Thai Tangerine Size Classification via Computer Vision

In the citrus industry in Thailand, fruit sizing is generally performed by machines, which entails mechanical damages to the fruits. Computer vision technology can serve as a non-destructive alternative. Our focus is on the size classification of Thai tangerines, regarded as major economic fruits, by means of this technology. Specifically, we apply five classification methods, namely area thresholding, diameter thresholding, K-nearest neighbors, support vector machine, and artificial neural network, to a dataset of Thai tangerine images created by our computer vision system. We compare these methods in terms of their classification accuracy and visualize the classification performance using cobweb representation. Our results provide insights into the potential of computer vision technology for fruit size classification.