Design And Implementation Of Image Acquisition System For Coal Samples In Coal Preparation Plant

In recent years, with the development of computer technology and imageprocessing, the machine vision technology is being widely used in industry,agriculture, medicine, military and other fields, as a high-precision, highly intelligentadvanced technology, it has drawn increasing attention. While, the machine visiontechnology hasn’t been applied much to the field of mining engineering,especially tothe field of mineral processing. But now the working environment of coal washingneeds the advanced online testing system with intelligent control urgently. Thetechnology of online coal quality rapid detection has become a technical bottleneck,which restricting coal preparation process automation. For a machine vision system,the most critical thing is the image acquisition; any defects in image acquisition willbring the difficult to the subsequent image analysis and understanding. Meanwhile theimage acquisition is also the foundation of image processing, image compression andimage recognition, so the research and design of image acquisition system has veryimportant practical significance and value.In this paper,the structure and characteristics of image acquisition system werediscussed firstly, and its application in industrial inspection and mining engineeringfield was reviewed，then the topic s background and significance of the paperselection and research content were discussed.According to the CPP’s production, process requirements and on-line detectionsystem functional requirements, the overall scheme of image acquisition system weredesigned based on the hardware aspects of whole constitute. There are image dateacquisition section, lighting, camera anti-vibration and dedusting three subsystems inthis system: And the main hardware of the acquisition part was completed a detailedperformance comparison and selection.In order to eliminate the pollution of light, vibration, dust around the belt in CPP,and build a good environment for image acquisition, shading and lighting system weredeveloped and the camera anti-vibration and dedusting were discussed.Finally, the developed image acquisition system was built on the fine cleanedcoal belt in1#Branch of Taixi Coal Preparation Plant, a portion of the test results wasgiven in this paper. For the collected180coal images of different ash content, sevenfeature values were extracted and normalized, the recognition system of coal ashfractions was built based on optimal parameters of SVM. The above system s accuracy was validated used a test set, and ultimately the end of the test set coal ashprediction mean absolute error is0.12%, and the predicted result was all right. Sothefeasibility of image acquisition system was validated.Machine vision technology will play an important role to promote coal processintelligent control. The image acquisition technology of coal used was designed anddeveloped in this paper, which provided a successful demonstration for the onlineprocess control based on materials composition in CPP.