Abstract: A FPGA-based binary connected component detection and labeling algorithm is proposed, which Extracting 9pixel information at one time, it makes full use of the parallel processing ability of FPGA, Complete 9pixel processing in one clock cycle and quickly marks the image while scans.Finally, build a simple test circuit, transmit the two values image to FPGA from PC, and then FPGA will mark the image.When the mark is finished, FPGA will return the marked image and the time it is used to PC.Comparing the experimental results with the existing image processing methods, it is found that the FPGA can be used as a special image processing equipment for it's low cost and main frequency, power consumption is much lower than Software processing equipment when the speed that FPGA processes images is equivalent with Software processing method.