In order to solve the problem of large-scale speed aliasing of radial velocity of millimeter wave cloud radar in the process of rapidly changing weather, according to the equipment performance and observation characteristics of scanning millimeter-wave cloud radar, a radar speed dealiasing algorithm is designed with the help of image recognition technology and tested and evaluated. With the help of the Sobel operator, the fuzzy index of the edge of the velocity blurred region is identified as the fuzzy line, the data is divided into different regions, and the remaining area is gradually extrapolated based on the most likely non-fuzzy region selected by fuzzy logic, and finally the correction result is checked twice to correct the velocity deviation caused by the fuzzy region recognition error. Selecting 200 sets of experimental data to test and evaluate the applicability and accuracy of the algorithm, the results show that the algorithm can achieve defusing processing of more than 90% of the speed data, and is suitable for various scanning modes such as PPI, sPPI and RHI of millimeter-wave radar, which provides new ideas and methods for improving the data quality of millimeter-wave cloud radar.