Classfication Of Mango Based On Computer Vision

In our country, detections of fruits are mainly depended on the human's sense organs, which lead to lowness of efficiency and bigness of error. The appearance quality of export fruits looks ugly. As a result, the fruits are short of competition in the international market. Comparing to human sense organs, computer vision technology is higher speed, larger information and more precision quality.The system of mango's classification based on computer vision is composed of acquiring information, image pretreatment, image segmentation, characteristics picked-up and designing classification.Acquiring information is to import in the standard image, pre-treatment is to boost up the image. The color system of HIS, which is accorded with human's apperception of color, is adopted. Then median filter is used to filter the image of I. the threshold which is used for image segmentation is computed with maximal square difference.When it comes to feature selection, several inferior to optimal solution methods and the branch and bound (BAB) algorithm are used to search the features for classification. The experiment shows that the exclusive method is BAB algorithm that is sure to acquire the optimal features. In order to increase searching efficiency, the improved branch and bound (IBAB) algorithm is mentioned. The experiment shows that IBAB is much faster than BAB and more and more effective with the increasing of the features.Three methods are discussed in this paper to devise the classification:...