Research Of New Application Of Image Processing Technology In The Mine Corporation

With the development of economy, domestic iron and steel enterprises are facing more intense competition due to the expanding of production scale and the entrance of foreign iron and steel group; therefore, the higher grade of ore is required by civil corporations. However the grade of ore is too low that it doesn't satisfy the requirement of production. The floatation technology is an effective method for enhancing the grade of ore at present. In floatation process, the achievement of recognition of status is very difficult on account of its complexity. Presently, the most of flotation corporations use two following manners. Firstly, many dear requirements are used for realizing on-line analysis in plants. But the cost of fixing, debugging and maintenance will be greatly going up. Secondly, the traditional flotation control depends on operators who inspect the state of froth surface by their experiences. However, the control precision of the method isn't guaranteed. Therefore, the foregoing manners are not fit for the requirement of development of flotation corporations.In the thesis, the image processing technology is introduced into the state recognition of flotation process for extracting the feature from froth image. And then the image recognition model is set up with parameters of froth image characteristics on the basis of rough set and artificial neural network theory.The main contents in the paper are as follows:(1)The paper first discusses the manufacturing technique and the requirement offlotation process. The shortages of actual state recognition way are analyzed and thefeasibility using image processing is debated. (2)The paper discusses the characteristics of forth image in the course of anti-flotation. According to its characteristics and the demanding of production control, the color and texture features are regarded as input parameters including saturation, intensity, energy, entropy and moment of inertia. Above these parameters, the foundation for realizing the mathematical model of flotation process control will be provided.(3)The dosage of flotation reagent is a chief aspect in flotation process. This paper combines rough set with learning vector quantization neural network, and utilizes their complementarities to set up the state recognition model based on rough set and LVQ net, which will be used for the control of dosage of flotation reagent.