Automatic Counting of Chili Ripeness on Computer Vision for Industri 4.0

Abstract

This study aims to determine and counting the number of ripe chili on a tree. The proposed system uses images taken in gardens with a distance of 40 cm and an image resolution of $\mathbf{576x864}$. This work implements RGB2HSV color transformation, image segmentation, threshold, and image morphology, and then implements a Blob Analysis method for detecting and counting the ripe chilies on the trees. The trial results of this study provide a boundary for the blob area that will be used. Blob area less than 400 will not be labeled for the bounding box of the object detected. The accuracy of the proposed method to detect and calculate the ripe chilies on tress is 89.7%.