In the field of computer vision, vehicle object detection has been a topic of significant and complex interest. With the rise of intelligent transportation systems and autonomous driving technology, the importance of vehicle object detection continues to be highlighted. Given the current issues of low precision, high miss rate, and poor robustness in existing algorithms, this study introduces an improved vehicle detection algorithm, SSB-YOLO, based on the YOLOv8 model. The SSB-YOLO algorithm integrates the Shuffle Attention mechanism to filter out unimportant factors and enhance model performance; it also incorporates the spatial and channel reconstruction convolution mechanism to reduce spatial and channel redundancy between features in convolutional neural networks. Furthermore, a new and better algorithm based on Wise-IoU optimization is proposed, which yields superior bounding box regression performance throughout the training period. The model demonstrated improved detection accuracy and reduced computational cost. The experimental results indicate that, compared to the YOLOv8n model, SSB-YOLO achieves a 1.6% increase in mAP@50. This approach outperforms other object detection algorithms, enhancing the overall system's robustness and accuracy and thereby providing higher precision in the field of vehicle detection.