Window Location for Envelope Image Based on Improved Hough Transform

To solve the problem of huge computation in the solution process of using Hough Transform (HT), an improved HT algorithm based on Freeman criteria has been presented and used to locate window region for the envelope image successfully. The improved algorithm detects the whole image using a template with size of N*N and the center pixel of template, replacing other pixels in template, is calculated by HT to decrease the number of computing pixels. Besides, the frame of the window has spatial geometrical relationship, the center pixels of template on the window frame could be transplanted within the scope of the parameter θ to decrease the computing times. On solving the problem of window location in envelope, experiment result indicates that the improved algorithm is 9 times as fast as the standard one, and more accurate and faster than other improved HT and Based on Connected Components algorithm.