Quality Grading Classification of Dry Chilies Using Neural Network

Abstract

Due to lack of precise information for quality of crops produce, the overall growth of agriculture sector in India is limited. In the present work, the deep convolutional neural networks (CNNs) were used for image classification of dry chilies with distinguishable quality of farm produce. Our unique approach facilitates traders to visually evaluate the quality of chilies from the high-quality images received from producers in farms. The quality of chilies was determined from the color. The database was prepared following image capture, data collection and image analysis. The results showed that our model achieved precision between 65 and 75%, for separate class tests, on average 66%. The global objective of this project was to develop an innovative tool to investigate the quality of chilies from their color using image analysis to improve its applicability for commercial use in edible products.