Rice chalkiness quality detection method and system based on computer vision

Abstract

This invention belongs to the field of agricultural product quality testing technology, and discloses a computer vision-based method and system for detecting chalky rice quality. This method enables high-precision, interference-resistant, automated quality testing of rice populations without the need for individual grain segmentation. The method first acquires image samples of hulled rice populations and their corresponding global quality labels. Then, a deep neural network regression model is constructed and trained using a twin consistency constraint strategy. During training, the main regression branch and the auxiliary consistency branch share weights, and a hybrid loss function is used to simultaneously constrain the difference between the predicted value and the true label, as well as the consistency between the prediction results of the original image and the transformed image. Finally, the trained model is used to directly perform end-to-end quality prediction on the images of the hulled rice populations to be tested.

Images