The detection system for stored-grain pests based on image processing technology has many advantages of rapidness, accuracy, visualization and etc. In order to take full advantages of the foreign and native research findings, improve its research and application in this field in China, the research progress of automatic recognition for stored-grain pests outside and inside grain kernels was summarized by utilizing computer image processing technology based on three methods, such as visible image, near infrared image and soft X-ray. The superiority and inferiority of each method was analyzed. Further key research work from five aspects was presented, such as new feature extraction, recognition of live and dead pests, detection of early pest infestation, dynamic image technique, development of automatic separating and transmitting machine, so the automatic detection of stored-grain pests could be actualized. 