Computer vision is used for monitoring and classification analysis in today’s society. The monitoring error rate of children’s picture books and the manual efficiency of classification are too slow, so it is difficult to find suitable books in a short time. Computer vision can help monitor and classify children’s picture books based on natural processing techniques. In this paper, the comparison between classical calculation and evolutionary calculation is used to prove that evolutionary calculation has a better accuracy, and the evolutionary calculation is further studied, finally reaching 97.8% accuracy in monitoring. This kind of evolutionary calculation should be vigorously developed in the future.