Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/15/2003
Publication Date: 5/15/2004
Citation: Bennedsen, B.S., Peterson, D.L. 2004. Identification of apple stem and calyx using unsupervised feature extraction. Transactions of the ASAE. Vol 47(3): 889-894
Interpretive Summary: Automatic apple sorters need to distinguish between the stem/calyx of apples, and defects. A new, self-adapting image processing method was developed to identify the stem/calyx end of apples in an existing, research sorting system. The performance of the image processing routine corresponded to 99.95% correct classification under practical implementation. The self-adapting image processing method has potential for making automated apple sorting feasible, given further work, in order to obtain a better understanding of the method.
Technical Abstract: Neural networks and unsupervised feature extraction were used to classify apple images based on whether or not they included a stem or calyx end. In one experiment, the system successfully classified 98.4% of a test set consisting of 254 near infrared-images, captured at 740 nm. The network was also tested on grey level images captured with light in the visible range. In this test of 242 images, 5% were not classified correctly. In another classification test that included apple images with prominent defects 28% were misclassified. However, the majority of the errors occurred because major defects were mistaken for the stem or calyx. In a practical implementation, errors that could lead to loss in a sorting system would amount to only 0.05%.