Comparative Study of Geometric Morphometrics and Machine Learning for On-Demand Rice Grain Grading
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Grain quality plays an integral role to agricultural productivity and global food security, influencing market price, customer preference and trade regulation.Rice grains had been traditionally categorized in 'full' and 'broken' by manual visual inspection which is laborious, time consuming subjective process leading to some errors.However, this traditional procedure needs the demands of modern agricul-tural supply chains, especially for entities such as FCI, CMR, rice millers,etc... where accuracy and efficiency are imperative.This paper proposes a novel mor-phometric image processing method for detection & classification of rice grains.The classification algorithm implemented is rudimentary, yet powerful as it con-structs its own "gold standard" by recognizing the largest grain present in every analyzed micrograph to serve as a reference for 'full' grains, all remaining grains are classified according to either geometric properties.In the framework two main types of approaches are compared: (i) geometric properties and (ii) Machine learning used such as Convolutional Neural Networks (CNN), Random Forest, Support Vector Machines (SVM) and XGBoost to improve classification accu-racy and applicability.This inexpensive, reliable and scalable alternative could be a solution to modernize the quality control in rice processing industry and provide justice to farmers with price premium for their products.
Fields of Study & Subjects
Smart Agriculture and AI
Industrial Vision Systems and Defect Detection
Agricultural Engineering and Mechanization
Publication details
- Journal / source
- Advances in intelligent systems research/Advances in Intelligent Systems Research
- Pages
- 364–374
- Publication date
- 1 April 2026
- Language
- EN
- Document type
- Book chapter