Cognitive Method for Apple Grading Based on Convolutional Attention and YOLOv8s Network

Manual sorting of similar fruits is greatly affected by the subjective factors of the sorting personnel. To address this issue, a cognitive approach for apple grading utilizing convolutional attention and the YOLOv8s network was investigated. A convolutional attention mechanism is designed within the convolutional block for apple feature extraction. This mechanism automatically learns and weights the importance of different spatial positions and channel features, helping the model to capture key information from input apple images better, suppress irrelevant data, and enhance feature extraction. First, a cognitive approach for apple grading utilizing convolutional attention and the YOLOv8s network is proposed to clarify the structure and function of the model. Secondly, considering the material transportation characteristics of the apple sorting site, image enhancement technology is used to preprocess the original input image to construct an apple sample dataset. Finally, the structure of the convolutional attention module is explained, and a cognitive algorithm for apple grading based on convolutional attention and YOLOv8s is provided. To determine how well this method works, a series of comparative experiments was performed. Experimental results show significant cognitive accuracy and calculation speed advantages, providing theoretical support for subsequent online sorting systems.