Research on Detection Method of Papaya Maturity Based on Yolov8n-Arc

The maturity of papaya is mainly judged manually at present, and there is a lack of fast and accurate automatic detection method for the maturity of papaya. Based on YOLOv8n model, this study proposes a method to detect the maturity of papaya in natural environment. Firstly, scale feature interaction mechanism AIFI based on attention mechanism is introduced into the backbone network, focusing on the advanced feature extraction of papaya images. Secondly, the public convolution in the backbone network is replaced by the receptive field coordinate attention convolution RFCAConv, focusing on the key areas of papaya skin with specific color changes, which enhances the spatial expression of features. The hybrid attention mechanism (CBAM-SKAttention, CB-SK) is incorporated into the neck network to enhance the model's ability to extract and integrate features and reduce the interference of cluttered background information. Therefore, a papaya maturity detection model (YOLOv8n-AIFI-RFCAConv-CB-SK, YOLOv8n-ARC) has been formed. The results showed that the mean average precision mAP of the improved model for the detection of papaya maturity was 87.83 %, which was 2.90 percentage points higher than the original YOLOv8n model. the average detection time is increased by $0.7 / \text{ms}$. Therefore, YOLOv8n-ARC has improved the accuracy and detection speed of the model, providing technical support on intelligent picking of papaya.