基于计算机视觉的农产品形状分级研究

Study on the Shape Classification of Farm Produce Based on Computer Vision

计算机视觉在农产品分级和品质检测领域中的应用越来越广。但由于水果等农副产品的形状不规则,如何准确地用较少的特征值定量描述农产品形状,是目前研究的重要课题之一。该文在分析了现有农产品形状描述方法的基础上,提出了采用傅立叶半径描述子对农产品图像轮廓进行描述和分类的方法。并给出了利用半径描述子计算图像区域面积以及利用欧几里德距离判定边界相似度的方法。

Computer vision technology is more and more applied to the field of farm produce classification and their quality examination.However,because of the irregularity of some agricultural produces,such as fruits,it is more difficult to describe the shape of agricultural produce quantitatively with lesser characteristics.Based on the analysis of current methods of produce shape description,a new method using Fourier radius descriptor to describe and classify the image outline of farm produce is proposed in this paper.A new method using Fourier radius descriptor to compute the area of image and using Euclidean distance to determine the similarity of boundary is educed too.