ASABE Technical Library - Abstract

Abstract.

Guignardia citricarpa Kiely (Anamorph: Phyllosticta citricarpa (McAlpine) Van der Aa) or citrus black spot (CBS) is a citrus fungal disease found in 2010 in south Florida that causes premature fruit drop and yield loss particularly on late harvested varieties such as Valencia. CBS symptoms on citrus fruit occur in various forms; however, hard spot and cracked spot are the most common symptoms. In this study, spectral characteristic of several CBS symptoms were investigated to determine the best spectral bands for diagnosis. Color images of CBS symptoms as well as healthy fruits were acquired using a digital single-lens reflex camera. Band selection was conducted with respect to spectral response of the camera. A set of color features were extracted from red, green, and blue channels of the color images for classification. A classification model was developed using various classifiers and k-fold cross validation which was able to accurately detect the CBS symptoms along with the level of severity. The results showed that the spectral analysis could provide useful information to develop an efficient color image classification algorithm for CBS detection. CBS-positive spots were classified with a 100% accuracy using this algorithm. CBS lesion type and the maturity of healthy fruits were also identified with accuracies of 69% and 98%, respectively.