Research On Ash Detection Method Of Flotation Tailings Based On Image Processing

Flotation tailings ash is an important indicator of flotation products.For the flotation tailings ash of the coal preparation plant,offline detection is used to achieve accurate online measurement,and the current flotation soft measurement mostly uses a single gray image,which leads to the problem of poor accuracy and adaptability of the soft measurement model.A soft measurement method for flotation tail coal based on color image processing method,and a soft measurement model for flotation tail coal ash based on adaptive particle swarm optimization least squares support vector machine(APSO-LSSVM).Therefore,three aspects of image feature extraction,feature data dimension reduction processing and tail coal ash soft measurement model construction of flotation tail coal slurry were studied.Firstly,this paper introduces the research background of this study.In the process of flotation control,there are many methods for ash detection of flotation products.By comparing the advantages and disadvantages of various methods,and analyzing the actual situation of the current production site,It is found that image recognition technology has great potential in the field of soft sensing of key parameters in flotation process.In this paper,a set of experimental devices for collecting images of flotation tail coal slurry is designed for the experimental objects,including color CCD industrial camera,fixed focus lens and 45° ring light source.According to the actual production situation,the coal sample allocation plan was formulated.The flotation clean coal dry powder,the flotation tail coal dry powder and the vermiculite dry powder were used as the basic coal samples,and the concentration was 10g/L-70 g by proportional mixing and water stirring./L interval 10 g / L,ash into 20%-70% interval 2% of 182 sets of sample pulp.Then,the main hardware and lighting methods of the experimental device were selected and designed.In order to maximize the difference of image characteristics of different sample pulps,this study analyzed the exposure time and aperture size of the camera based on the characteristic difference of 20% ash and 70% ash and found the most suitable combination of parameters for this experiment.By analyzing the original image of the flotation tail coal slurry,comparing the black and white image information and color image information of different ash,it is found that the tail coal slurry image contains a large amount of color information due to the presence of non-coal materials.Therefore,this paper chooses to collect tail coal.The color image is analyzed.On the other hand,by analyzing and judging the factors affecting image feature extraction in the original image,there are noise and specular reflection of coal particles.To eliminate their interference,Gaussian filtering,mean filtering and multi-image averaging are used in this paper.The color information of digital images has different standard definitions in different color spaces,so this paper briefly introduces RGB color space,YUV color space and HIS color space,respectively,and then extracts the mean value of the slurry image in the above three color spaces..In order to simplify the computational complexity of the soft-measurement model,the kernel principal component analysis method is used to reduce the original sample data from 10 dimensions to 5 dimensions,which provides the subsequent model construction.Convenience.Finally,this paper studies the basic idea and principle of support vector machine,and introduces the support vector regression machine which is suitable for this research.Because the parameters of support vector machine have a key influence on the accuracy and adaptability of the model,the particle swarm optimization algorithm is used to optimize the model parameters of support vector machine.Finally,an adaptive particle swarm optimization support vector regression model for the soft-sensing of coal ash in flotation tail is constructed.As a contrast analysis,the soft-sensing model of coal ash in flotation tail is constructed by using artificial neural network toolbox in MATLAB.The comparison results show that the APSO-LSSVM model has higher accuracy and applicability.