Research And Application Of Characteristics Extraction Methods Based On Froth Flotation Image

Froth flotation is the most common method for extracting useful minerals from ores.The surface characteristics of flotation foam is an important indicator of flotation performance,and there are many problems in the way of manual observation.Nowadays,machine vision monitoring system has been widely used,and it is the trend of industrial intelligence and innovation.Accurately and rapidly extracting froth characteristics is the key of intelligent monitoring system.In this paper,we analyzed the features of industrial environment and foam image,the main work as the following aspects:(1)An adaptive MSR image enhancement method proposed to reduce the influence of the fog and the uneven illumination on the image.(2)This paper improved watershed segmentation method based on multi mark.First,We used K-means clustering automatic extraction of high gray area,and calculated the gray difference between the maximum and minimum area,thus,we got the optimized internal markers.Second,we used morphological open close transform to improve the foam image.Finally,the watershed segmentation based marker method complete the precise segmentation of different processes.(3)ORB method designed to speed up the velocity of extracting local feature points,and the methods of matching are improved.In addition,and improved feature point matching method,and according to the similarity between the froth the proposed block and reduce the search matching method based on window,to achieve the precise detection of bubble velocity field.In order to solve the problem that the similarity between different foam is very large,the method of dividing the block and reducing the search window is aim to realize the accurate detection of the foam velocity field.(4)This paper design is a software based on machine vision,which is used in flotation plant to help the reduce operator workload.Our system is realization of video capture,image processing and characteristics extracting(e.g.number of bubbles,size distribution,velocity and stability),results display and alarm,data storage and so on.Thus,the machine vision system improve the level of industrial automation.