Research on intelligent vehicle object detection algorithm based on DC-YOLOv8s

Aiming at the shortcomings of current mainstream detection algorithms in intelligent vehicle environment awareness, a design scheme of intelligent vehicle object detection system based on DC-YOLOv8s (Deformable Convolution-You Only Look Once version 8s) is proposed. The intelligent vehicle object detection system based on DC-YOLOv8s is composed of the preprocessing module, the feature extraction module, the feature fusion module and detection module. The preprocessing module performs geometric transformation and data enhancement on the input image. The feature extraction module is used to extract the feature information of the image to be detected. The function of the feature fusion module is to fuse low-level features with high-level features, and the detection module predicts the category and location of the detection target according to the output of the feature fusion module. By means of experiments, the feasibility of intelligent vehicle object detection system based on DC-YOLOv8s is verified. Through the analysis of experimental results, the detection system can give full play to the robustness of deformable convolution and the effectiveness of ICIoUL (Improved Complete Intersection of Union Loss) for improving positioning accuracy. Experimental results show that our system's object detection accuracy can reach 94.6%. Compared with the YOLOv8s model, the Precision is increased by 7.4%, the Recall is increased by 6.9%, and the mAP@0.5 is increased by 6.5%.