Vehicle Small Target Detection Algorithm Based on Improved YOLOv7

In response to the intelligence of modern transportation, the demand for vehicle detection has increased dramatically, and vehicles on urban roads often appear in the form of small targets in the video. However, encountering challenges such as limited image resolution, subtle target features, and a narrow field of view, which collectively hinder the effective detection of small objects, we have refined a YOLOv7-Tiny-based algorithm specifically tailored for small target detection. As a result, we have enhanced a small target detection approach utilizing the YOLOv7-Tiny framework to address the difficulties encountered in detecting minute objects.The CBAM attention mechanism is introduced into YOLOv7-Tiny, and the loss function is optimized using the NWD strategy. Finally, it is tested on the traffic camera detection dataset and compared with other target detection algorithms, and the improved network model can achieve 0.737 of [email protected], which is 1.2 percentage points better than the baseline. The detection effect is better than other classical networks.