Soft Sensing Of Pr/Nd Element Component Content Under Different Illumination Conditions

Rare earth elements are indispensable materials in the production process of various industrial products.Because of this reason,many countries attach great importance to the industrial production of rare earth elements.In China,cascade extraction is mainly used in rare earth industrial production.This method is not easy to control because of its complex process,many control variables,strong coupling and large time delay.Rapid and accurate detection of rare earth element content is the key to flow control of extractant and detergent in rare earth extraction process.In this paper,use the color constancy to solve the problem of get the content of Pr/Nd rare earth under different illumination conditions.The weighted least squares support vector machine(WLSSVM)is used to establish a prediction model for the rapid determination of components in rare earth mixed solution.The main research contents are as follows:1.The characteristics of color description in different color spaces are analyzed.The image of Pr/Nd mixed solution was taken as the research object.The first moment values of H,S and I in HSI color space are selected as image features to distinguish different solution images.2.The theory of color constancy is introduced.In Grey Edge algorithm,region partition parameters are added,and genetic algorithm(GA)is used to optimize the parameters of the algorithm.The images under different illumination conditions are corrected to standard illumination.Through the simulation experiment,the better correction effect is obtained.3.The least squares support vector machine(LSSVM)algorithm is introduced.According to the different influence of each color information on component content,different weights are given to each color information,and weighted LSVM component content prediction model is established.Finally,the comparison and validation of the modeling method show that the model has high prediction accuracy and good generalization ability,and is suitable for the accurate and fast prediction of component content in rare earth extraction process.In this paper,color constancy is added to detect the content of rare earth components.The content of rare earth components is obtained after correcting the images of rare earth solutions under different illumination conditions.It plays an important role in the detection of component content in industrial production.