In order to instantly acquire the information of hazard levels of plant diseases and pests, a method of segmentation of corn spot image based on RGB color model was proposed. The hazard levels of the whole plant could be calculated through hazard levels of the corn canopy acquired by the results of image segmentation. During the segmentation of crop leaf and disease spots, the method was separated into two steps: ①R, G, B three components were gotten from RGB color images which were caught by numeral camera, and the corn leaves were segmented from the background of the 2G-R-B images by choosing threshold automatically via iterative method. ②The disease spots were segmented from leaves based on R-G images. The results showed that the corn leaf blades and the sickness spots of the 30 images could be picked up except a few non-etiolation infection courts.