CAM-YOLO: tomato detection and classification based on improved YOLOv5 using combining attention mechanism

Several images from the dataset were chosen for testing to assess the model's performance, and the detection performance of the CAM-YOLO and standard YOLOv5 models under various conditions was compared. The experimental results affirms that CAM-YOLO algorithm is efficient in detecting the overlapped and small tomatoes with an average precision of 88.1%.