Geometry-based mass grading of mango fruits using image processing

Abstract

Mango (Mangifera indica) is an important, and popular fruit in Bangladesh. However, the post-harvest processing of it is still mostly performed manually, a situation far from satisfactory, in terms of accuracy and throughput. To automate the grading of mangos (geometry and shape), we developed an image acquisition and processing system to extract projected area, perimeter, and roundness features. In this system, images were acquired using a XGA format color camera of 8-bit gray levels using fluorescent lighting. An image processing algorithm based on region based global thresholding color binarization, combined with median filter and morphological analysis was developed to classify mangos into one of three mass grades such as large, medium, and small. This system achieved an accuracy of 97% for projected area and Feret diameter, 79% for perimeter, and 36% for roundness. To achieve a finer grading, two different grading features could be used in sequence. The image grading system is simple and efficient and can be considered a suitable first stage to mechanizing the commercial grading of mangos in Bangladesh. Moreover, the method has the potential to be applied to other crops with suitable adjustments.