Improved Generalized Hough Transform for the Recognition of Irregular Objects
GE Qing-ping · Jisuanji gongcheng · 2007
The generalized Hough transform for the recognition of irregular objects is a powerful method,since it can detect overlapping objects under severe noise.The conventional computation method of the R-table’s index is sensitive to discontinuity and distortion.So the GHT’s robustness is degraded a lot.To solve the problem,this paper proposes a new procedure for calculation of the normal direction of each point.Characterizing the edge points by their normal directions can improve the GHT’s robustness to discontinuity and distortion.Experimental results demonstrates that the improved method does work well for objects recognition.And the shoe sample could be detected accurately despite of noise,discontinuity or geometric distortion in the system of shoe sample CAD.