Yield Estimation of Citrus Fruit Using Color Based Segmentation and Circular Hough Transformation

Abstract

Yield estimation using image processing techniques is an emerging domain of research. The yield estimation for citrus fruit is very important before harvesting fruit. The purpose of this study is to estimate the yield of citrus fruit using image processing techniques. Different researchers have developed different methods for yield estimation by using Image processing, Machine learning, and deep learning to estimate the yield of citrus fruits on time and to help the farmer to make timely decisions. A lot of work has already been done, but the accuracy in segmentation of fruits, occlusion, and different lighting conditions are still challenging problems to be addressed. The proposed technique eliminates these problems and improves overall detection accuracy. The proposed technique is based on two steps (i) color base segmentation, to segment the region of interest (Oranges), and (ii) circular Hough transformation for detection and counting of oranges. The overall accuracy of the system is 87% and the coefficient of determination R2 is 0.99 which shows the efficiency of the proposed approach