A Scattered Liquid Component Analysis Approach Based on Spectral Visual Encoding and Fusion

In recent decades, spectral analysis has become a key research field to determine product components. Ion concentrations in metallurgical liquid are crucial component parameters for guiding the stable process operation in zinc hydrometallurgy. Its rapid and accurate analysis plays a critical role in industrial informatization. However, on the one hand, due to the complex physical and chemical properties of metallurgical liquid, the suspended solid particles in the liquid cause incident light scattering, which violates Lambert-Beer's Law, and results in a considerable determination deviation in the traditional methods; on the other hand, the optical-based particle modeling relies on precise optical parameters and complex calculation, which is a costly endeavor. To get around these issues, a scattered liquid component analysis method based on spectral visual encoding and fusion is proposed. It unifies the traditional three spectrum analysis steps and provides a new perspective on spectral variation characterization. First, a multi-spectrum acquisition system based on region scanning was developed to mitigate the effects of suspended solid particle sensitivity. Second, a spectral visualization framework based on image encoding is established, and the dynamics variation of spectral curve shape and absorbance amplitude are characterized. Third, a multiple-channel mechanism is designed, which enables the proposed method to extract and fuse feature information of different encoded images. The experiment results proved that the proposed method yields higher determination accuracy.