Persimmon's surface defect recognition based on machine vision fuzzy clustering

'Jiro' sweet persimmon surface defects were studied in this research. The color parameters and the texture parameters of the images were analyzed by the method of fuzzy pattern recognition, and the fuzzy vision set of the sweet persimmon's surface defect recognition was proposed, on the ground of that we designed and computed the vision membership degree function value of the different defects on the fruit surface. The closeness examination indicated that the division of different surface defects' membership degree function value and membership degree interval can fit the requirements of fussy cluster. The confirmatory experiment also showed that the fuzzy machine vision membership degree function and the relevant parameters interval as the tool of persimmon's surface defect recognition have a high accuracy in the aspect of recognition rate.