以山西省种植的英国红芸豆为研究对象,参考芸豆的行业标准和专业标准,采用计算机视觉的方法对霉变芸豆进行检测和识别,提出使用表面粒度和霉变区域的面积所占的比例的方法对芸豆的外观品质进行评价。试验表明在所选的样本条件下,正确率可达90%,为芸豆霉变品质的计算机视觉检测和评价提供了理论基础。 Taking the red kidney beans planted in Shanxi Province as the research object, with reference to the industrial standards and professional standards of kidney beans, the method of computer vision was used to detect and identify moldy kidney beans, and the proportion of surface area and area of ​​mildew Methods The appearance quality of kidney beans was evaluated. Experiments show that under the selected sample conditions, the correct rate can reach 90%, which provides a theoretical basis for the computer vision detection and evaluation of the quality of the mold of the kidney fungus.