Image Characteristics Of Foam Velocity And Collapse Rate In Flotation Of Baiyun Ebo Rare Earth Mine

The dynamic characteristics of flotation foam include velocity,collapse rate and so on.This paper studied the flotation process of bubble collapse and the speed of the rate and grade,the relationship between the baiyun obo mine as raw material,and carries on the flotation foam of each time period after taking pictures and make a series of image processing and extraction bubble rate and collapse rate and analyze the correlation between the two and grade.The optimal dosage of flotation agent and flotation temperature were found out through flotation condition experiment before the photo collection.The optimum flotation conditions can be obtained:the flotation temperature is between 50?and 55?,the pH of the pulp is 9,the dosage of inhibitor is 6mL,the dosage of collector is 20mL,and the dosage of foaming agent is 3 drops.At the same time,in order to avoid the chance of the experiment,5 groups of conditional experiments were carried out again under the optimal reagent condition.Bubble scraping and grade measurement were carried out for each group of flotation experiments.After recording the data,it was concluded that the rare earth grade reached its maximum at 1min,then gradually decreased,and reached its minimum at the last minute.In the whole flotation process,the flotation was in the best state at 1min.In this paper,the computer image processing technology is used to capture the dynamic feature information of foam,analyze the foam speed and collapse rate,and carry out quantitative diversification processing.SIFT feature matching method is used to extract dynamic features.First of all,flotation photos were collected under the optimal conditions.The total collection time was 4min,and the photos were taken every 59s,with the interval of 0.5s.After that,image features are processed for the photos taken.In order to prevent off-site noise and possible distortion in the process of photographing,a series of pretreatments such as grayscale,denoising and edge detection are needed.After that,the characteristic value of the velocity was extracted.Among the 50 groups of data collected,the grade value corresponding to the velocity of 1.9mm/s and the collapse rate of 0.33%was the highest,reaching 60.45%,and the grade value corresponding to the velocity of4.1mm/s and the collapse rate of 1.3%was the lowest,only 48.6%.Finally based on the BP neural network for optimization of the rare earth minerals flotation process and the dynamic characteristics of the matching model and forecast results are concluded:the return of the bubble velocity and grade value R~2=0.9678,the return of the collapse rate and grade value R~2=0.95882,the R~2 values are close to 1,the prediction error is less than 0.06 at the same time,the speed of the flotation foam and collapse rate and grade the correlation between the strong and the prediction error values in a small range,using the BP neural network to forecast has good prediction effect.In this paper,a new method based on image dynamic feature extraction is proposed to detect the flotation grade,which has the characteristics of intuitive results,convenient data analysis,objective and accurate,providing a new idea for the software measurement and modeling of the flotation grade of rare earth minerals.