Quality Analysis of Rice Grains Using Ann and SVM

Abstract

Rice is the most favorable and most consuming food for all the human being in all over the world. Market for rice depends on the quality of it. Currently the type and quality of rice are assessed by visual inspection method through naked eye. This process is however tedious, time consuming, needs human expertise and depends on physical fitness of the inspector. To overcome these drawbacks, in this paper, an automated system is introduced which identifies and classifies the rice grains based on digital image processing techniques. Image processing method is most suitable as it is a non-contact technique, where in the image of the rice grains are captured. The captured images are pre-processed, segmented and features are extracted through MATLAB. From the extracted features the quality of rice is assessed based on Neural Networks (NN) and Support Vector Machine (SVM) classifier algorithms. A comparative study is made between these two methods and the results infer that SVM based classification outskirts its counterpart.

Thai Hom Mali rice grading using machine learning and deep learning approaches

“…There were a number of researches about rice classification using image processing [11], [12]. Many techniques were used such as k-nearest neighbors (KNN) classifier [13], [14], support vector machines (SVMs) [15], [16], neural network (NN) [17]- [21] and convolutional neural network (CNN) [22]- [24]. The most of the research based on grain separation features such as morphology, shape, color, and texture.…”