Title: Machine learning-based approach for degree of milling analysis of Indian rice variety

Authors: S. Harini; Saritha Chakrasali; G.N. Krishnamurthy

Addresses: B.M.S. College of Engineering, Bull Temple Rd, Basavanagudi, Bengaluru, Karnataka 560019, India ' BNM Institute of Technology, 12th Main Road, 27th Cross, Banashankari Stage II, Banashankari, Bengaluru, Karnataka 560070, India ' BNM Institute of Technology, 12th Main Road, 27th Cross, Banashankari Stage II, Banashankari, Bengaluru, Karnataka 560070, India

Abstract: Image processing and machine learning has a wide application in the field of agriculture and food industry. This is because of the non-destructive evaluation process, performance and low cost compared to manual methods. Analysing grain quality manually is laborious and also subjective. It depends on the knowledge and experience of the experts. Rice is one of the staple food grains in major countries of the world. India being one of the top most exporters of rice grains, the quality analysis is very crucial. The food industry and consumers suffer from the lack of a fast, automated solution for identifying the quality of grains. To address this problem, this work proposes a machine learning-based solution for automatic analysis of quality of rice grains using degree of milling (DOM). Various machine learning algorithms are used for the analysis. A noticeable result is obtained for SVM, KNN, decision tree and CNN algorithms with an accuracy of 96%, 90%, 88% and 100% accuracy, respectively.

Keywords: convolution neural network; CNN; SVM; decision tree; KNN; Indian rice; degree of milling; DOM.

DOI: 10.1504/IJAITG.2023.135052

International Journal of Agriculture Innovation, Technology and Globalisation, 2023 Vol.3 No.2, pp.177 - 192

Received: 24 Feb 2023
Accepted: 29 Mar 2023
Published online: 28 Nov 2023 *