A fast algorithm for the generalized Hough transform (GHT) based on the use of a hierarchical processing scheme and the inverse generalized Hough operation is proposed. By reducing the size of the image portion which need be processed in the proposed fast GHT, not only the computation time but also the number of processing elements for parallel processing can be reduced. The way to apply the proposed fast recursive GHT on pyramid machines is discussed. Some experimental results are also included to demonstrate the applicability of the approach.
