一种基于特征分类的液体内杂质检测方法

在酒类产品的杂质检测过程中,不可避免地会产生气泡,而当前的杂质检测算法并不能有效消除气泡对检测的影响,尤其是在大量气泡存在的情况下。针对此问题,提出了一种基于特征分类的液体内杂质检测方法,通过提取目标的细微特征来区分杂质和气泡,算法通过双边滤波来预处理图像,改进了多尺度小波变换边缘检测算法,并用其来检测目标边缘,最后通过特征分类的方法来判定杂质。实验结果表明,该方法能有效消除噪声和气泡对检测的干扰,杂质检测的准确率达到了95%。

A Method for Detecting Liquid Impurities Based on Feature Classification

In the process of impurity detection of alcoholic products, bubbles will occur inevitably, and the existing detection algorithms cannot effectively eliminate the influence of bubbles on detection, especially when a large number of bubbles exist. To solve this problem, a method based on feature classification is proposed for impurity detection in liquid, which can distinguish impurities and bubbles by extracting subtle features of the target. The algorithm preprocesses the image by bilateral filtering, improves the multi-scale wavelet transform edge detection algorithm, and uses it to detect the edge of the target. Finally, the impurity is determined by the method of feature classification. Experimental results show that the proposed method can effectively eliminate the interference of noise and bubbles, and the accuracy of impurity detection reaches 95%.