Study Of Edge Detection And Image Restoration And Its FPGA Realization

Image edge detection and restoration are very important in the image processing, which are used widely in the pattern recognition, image segmentation, and image enhancement. After studied for several decades, there have appeared several classical edge operators, such as Roberts, Sobel, Laplacian, Canny and LOG. Meanwhile, there also have presented several classical algorithms of image restoration, such as Inverse Filter, Wiener Filter and Constrained Least Squares Filter. Along with the developing of the new technology and the related theories, new algorithms for edge detection and image restoration have being come forth increasingly. The study of new methods of edge detection and image restoration is showed in this paper, which based on the model of Cellular Automata (CA) and Statistical Mechanics.CA is a kind of dynamical systems in which space, time and state are discrete, having the parallel processing framework. As an ideal model to implement a complicated system, CA becomes one of the hotspots of study owing to its capability to simulate and solve many actual problems. Because the two-dimension square model of CA can correspond to the two-dimension image well, it is possible to put the CA model to the field of image processing. The neighborhood of Moore CA model is applied to make a new algorithm of edge detection in this paper. By testing several images and comparing with the result of classical method, the characteristic of the new algorithm can be proved.According to the fast developing of micro-electronics technology, it is possible to make an algorithm turn into hardware model. For its characteristic of parallel logic and dispersed state, CA can be realized easily in hardware. Thus, the VHDL is carried to make a design for the forenamed algorithms, and the design is validated by simulation. The result of simulation shows that the new algorithm of edge detection which is set up in this paper has done a good effect on image's edge detection, and the processing speed has improved.In addition, a new algorithm of image restoration which combines CA edge detection method and statistical mechanics model is discussed. By applying statistical mechanics to image restoration, it is likely to modeling for a given...