A Paper-counting Device For Ultrahigh Stack Based On Machine Vision

In paper processing factories,paper quantity detection is very important to profit and survival.Both primitive measurement and mechanical measurement have advantages and disadvantages,but it will not conform to modern efficiency mode of production soon.Although paper-counting devices based on machine vision are springing up in the market,they can not detect ultrahigh paper stacks.Therefore,we design a device for detecting ultrahigh paper stacks’ quantity,and its main functions have been basically achieved.The research contents are as follows:Firstly,in the beginning chapter,the background and significance of research are introduced;the concept and application of machine vision are described.The author presents research status of paper-counting at home and abroad.The description of large-scale imaging technology is emphasized.Secondly,imaging system is designed through analyzing characteristics of paper stack.The core components of imaging system including camera,lens,light source and linear module are selected combining with the main parameters and experimental data.This paper introduces the importance of lighting system through changing illuminating angle.And it also puts forward that using linear module to move a camera collecting sequence images so as to set up a proper imaging system.Thirdly,sequence images are stitched with external added feature in the thesis.Because the content of images is highly consistent,the author gives a black ruler as a special feature in images so as to do successful registration with calculating correlation.Fourthly,the modified line detector is applied to enhance effective information in images.It simplifies the computation of original line detectors and improves the performance using the paper characteristics of single direction and slender shape.Finally,through practical applying and field testing,both mathematical morphology method and peak-finding method are discussed.Mathematical morphology method extracts the peak points in enhanced images.Due to peak points generally exist in the center of paper,extracted result is fracture lines with single pixel width.After a series of mathematical morphology operation,the number of paper could be computed from the processed image.Peak-finding method is to use large-scale slender filter to eliminate pixels’ personality and extract the common characteristic of each column of pixels.At last,paper number can be gotten through analyzing the number of peaks in waveform.