The method of parallel Gabor wavelet transform edge detection based on CUDA

In view of the limitations of classic differential operators and the defects of traditional Gabor wavelet transform algorithm for edge detection, we present a parallel Gabor wavelet transform method for edge detection based on CUDA in this paper. Firstly we introduce the wavelet transform algorithm, and then discusse the feasibility that parallel detect edge in wavelet transform and relize it,the experimental results show that the parallel algorithm can take full advantage of GPU´s parallel processing capability, not only to achieve the automation of image edge detection, but also greatly shorten the processing time of the algorithm.