Computer vision technology has been widely used in image processing. It provides aresearch space for the automation classification of fruits. In this paper, the application ofcomputer vision methods was cited for establishing an automatic classification of qualitygrade based on the color and gradient features of the fruit surface in HSV color space.The research contents including the image acquisition, image preprocessing,feature extraction, and classification recognition. For extracting the fruit surfacefeature effectually, the image color space of RGB space was converted to HSV spacefirst. The H mean value was extracted as the color features of the fruit surface in HSV space.And H space mean gradient was also extracted as the depict features of the fruit surface byusing Laplace in HSV color space. Finally, a classifier was established using SOM neuralnetwork based on the color features and depicts features. The experiment experimental resultswere reported. |