Modeling Of Mineral Grade Based On Flotation Froth Image Color Feature Extraction And Engineering Realization

With the further development of the metallurgical industry and the exploitation of polymetallic complex ore.Although various beneficiation methods have a long history of production,due to the increasing complexity of ore,the current mineral processing technology has been unable to meet the selection requirements.In the flotation production process,the process is complex,the influencing factors are uncertain,and the nature of the ore will directly affect the production index of the flotation operation.Therefore,online pulp grade analysis and automatic addition of pharmaceuticals and production process testing are of great significance for guiding production,saving medicaments,and improving flotation indicators.Due to the irregular nature of ore and the distribution of resources,China’s mineral resources utilization rate is not high.Therefore,in the flotation production process,the grade analysis equipment capable of online inspection is urgently needed.Domestic and foreign mineral processing and automation workers have made in-depth parameters for the grade.However,due to the complexity of the flotation process mechanism and the uncertainty of the environment,there is currently no suitable online grade analysis equipment successfully applied in the industrial field.The widely accepted ore flotation foam image of the specific ore dressing can reflect the flotation state.In this paper,the following research is carried out on this issue:Firstly,it demonstrates the feasibility of flotation foam image processing in the concentrator.Combined with the flotation process flow and the principle of flotation solid material separation,XRF qualitative and semi-quantitative analysis of flotation feed and tailings,Analysis of the color of various substances and the feasibility of image processing technology.Then,digital image processing technology is used to extract the color characteristics of the foam.The image is preprocessed by flotation foam image,and the modified Tenengrad sharpness evaluation function is used to screen the blurred image.The method of total variation repair is adopted for the bright spot on the top of the foam.To achieve the repair of bright pixel;finally,the image of the restored foam image is extracted by multi-color space,and the feature of the extracted color component is simplified by analyzing the macroscopic feature of the flotation bubble image.Finally,the color features of the more suitable flotation bubble images were selected.Finally,using SVR as the modeling method,through the comparative analysis of kernel function and the optimization of model parameters based on GWO,a tailing grade prediction model suitable for the iron ore beneficiation plant was finally studied.The design and development software platform was successfully applied to the flotation industrial site,and docked with the flotation automatic dosing system to realize the intelligent optimization control of the flotation dosing process.At the same time,it has successfully put into production on the spot and brought considerable economic benefits to the company.