Image Edge Detection Algorithm Research Based On Blackfin

Edge detection is used widely in image processing, machine vision and pattern recognition. It's significant for feature extraction, model identification and image understanding. Wavelet is a powerful mathematical analysis tool for time-frequency, which are emerged in recent years. Image processing methods with wavelet have achieved good results in image signal analysis. Due to the feature of muti-analysis, wavelet is also suitable for edge detection.After reviewing the existing edge detection algorithms and analysing the advantages respectively, this paper proposed a noval algorithm of edge detection. Expoilting the phase invariance of stationary wavelet, detect each sub-band's edge with Sobel algorithm and integrate the information through image infusion technology. The experiments show that the algorithm can preserve more detail information; and exploring pulse coupled networks we propose a pre-processing alogorithm. And we test our scheme by non-sample contourlet again.By comparing performance with other DSPs and according to the application in this paper, Blackfin 531 DSP is choosed as the central processor of the image processing system. And the proposed wavelet edge detection algorithm is complemented on this platform. The results of simulation proved that the proposed algorithm worked well on the hardware platform, and better edge detection effects are achieved.In the end, the work in this paper is summed up, and the further research topics is mentioned.