Abstract: In this paper, by using computer vision technology, danger degree of cotton insect pests was automatically measured based on inside hole and irregular edge of cotton leaves. The local threshold algorithm was used to segment the cotton leaf from the background and the LOG algorithm was applied to recognize the edge of cotton leaf. The deformity of leaf edge was certified by using dilatation algorithm, and the hole of the cotton leaf was determined by using the edge tracking algorithm. The experiment results showed that the method can effectively judge the danger degree of insect pests of cotton, and the measurement error of danger degree was less than 0.05.