A Lightweight Cherry Tomato Maturity Real-Time Detection Algorithm Based on Improved YOLOV5n

To enhance the efficiency of mechanical automatic picking cherry tomatoes in a precision agriculture environment, this study proposes an improved target detection algorithm based on YOLOv5n. The improvement steps are as follows: First, K-means++ clustering is utilized to update scale and aspect ratio anchor box, adapting it shape characteristics tomatoes. Secondly, coordinate attention (CA) mechanism introduced expand receptive field range reduce interference from branches, dead leaves, other backgrounds recognition tomato maturity. Next, traditional loss function replaced by bounding box regression with dynamic focusing (WIoU) function. outlier degree nonmonotonic address boundary balance problem between high-quality low-quality data. This research employs self-built dataset train algorithms before after improvements. Comparative experiments conducted YOLO series algorithms. experimental results indicate that model has achieved 1.4% increase both recall compared previous model. It achieves average accuracy mAP 95.2%, time 5.3 ms, weight file size only 4.4 MB. These demonstrate fulfills requirements for real-time lightweight applications. highly suitable deployment embedded systems mobile devices. presented paper enables maturity provides rapid accurate guidance achieving