Environmental Target Detection Technology for Autonomous Driving Based on Improved YOLO v7

Aiming at the problem of low accuracy of road environment detection in automatic driving technology, an automatic driving road environment target detection technology based on improved YOLO v7 algorithm is proposed. The road environment data set is constructed by image acquisition. The CBAM is added to the YOLO v7 backbone network to enhance feature extraction capability. The loss function is replaced by Focal-EIOU Loss to improve detection accuracy. The Deformable ConvNets v2 is added to the YOLO v7 backbone network to enhance the performance of the model by enhancing the modeling ability. After testing, the accuracy of the improved model reached 85.6%, which was 3.9 % higher than that of the original model. The results show that the improved model can detect a series of objects such as cars, pedestrians and various traffic lights in the road environment more accurately, which is of great significance to ensure the safety and practicability of automatic driving and has broad development prospects.