Abstract: It was difficult to on-line detect the sizes of fine particles that were prepared from industrial production processes. An on-line detection method was developed based on light scattering extinction to measure particle sizes. The law of image pixel value data of different sample particle concentrations under on-line detection was analyzed and discussed using image analysis and data feature analysis methods. Under the conditions of the in-situ experiments of the ball mill, the scattered light spot images were collected, which determined sample particle mass concentration within a certain particle size range. The difference of brightness of the light spot images was analyzed, and the distribution law of the minimum value of the pixel value was statistically obtained. The standard control parameter models were established for different particle size ranges and different sample concentrations. The models were verified by the sample particle liquid with known parameters, and the similarity matching degree is more than 0.95. The models were applied to in-situ measure fine particles sizes which were prepared by ball mill showing that the method may be used for the primary detection of in-situ particle sizes.

Key words: in-situ detection, extinction method, speckle imaging, image feature analysis

摘要： 针对工业化制备微细颗粒原位在线检测的难题，基于光散射消光法测量颗粒粒径，利用图像分析和数据特征分析方法分析讨论了原位效果下不同样本颗粒浓度的图像像素值数据规律。在球磨机原位实验条件下，对满足一定粒径范围下确定样本颗粒质量浓度的散射光斑图像进行采集，通过对比光斑图像明亮程度的差异性统计分析像素值极小值的分布规律，建立不同粒径范围和不同样本浓度的标准对照参数模型。利用确定参数的样品颗粒液对模型进行验证，分布规律区间相似匹配度在0.95以上。对球磨机制备微细颗粒过程中的颗粒进行模型对比判断，结果表明该方法适用于原位粒径初级检测。

关键词: 原位检测, 消光法, 斑图成像, 图像特征分析