Research on Fresh Quality Detection of Citrus Based on Computer Vision and Neural Network

摘要

In this study, the Tribute Citru was taken as the research object. The color image of Tribute Citru was segmented under B component of color image collected by computer vision system, and the hue frequency sequence of Tribute Citru was extracted under H component of color image as the color feature of Tribute Citru epidermis. Then the structure of the neural network was constructed and optimized, and the mapper structure for nondestructive detection of Tribute Citru fresh quality was 141-8-1. The results showed that the accuracy of computer vision and neural network technology in detecting the taste parameters of fresh Tribute Citru was 86.67%.