基于机器视觉的苹果果形分级方法研究

Research on Apple Shape Classification Based on Machine Vision

为了提升对苹果果形分级的准确率,文中采用了改进的圆度值分析法对苹果的果形进行分级.搭建硬件平台,获取苹果样本的任意9张图像,对图像进行二值化和中值滤波预处理,计算出预处理后的二值图像的面积和周长,求出圆度值,再求出圆度值的平均值和方差,与设定的阈值进行比较,最终得出分级结果.经实验证明,改进的圆度值分析法简单易行,分级正确率高,达97.5%,为苹果果形分级提供了参考依据.

In order to improve the accuracy of apple shape classification,the improved roundness value analysis method is used to classify apple shape.Build a hardware platform,get any nine images of apple samples,preprocess the image with binary and median filtering,calculate the area and perimeter of the binary image after preprocessing,and find out the roundness value.After calculating the average and variance of the roundness value,compare with the set threshold value,and finally get the grading result.The experiment shows that the improved roundness value analysis method is simple and easy,and the correct rate of classification is high,up to 97.5%,which provides a reference for apple shape classification.