Design Of Lighting System And Flotation Video Image Processing

The flotation is a key step in the separation of industrial minerals extraction.Within the limited mineral resources,how to improve the utilization is an international problem.For this reason,the global researches and applications have highlighted how to increase resource utilization which is combined with machine vision and digital image processing technology.In this paper,it is focused on the key techniques of the lighting system and the bubble detection,such as uniform illumination,image enhancement and bubble extracting.The abovementioned techniques are used to overcome the difficulties of bubble detection in the condition of uneven illumination,noisy and complex pavement image.These techniques are combined with non-imaging optics,digital image processing methods and fractional calculus.The main research content and the contributions of this thesis can be summarized as:1.Design of uniform illumination distributionAs for the problem of the uneven illumination image in the machine vision system,we use non-imaging optics theory to process secondary light distribution and focus on the method that structure the free curved lens by solving equations.The uniformity of image intensity is increased and texture details are richer by physical modification.This will establish the theoretical basis for image enhancement by the algorithm.2.Image preprocessingAlthough the uniformity of image intensity is improved by physical modification,image enhancement algorithm is necessary for the reason of the existence of shadow and superposition of three-dimensional bubbles.A new enhancement algorithm based on adaptive fractional differential,which integrates the advantages of the algorithm,has been proposed in this paper.The fractional differential is beneficial in enhancing image texture detail and enhancing the edges and adaptive to choose the optimum differential order times through the information of the bubble boundary pixel texture able to meet the requirement of real-time monitoring of flotation.3.Image classification and discriminationBased on the analysis about the surface visual feature of flotation bubble to classify images,this will provide the foundation of multi-scale transform for accurate image segmentation.Compared with Otsu operator and Canny operator,The method which adopts gray symbiotic matrix to obtain texture features,and then analyze texture data by multiple linear regression to identify bubble image type performs better classification results.4.Bubbles extractionSince the boundary pixel gray of the bubble is a set of partial minimum to its nearby pixels.This paper is proposed a new algorithm based on the valley edge detection of fractional integral.The proposed algorithm not only has the good anti-noise performance but also obtains satisfactory effects in both improving the weak boundary of the bubble and accurately positioning bubble information in different pavement images.Besides,this paper improves the classical algorithms of burr noise elimination and gaps connection.The improved algorithm is proved to obtain good effects in removing burrs and can accurately connect bubble gaps according to the original bubble information.