Vehicle Detection in Low Illumination Based on Attention Mechanism and RetinexNet

In order to solve the problem of low detection ability of target detection algorithm in low illumination environment such as night time, this paper puts forward a method to detect vehicles by improving YOLOv5 network structure and introducing RetinexNet. CBAM attention module is added to the Neck detection layer of the network to extract the main features of the vehicle, reduce the unused feature extraction, and enhance the detection capability of the vehicle. The DIoU is introduced as the loss function of the model to solve the imprecise location of the prediction box and speed up the convergence of the model. RetinexNet is used to enhance and denoise low illumination images so as to improve the detection ability of low illumination images. The experimental results show that the improved model achieves 90% accuracy in vehicle detection and has good detection performance in low light environment.