Automated Classification and Quality Analysis of Rice and Wheat Grains Using Deep Learning

The manual inspection of cereal grains such as rice and wheat for classification and quality control is a traditionally time-consuming, labor-intensive, and subjective process. This paper presents a robust, automated system designed to provide a rapid, objective, and accurate analysis of grainqualityfromdigitalimages.Oursystememploysahybridapproach,combiningthesophisticated feature extraction power of deep learning with precise computer vision algorithms. A Convolutional NeuralNetwork(CNN)istrainedtoclassifyriceintofivedistinctvarietiesandwheatintofourquality categories, achieving high classification accuracy. For a more granular quality assessment, the system integrates two specific algorithms: (1) A geometry-based method using rotated bounding boxes to accurately measure the morphological characteristics (length and width) of individual rice grains for grading purposes. (2) The Watershed segmentation algorithm to effectively separate and count touching or overlapping grains, allowing for the quantitative distinction between full and broken grains. All components are integrated into a user-friendly web application developed with the Django framework, demonstrating a practical and efficient end-to-end solution for automating grain quality analysis in the agricultural and food processing industries.