Citrus Yield Mapping System Using Machine Vision

Abstract

This research was conducted to develop an image processing algorithm to identify andcount the number of citrus fruits in an image. Once this algorithm is completed it will be incorporatedinto a machine vision system consisting of a GPS receiver and distance measuring devices in apick-up truck to estimate yield of a citrus grove on-the-go. A total of 90 images were acquired in anexperimental citrus grove. Images of the citrus grove were analyzed and histogram & pixeldistribution of various classes (citrus, leaf, and background) were developed. The threshold ofsegmentation of the images to recognize citrus fruits was estimated from the pixel distribution of hueand saturation color plane. A computer vision algorithm was developed to enhance and extractinformation from the images. Preprocessing steps for removing noise and identifying properly thenumber of citrus fruits were carried out using a combination of erosion and dilation. Finally thenumber of fruits was counted using blob analysis. The total time for processing an image was283 ms excluding image acquisition time. The algorithm was tested on 59 validation images and theR2 value between the number of fruits counted by the machine vision algorithm and the averagenumber of fruits by manual counting was 0.76.