Digital Image Processing Based On Dynamic System Compute

Digital image is a 2-D matrix. Image processing may be realizedby linear matrix transform method. An image matrix could be seen a2-D linear discrete dynamic system. The algorithm and method of thedynamic system can be used into image processing.This dissertation discusses the algorithm and the application ofautomaton theory on black-white image processing. Image processingand automaton are combined in the research on image edge problem. Anew automaton based on Freeman chain code are presented. The modelof traffic is proposed as the application of automaton theory. At last, asoftware has programmed to solve the image form problem.In this dissertation, research works have been launched in thefollowing aspects:1. The automaton for image edge is proposed based on automatonand eight-direction Freeman chain code. The realization of the chaincode automaton has been used in the research on image edge processing.2. The raster filling algorithm of computer graphics theory istransplanted into a new filling algorithm based on chain code. A newboundary point classifying method is defined by using boundaryFreeman chain code. The automaton fills pixels between left point orright point of contour and raster line with complement operation, thus itcan handle arbitrary complicated region without supplementary memoryspace or additional labeling.3. A new perfect rule based on one-direction and two-directionimage edge labeling is proposed. This rule considers not only the pixelsof the left and right edge but also the pixels of the top and bottomregion contour. The left and right boundary of the image can be decidedaccording to the direction of boundary tracing. The length of every line of the closed region can be got by the weights of horizontal coordinates.The length of the image line can be used to compute region area andgeometric moment of the image.4. The new software of form recognition and text filled in theblanks is realized. Firstly, the original form transfers to digital formimage by scanner. Secondly, fill the formatted text content in the formimage, then save the image data and formatted text content separately inone file. The software has defined a new efficient file format, whichcontains image data and formatted text.5. In the programming of the form software, an automaticrecognition algorithm based on image edge automaton is given. And themethod which adjust skew form has been used in the application of theform processing. The software has finished, and has a good performancein the application of the form.