method called fast approaching algorithm using changing pace nonius length is also presented to determine the scale of targets quickly. Self-adaptive precision is also used in the processing.