Shape Detection based on Improved Hough Transformation

Hough transform is a common algorithm used to detecting geometry shape of objects in computer vision, lines circles and other shape objects will be detected easily. But classical Hough transformation exists many shortcomings, this paper proposed one kind of improved HT algorithm, it may greatly enhance the recognition speed and the examination precision, reduces the storage space, reduce the amount of computation , and still has characteristics of the classical Hough transform. It also realized one kind of method to withdraw the curve gradually is realized, withdraws the curve in turn from long to short. The computational method of the curve initial and terminal station are given, the experiment has proved the method validity. Based on this, but also has analyzed the shape invariable feature, after obtaining the graph translation, revolving, scaling, the number of Hough transformation peak value has not changed, after the angle normalization, the beginning position of the peak value has the change, but the peak value gap does not change.