煤浮选柱泡沫图象的识别

Flotation is a major coal preparation method. The research of flotation theory was mainly interior of froth，the visual information for the surface of froth did not be concerned before. Froth is the final product of flotation. The character of surface of froth can directly and rapidly indicate the flotation result. Operators usually adjust the flotation process according to the information watched by eyes, measurable evaluation could not be obtained. The adjustment done by operator is only according to man's feeling, in this way it is difficult to achieve the high level of automation. .Computer-based vision technology has been used to monitor flotation process, the characteristics of image of froth for column flotation of coal have been systematical evaluated in this study. .Major works of this project include: image of froth on-line monitor system for hardware and software; laboratory and industrial column flotation tests for different flotation conditions; physical character and textural feature extraction of image, pattern classification of image by means of neural network called Self-organizing Feature Mapping (SOFM) for coal flotation froth. .The major conclusions of the study are: there is no significant different of color for coal flotation froth; if the edge of bubble is clear, the information of individual bubble can be extracted, therefore the number of bubbles, the size distribution, average size and coefficient of shape of the bubbles can be calculated by means of image enhancement, binary method, edge detection and so on. Textural features of spatial gray level dependence matrix (SGLDM) and neighboring gray level dependence matrix (NGLDM) have been extracted within the study, and relationship of all these features with froth character has been analyzed. Images of laboratory and industrial froth of column flotation for coal have been classified by SOFM, the total correct rate is over 70% comparing with the observation of experts. The moving speed of bubbles on the surface of froth has been detected by laboratory on-line monitor system..This study has successfully introduced computer-based vision technology to monitor coal flotation and got evident achievement. This method can be used to control flotation performance, the advantage of this method is: low investment requirement, improve the quality of products, raise the recovery of products, increase the productivity and saving manpower. Therefore significant profit can be obtained. This method is the combination of high technology with traditional mineral processing industry. .Flotation is a major coal preparation method. The research of flotation theory was mainly interior of froth，the visual information for the surface of froth did not be concerned before. Froth is the final product of flotation. The character of surface of froth can directly and rapidly indicate the flotation result. Operators usually adjust the flotation process according to the information watched by eyes, measurable evaluation could not be obtained. The adjustment done by operator is only according to man's feeling, in this way it is difficult to achieve the high level of automation. .Computer-based vision technology has been used to monitor flotation process, the characteristics of image of froth for column flotation of coal have been systematical evaluated in this study. .Major works of this project include: image of froth on-line monitor system for hardware and software; laboratory and industrial column flotation tests for different flotation conditions; physical character and textural feature extraction of image, pattern classification of image by means of neural network called Self-organizing Feature Mapping (SOFM) for coal flotation froth. .The major conclusions of the study are: there is no significant different of color for coal flotation froth; if the edge of bubble is clear, the information of individual bubble can be extracted, therefore the number of bubbles, the size distribution, average size and coefficient of shape of the bubbles can be calcul

精矿泡沫是浮选的最终产品，泡沫表面性质能最直观、最迅速的反应浮选的分选效果。利用图象处理技术，研究浮选柱表面泡沫状态特征和差异与浮选柱操作变量、浮选柱结构的关系；与浮选柱分选效果的关系，研究基于浮选柱泡沫状态实时图象处理系统的浮选柱专家系统，加深对浮选机理的了解，提高浮选柱自动控制水平。.