基于图像处理的红枣分级检测方法设计

Design of Jujube-grading Detection Method Based on Image Processing

为提高红枣分级速度和分级检测的精度,实现无损分级的效率最大化,按照分级要求对红枣的大小、形状、颜色、表面褶皱程度及外轮廓等进行检测。采用图像预处理算法对红枣图像进行平滑去噪、图像增强,多种检测算法比较后采用自适应Canny算法对红枣图像进行外轮廓边缘检测,拟采用傅里叶算法对红枣的尺寸进行计算,由极半径函数确定红枣的形心位置坐标,配合Euclidean算法检测红枣距离尺寸,对干枣、鲜枣以及大小形状不规则的红枣混装进行分级处理。对若干大小、形状等特征不同的红枣进行分级检测试验,结果表明在提高红枣检测速度的同时准确率均在90%以上,具有较好的鲁棒性和准确率,在红枣分级要求精度较高的情况下,减少了整体运行时间,满足红枣分级的实际要求。

In order to improve the classification speed and accuracy of jujube,and maximize the efficiency of non-destructive classification,the size,shape,color,surface wrinkle and outline of jujube were detected according to the classification requirements.The image pre-processing algorithm was used to smooth and denoise the jujube image,and enhance the image.After comparing various detection algorithms,the adaptive Canny algorithm was used to detect the outline edge of the jujube image.Fourier algorithm was used to calculate the size of the jujube.The centroid coordinate of the jujube was determined by the polar radius function,and Euclidean was used to coordinate the centroid position of the jujube.The algorithm detected the distance size of jujube,and classified dry jujube,fresh jujube and jujube with irregular size and shape.The grading test of jujube with different sizes and shapes was carried out.The results showed that the detection speed and accuracy of jujube could be increased by more than 90%.It had better robustness and accuracy,and the overall running time could be reduced to meet the practical requirements of jujube grading under the condition of high grading accuracy.