NIR Computer Vision for Rapid Measurements of Some Physical Properties of Mangoes

Abstract

Quality assessment through quick sorting and grading of mangoes is today’s need, for not only to enhance the post harvest management process but also to minimize the post harvest losses and to recover more economic and nutritional values. Near infrared (NIR) based computer vision (CV) which is very close to human vision was used for rapid measurements of some physical properties of mangoes (Chausa, Dashehari). Lab View s/w was used to process the image and development of algorithm steps for analysis of the image properties. Fruit’s diameters were measured and compared with manual measurements using paired t-test, the 95% limits of agreement (Bland-Altman plot) and both measurements correlated using regression analysis. The results obtained were consistent and the correlations of the mean dimensions between both methods were found to be noticeably high. Coefficient of determination (R2) for all diameters was found between 0.951 and 0.90. Various fruit’s shape attributes important in design of harvest and post harvest equipment were also evaluated using NIR-CV technique. The present research, thus, could be the basis for the development of real time CV systems which would be near to human vision.