Image Processing Method Dased On Pde And Numericai Solution

With the fast development of computer and a microelectronic technology, digitalimage processing emerged and became a new discipline. In the finance,military, theindustry, the medicine, domains, agriculture, transportation, public security, survey,meteorology, address and so on ,it obtained the universal application. Moreoveralong with unceasing development of the theory, the method and the technical ofdigital image processing,its application will expand to further scope.In imagery processing method, misalignment method particularly based on PDEimagery processing method receives more attention because of the following tworeasons. On one hand, this method has a better accurately, when it compares with thetraditional linear processing method to. On another, it can maintain characteristic ofimage directly, such as the gradient, the geometric buckling and so on. It means thatimage processing of PDE has advantageous in establishing kind of numerical modelto express nimbly, therefore it becomes one important method of imageryprocessing.But compared with the additional nonlinear method, the calculation of PDEimage processing is large and slow, and the general method of PDE image processingcan lead to blurring of edge and image distortion. Based on these deficiencies of PDEimage processing when processing image, this thesis proposes an important newmethod of image processing of PDE, which can better maintain the image. Corollaryof the new method is based on an important feature that scale space is controlled by apartial differential equation, and has learned the concept that how does the heatconduction control the scale space. Compared with the other classical processingmethods, the new method not only has theoretical advantages, but also effectivelysolved the contradiction between the image edge features and the image being too smooth.Experimental results show that the new method not only can recovery theimage better, but also can keep the feather of the edge of image better. Furthermore,the equation derived does not contain mixed derivatives of mixed items, so algorithmis simplified and the computing speed is speeded up.