Froth Flotation Characteristics Based On Analysis And Applicationof Image Processing

Flotation is a very important part in the production of metal. Under certainconditions，put a certain amount of flotation agent in the pulp and filling the air，eventually produce large bubbles by stirring to improve the ore grade。Most frothflotation have features，such as sizes with dimensions, strong bubble adhesion, fielduneven illumination, surface texture and complex characteristics. Up to now，Most ofthe flotation process is still mainly rely on manual observation status froth，the processof flotation make too many uncertainties arising which cause decreased recovery.Flotation process involves many variables whose process is extremely complex andthere is a delay in the detection process. Therefore, it is necessary to study thecharacteristics of froth to achieve precise control of the flotation process, improve theutilization of mineral resources.In the article, analysis to extract image texture, research image segmentationalgorithm applicable to features copper-molybdenum mine. Conduct feature analysisthrough the copper-molybdenum mine image. The reaction can be effectively extractedcopper and molybdenum ore flotation process image related parameters to use theseparameters to map the flotation process image. In this thesis work is as follows:(1) Summarizes the characteristics of copper flotation froth image molybdenummine, then extract high highlight regional of froth flotation by Otsu Thresholdingmethod, mapping out the relationship between highlights and bubble size, which Laythe foundation for the subsequent structural elements selected.(2) Bubbles for flotation froth image and strong adhesion, uneven illumination,highlights the presence of high top features, propose an improved algorithm to removea highlight vector morphological. The method of adaptive selection of structural elements corresponding to effectively preserve the color and contour of the imageinformation, combined with the mean filter to remove image noise of imagepre-processing.(3) Summed up the bottom edge detection algorithm for image segmentation, afterimproved, its template should select on uneven bubbles can’t be accurately divided.Details bubble surface and some small noise points will form the minimum area,resulting in over-segmentation phenomenon. For this, top-hat based on the use of thearea h top reconstructed image as seed points extracted. On the basis, We propose areconstruction based on morphological watershed segmentation algorithm, and verifythe effectiveness of the algorithm through experiments.(4) Briefly discussed foam surface texture, the size of the froth, stability andbearing capacity of the four main features. Verified by various experiments todistinguish the texture feature image capabilities.