The Research Of Apple Classification Based On Image Recognition

Presently, there are many examples that using image processing in detection, such as detect to product appearance, detect to infrared image in jointing. It is a process of image recognition. Firstly we receive the image signal and process the image, pick-up the diagnostic parameter in the result. Then recognize based on the diagnostic parameter, classify the product to excellent or inferior. This paper tells of the appliance of Image recognition in auto-detection, especially in apple image(designed the system of apple image recognition in size and shape).This paper narrates the lower layer image processing of the apple image to classify the apple based on the image. Detailed introducing image enhancement, edge detection, thresholding, thinning etc. And many arithmetic such as median filter, Sobel arithmetic operator, betterment gray scale histogram etc. Applied or betterment applied it to the graduating project. The thresholding of betterment gray scale histogram is a method that improve on the traditional thresholding of gray scale histogram, which result is very well.It is also detailed explained the method or step in image recognition during the image recognition system design. The apple feature parameters include size and shape. Picking up the average radius as the size feature. Picking up the average radius error as the shape feature. Founding the feature function that classify the apple according on size and shape. Present the datasheet of the result of application, which image recognition system classifies the apple. If not establish the feature function, we are also recognition through NN recognition. The paper probes into the appliance of NN in the apple system of recognition.