Design And Implementation Of On-line Paper Surface Defects Detection System Based On Image Processing Technology

In our life, people use the paper every day. Paper, on the other hand, is also indispensable raw materials in industrial production. It is complex procedure to convert raw materials to paper, ariability of process and the environment makes the paper surface may cause spots, holes, fold, ragged edges, which will reduce the paper quality and quality. The traditional detection methods are mainly rely on people’s eyes. Because those methods have many disadvantages, for example, low efficiency, high failure rate, low real-time and high labor intensity, the methods are not widely adopt in paper-making procedure. In order to improve the competitiveness of the paper products, an on-line paper surface defects detection system based on image processing technology is designed in this paper. The main processe of on-line paper surface defects detection system is as follows, the CCD acquisits paper surface image and transport the image to the detection system for preprocessing, which include filtering algorithm, image segmentation and simulation. With pre-designed algorithm, the system decides whether paper surface has defect. This paper adpots Microsoft Visual C++6 to program, combing advantages and disadvantages of iteration method and Otsu, we put forward an improved algorithm, which segmentates the gray-scale image into sub-regions and uses improved iterative method to calculate threshold for each sub-region and the average gray value, then obtain the optimal threshold of the whole image. Experimental results show that improved threshold segmentation method not only can split out high-contrast paper disease images, can also split out most of the low-contrast image paper defect. The improved threshold segmentation algorithm is integrated into the software system. In the guarantee of the light source system, image acquisition system is stable under the test of system, the results show that the detection system can achieve the expected technical indicators, and realize paper defect. The recognition rate of the designed on-line paper surface defects detection system is as high as 90% in filed testing.