Vehicle Detection in SAR Imagery Using YOLOv7: Challenges and Performance Evaluation

The acquisition of electro-optical imagery is often disrupted by unfavorable weather conditions, such as cloud cover, fog, or the absence of daylight. In such situations, radar imagery becomes an indispensable data source, enabling effective monitoring regardless of weather conditions. This study presents research on vehicle detection in radar imagery using a deep learning model based on the YOLOv7 architecture. The model was trained and validated on radar data to evaluate its effectiveness in detecting vehicles under various conditions. Experimental results demonstrated that the model successfully detected approximately 70% of vehicles in test scenarios, confirming its robustness in various situations.