In order to solve the problem that the traditional bridge crack detection algorithm cannot extract cracks accurately, a bridge crack detection algorithm is proposed based on image processing, which is suitable for complex scenes. According to the principle of the deep convolutional generative adversarial network, the bridge crack image generative model is proposed and used to amplify the dataset. For the characteristics of bridge cracks, a bridge crack image segmentation model is constructed based on semantic segmentation. The bridge crack image segmentation model is used to extract the bridge cracks from the high-resolution crack images. The research results show that the proposed algorithm has a better detection effect and a stronger generalization ability in the complex road scenes compared with the existing algorithms.