The invention discloses a carrot surface defect detection method based on image processing. The method comprises image preprocessing and detection of four kinds of defects comprising fibrous roots, bending, cracking and fracturing influencing carrot appearance quality. A carrot image is preprocessed at first, and then fibrous roots, bending, cracking and fracturing are detected respectively. In fibrous root detection, a concave point detection method is adopted for gradually judging points on the carrot contour; in bending detection, a circumscribed convex polygon in the carrot area is obtained, and the area ratio of the circumscribed convex polygon to the carrot area is calculated; in cracking detection, Canny edge detection and Hough conversion are carried out on an R component image. Upper portion cracking and lower portion cracking are detected on the basis of judging the carrot direction. The carrot surface defects are automatically detected by adopting the imaging processing technology, carrot appearance quality detection efficiency and accuracy can be greatly improved, and the labor cost is greatly saved.