Rice Quality Analysis Based on Physical Attributes Using Image Processing and Machine Learning

This paper proposes a methodology to automate the quality control of rice grain, as its been manually analysed by veteran rice inspector which are inaccurate, collection of data of various rice types to analyse quality based on their physical attributes such as Area of the grain, Perimeter, Width and Height, Aspect Ratio, Major and Minor Axis. We have used techniques such as Digital Image Processing, Computer Vision - which involves pre-processing, morphological operation, edge detection, object detection and finally object measurement. Training of the system involving manual and Machine Learning which involves feeding of all data results from image processing into CSV file and formulating hypothesis for manual and SVM for machine based training, and then test to classify the quality into higher or lower grades based on the two methods and concluding the best methodology from observation. © 2023 IEEE.