Mango Fruit's Maturity Status Specification Based on Machine Learning using Image Processing

Abstract

The notion of maturity is very crucial to obtain a good storage period of septic fruits and vegetables. It is possible to profess the maturity of fruit by various characteristics where color of the skin is the most standard measure for judging maturity. Typically, human's perception can be wrong about the maturity while the perception being made by visualizing the skin color. This research aims to develop a technique to detect and specify the status of mango into different stages. The collected RGB images are converted to HSV color space at the very first phase of the conducted research. By considering the “S” channel, the obtained image is segmented where thresholding technique is used. From the segmented image fifteen vital features are extracted. Three as well as six stage maturity classifications are performed based on these features with 94 and 88 percent of accuracy accordingly. The accuracy of result indicates that the proposed technique can be a helping hand to promote our mango fruit industry as well as our economy.