Based On Vector Morphology Reconstruction Of Copper Flotation Foam Image Segmentation Method Research And Application

Flotation is a mineral separation technology by use of the difference of the chemical and physical properties of the mineral surface. At present, the domestic copper ore flotation process is mainly operated by experienced workers through observing the visual appearance of the froth phase. As such, it is difficult to work in optimized running state. Therefore, the research of flotation froth image monitoring system is great significance for optimal control of flotation process.Froth morphological characteristics is the important parameters of froth physical characteristics, and the segmentation of the froth images is an essential part of morphological feature extraction. Based on the characteristics of copper flotation froth images, research an image segmentation method of copper flotation froth. Main research work are as follows:1. Analyzing the characteristics of the copper flotation froth images, and considering the limitation that manual selecting structural element in image segmentation, the froth images is classified based on the average size of the white spot in the froth images, which is extracted by Otsu algorithm, to provide prior knowledge for image segmentation.2. The uneven illumination and the mass white spot has great effects on image segmentation. An improved algorithm of vetorial morphological opening reconstruction is introduced for brightness elimination from the froth images. The structural elements is selected on the basis of image classification. The method not only simplifies the calculation, but retains colour and outline information in the image.3. Through analysis on the froth image segmentation method based on valley-edge detection, it is pointed out that the size of template is difficult to determine, and the method can not accurately split froth image, in which there are too big or too small bubbles. As the result of the high noise of the copper flotation froth image, the over-segmentation phenomenon occurs easily. The tag image is extracted by use of the Top-hat based on h done area reconstruction algorithm, as the seeds of the watershed segmentation algorithm, the morphological opening closing reconstruction algorithm and the smoothing filter is adopted to filter out the noise in the froth image. On the basis, the new segmentation of the copper flotation froth image based on vetorial morphological reconstruction is proposed, it overcome the over-segmentation problem of the copper flotation froth image segmentation. After a large number of experiments, the method can be effectively used to split the images.The segmentation method in this thesis has been applied in a machine vision based flotation froth monitoring system. The method is able to split different type of froth image availably, and extract the morphological characteristics of the froth bubbles, it can direct the practical production.