【摘要】 该试验基于模糊聚类分析方法结合机器视觉技术,开展结球甘蓝等级检测方法研究。国标结球甘蓝等级规格(NY/T 1586—2008)中,根据叶球大小等9个等级评价指标,将同批次的结球甘蓝分为二级、一级、特级。提取结球甘蓝图像的形状特征、颜色特征和纹理特征,定义了平滑度等6个图像特征,共计21个图像特征参数,替代表示等级评价指标。通过模糊聚类方法将聚类集样本分为3个类别,分别统计3个类别特征参数的均值和标准偏差,按照各个类别拥有均值和标准偏差优异值的相对数量,确定3个类别的对应等级。通过计算测试集样本与各个类别样本中心的欧氏距离,以最小距离确定待测样品的等级归属。试验共108个样本,按照国标方法进行等级检测,二级、一级、特级的样本个数比例为1:2:2。其中80个组成聚类集,28个组成测试集,聚类集和测试集中各个等级的比例接近1:2:2。模糊聚类后,按照各个类别拥有均值和标准偏差优异值的数量,第1类样本是二级,第2类样本是一级,第3类样本是特级。聚类集和测试集中各个级样本占比与国标方法检测的3个等级的数量比例1:2:2一致。试验结果表明,该试验方法可用于结球甘蓝的等级检测。更多还原

【Abstract】 Based on fuzzy clustering analysis combined with machine vision technology,the detection method of cabbage grade was studied.According to the national standard grades and specifications of cabbage(NY/T 1586-2008),the same batch of cabbage was divided into second grades,first grade and super grade according to the nine grade evaluation indexes of leaf bulb size etc.The shape features,color features and texture features of cabbage image were extracted.In addition,smoothness etc.were defined,21 image feature parameters represent the grade evaluation index.The cluster samples were divided into three categories by fuzzy clustering method.The mean and standard deviation of the characteristic parameters of the three categories were counted respectively.The corresponding grades of the three categories were determined according to the relative number of excellent values of the mean and standard deviation of each category.By calculating the Euclidean distance between the test set samples and the centers of each class of samples,the classification of the samples to be tested was determined by the minimum distance.108 cabbage samples were tested according to the national standard method.The proportion of the number of second grade,first grade and super grade samples was 1:2:2.Among them,80 cluster sets and 28 test sets were composed.The ratio of each level of cluster set and test set was close to 1:2:2.After fuzzy clustering,according to the relative number of excellent values of the mean and standard deviation of each category.The cluster set samples were divided into three categories.The first category was the second level,the second category was the first level,and the third category was the super class.The proportion of samples of each level in the cluster set and the test set was the same,which was consistent with the number ratio of three levels detected by the national standard method of 1:2:2.The results showed that the test method could be used to detect the grade of cabbage.更多还原