Single-pixel margin detection method based on FPGA

The invention relates to a single-pixel margin detection method based on an FPGA, comprising firstly converting a RGBA space image into a gray image, then performing the pre-processing operation, such as fast median filtering and margin enhancement and the like, to a video image; using a Canny margin detection algorithm to compute a gradient of the image and a gradient direction, and performing the processing, such as non-maximum suppression and dual-threshold segmentation and the like, to obtain a relatively fine video image margin. On the basis of the Canny margin detection, a morphological processing method is used for performing the corrosion operation to a margin image by a detection method to further refine the margin so as to obtain the margin processing effect of the single-pixel. The video space conversion, the fast median filtering, the margin detection and the margin refining are integrated in one FPGA by the detection method so as to realize the single-pixel margin detection of the video image as well as customize the FPGA as an IP core, which is convenient for system upgrade and expanding, for realizing adjustable parameter according to specific requirement by utilizing the programmable characteristic of the FPGA.