论文部分内容阅读
根据精细化工生产过程中工艺介质密度分布、液位/界面和酸碱浓度等参数的实际检测要求,分别对X射线数字成像、机器视觉和近红外光谱等非接触式检测技术进行了研究,构建了相应的检测系统,研究了相关图像处理和光谱校正建模技术,取得了良好的检测实验结果。基于低能X射线数字成像和图像处理技术,进行了粉体工艺介质相对密度分布的数字成像及图像重建,通过多项式拟合进行硬化伪影校正,实现了粉体工艺介质相对密度分布的准确估算。选用智能相机,设计了1套液位/界面机器视觉检测系统,利用图像预处理、Otsu阈值化和Harris角点检测识别液位,结合图像标定获取待测液位/界面。通过采集大量样品,用Antaris II傅立叶近红外光谱仪测试,并经光谱数据处理,建立了HCL水溶液摩尔浓度、Na OH水溶液质量百分数和摩尔浓度的近红外光谱校正模型,已用于某精细化学品的生产现场测试,提高了酸碱原料的检测鉴定效率。
According to the actual testing requirements of process media density distribution, liquid level / interface and acid-base concentration in fine chemical production process, the non-contact detection technologies such as X-ray digital imaging, machine vision and near infrared spectroscopy were studied respectively, The corresponding detection system was studied, the relevant image processing and spectral correction modeling techniques were studied, and good experimental results were obtained. Based on the low-energy X-ray digital imaging and image processing technology, the digital imaging and image reconstruction of the relative density distribution of the powder process medium are carried out. The hardening artifact correction is performed by polynomial fitting, and the relative density distribution of the powder process medium is accurately estimated. Using a smart camera, a set of liquid level / interface machine vision inspection system was designed. The liquid level was identified by image preprocessing, Otsu thresholding and Harris corner detection. The liquid level / interface was obtained by image calibration. A large number of samples were collected and tested by Antaris II Fourier Transform Infrared Spectroscopy (FT-NIR). After spectral data processing, a calibration model of molar concentration of HCL, mass percent of NaOH aqueous solution and molar concentration near-infrared spectroscopy was established. Production site testing to improve the detection of acid-base material identification efficiency.