On-Line Detection System Of W Ion Concentration During Tungsten Smelting Process Based On Machine Vision Technology

During the process of w ion exchange smelting production,the concentration change of WO₃ in feed solution directly reflect the process of smelting production and its quality.While in the actual production,the detection of w ion concentration was relied mainly on the operation of the workers taking samples to laboratory for analysis.This method has problems of discontinuous,non-real time,resulting in phenomena like not timely production control,and inconsistency between production process and product quality.W ion concentration on-line detection method was studied using machine vision technology,and the characteristic and influencing factors of manual sampling and chemical examination analysis process was analyzed,according to the problems existed in manual sampling and chemical examination analysis process,taking tungsten smelting process as the background.Testing process that adapted to on-line analysis was optimized and restructured according to the requirements of on-line analysis.The w ion concentration on-line detection system consist of sampling and on-line detection was constructed,and the analysis of corresponding image feature and LSSVR optimized by PSO were studied.The main contents and conclusions are as follows:（1）The basic principles of w ion concentration on-line detection system are described based on machine vision,overall program and hardware design of w ion concentration on-line detection and analysis system that consist of sampling and on-line detection and analysis were established,and the influence of ambient light and feed solution temperature have on the accuracy of detection and analysis was analyzed.（2）Image processing function of the on-line detection and analysis system was realized by using HALCON.The environment and equipment noise was eliminated by using median filtering method;illumination compensation of the image was achieved by transforming the image into HSI space,performing liquid solution image decomposition by using wavelet transform,and low-frequency image enhancement based on multi-scale Retinex algorithm,according to the corresponding relationship between w ion concentration and feed solution color;S component segmentation of the image was performed by using OTSU algorithm,making it easy to extract and analyze the characteristic of the target area,thus acquiring the image parameters that characterize w ion concentration,such as the mean value of S,H,I.（3）W ion concentration prediction model was established by using LSSVR optimized by PSO to deal with interfering elements existed in feed solution concentration image information,to improve accuracy and real-time ability of detection and analysis.（4）W ion concentration on-line detection and analysis system was developed by using LabVIEW.The system is capable of real-time detection,excess capacity analysis,alarm and data sharing.The experiment showed that the w ion concentration on-line detection and analysis system based on machine vision technology has a good real-time ability,better accuracy,and the prediction model has a strong generalization ability.Its error is within 3%,and it is in line with the technological requirements.So,the system is able to replace the tedious manual sampling and laboratory analysis and achieve w ion concentration on-line detection and analysis.