Research on Winter Jujube Object Detection Based on Optimized Yolov5s

Winter jujube is a popular fresh fruit in China for its high vitamin C nutritional value and delicious taste. In terms of winter object detection, machine learning research, small size fruits could not be detected with accuracy. Moreover, deep due to the large model network slow detection speed, deployment embedded devices limited. this study, an improved Yolov5s (You Only Look Once version 5 model) algorithm was proposed order achieve quick precise detection. algorithm, we decreased parameters by reducing backbone improve speed. Yolov5s’s neck replaced slim-neck, which uses Ghost-Shuffle Convolution (GSConv) one-time aggregation cross stage partial module (VoV-GSCSP) lessen computational complexity while maintaining adequate Finally, knowledge distillation used optimize increase generalization boost overall performance. Experimental results showed that accuracy optimized outperformed occlusion target discrimination, as well Compared Yolov5s, Precision, Recall, mAP (mean average Precision), F1 values were increased 4.70%, 1.30%, 1.90%, 2.90%, respectively. The Model Parameters both reduced significantly 86.09% 88.77%, experiment prove from can provide real time method robot harvesting.