Managing the illegal use of high beams can reduce the occurrence of night traffic accidents. However, at present, there is no efficient detection method for high beam of night, and relevant traffic regulations cannot be effectively implemented. In order to solve this problem, an algorithm to detect the high beam at night was proposed in this paper. Based on YOLOv3, this algorithm optimizes the network structure of Darknet53, retains the information integrity of the object, accelerates its operation speed, uses standard residual components and dilated convolution to enhance the feature expression ability of the network, and uses K-means++ algorithm to optimize the prior bounding boxes clustering. The experimental results show that the mean average precision (mAP) of the algorithm designed in this paper is 98.88%, which is 50% higher than the efficiency of YOLOv3. The algorithm satisfies the practical requirement and could detect the violation effectively.