On-line Technical Research Of Paper Lacuna Inspection

In most of the paper manuafacturers,the task of the paper defect inspection is compeleted through the manual detection method and this method can not meet the production requirements, because the width of paper is becoming wider and the speed of machine is becoming faster,so achieving the automation of this process is necessary. In order to achieve real-time paper image acquisition and processing, accompanying with image processing features which include the quantity of data, time-consuming process and requiring a lot of system resources to support, the paper studies some techniques associated with paper lacuna.The paper designs sheet inspection experimental platform which is based on studying and referencing the theory and experience on On-line sheet inspection system used in paper industry, and tests this sheet inspection platform by experiment. First, introduces the principle of defect inspection, builts the based-on acquisition card an PC inspection platform to meet requirement;secondly, develops a variety of programs to test this inspection platform,and finds platform inherent problems through analysising the results of the test.According to the problem of the inspection platform, combining with the characteristics of the sheet defect inspection, the paper developes a detection aloforithm and tests it.First, analyzes the main factors which caused errors of the CCD camera output video signal, and develop relevant solutions; secondly, through the normal paper texture image analysis, the paper shows the improved threshold to extract the region of disease paper, this method can extract useful information more effectively, and overcome the shortcoming of the traditional threshold detection method which can not be a good at defecting low-contrast disease; then, the paper introduces the thinking of the multi-scale wavelet analysis to detect the edge of the paper disease, and experiment shows that this method is well positioned edge of the disease; and uses fuzzy method to achieve classification of defferent disease by extracting the paper texture features; Finally, the paper tests this algorithm,and the experiment shows that acquisiton card and PC-based collection system combined with the algorithm can not meet the sheet On-line real time defect inspection.To improve the detection speed, combined with the application of DSP in image processing advantages, the paper try to design the test platform based on TMS320C6711B. This paper gives the overall framework of disease detection system, a brief introduction about TMS320C6711B hardware structure and function, according to the need for paper inspection system, introduces the power supply modules, clock modules, memory modules and the use of FIFO buffer pool as a high coordination with low-speed transmission of video information, and communication between the host and the DSP and other design ideas and specific schematic.