Image Process And Three-dimension Reconstruction Technical Research Based On Human Entity Slices

Image process and three-dimension reconstruction technical research that bases on human entity slices is one of the important contents of digital visual human researching and mechanics equivalent artificial dummy. The research of the digital visual human body is one of the hot technologies of mankind's scientific technical research in the 21st century; it has a great meaning in the field of medical research. Developed countries begin on his field's research in later stage of the 80s, and make certain achievement. Our country formally listed "several key technology of digitized visual human body" in 863 projects just in 2001, entered in the starting stage of the digital visual human body research. The mechanics equivalent bionic dummy is a "substitute" under various kinds of complicated conditions of human body; it is the "steady receptor" to replace the mankind carried on the safe experiment. Which get across simulation various kinds of biomechanics' characteristics of the human body's, widely applies such fields as safety testing, spaceflight aviation, dress designing, etc... In the field of the mechanics equivalent bionic dummy, the developed country has already carried researches on the automobile seat, lathe safety protection, design of the assembled production line, dress designing, etc. through the bionic dummy, and make great progress, but it is still in fetch in and modeling stage that our country's research on the mechanics equivalent bionic dummy, and the theoretical research and development all far lag behind the developed country.To this current situation, the article has done some researches of image process based on the human entity slices, and have realized fringe application in manufacture the mechanicsequivalent bionic dummy by the result of the slices reconstruction.First, the article does coherence disposal to the serial slices images' brightness, contrast, size of area and relationship of position. It through calculate the color marks difference between the serial images, establishes mathematic model of the image color adjust, programs the adjust program of the serial images under VC++, and makes originality handle to the bitmap of the serial image, it does differential stretching compensation to the colorific components (RGB) of the DIB Pixels and makes the brightness and contrast of the serial images into a unitary form. It improves the image registration method that based on the external signs. The coordinates of the external mark point centers can be got accurately with Fourier convolution transform; realizing rapidly image aberration adjustment and cutting the superabundance areas with double linear coordinate transform; the precision registration images can be got by getting and refilling the pixels deviating from grids with cube convolution progresses pixels. Those problems can be eliminated that lead by the equipments and the setting when we gain the images of human entity slices.The article has realized exactitude picking-up the target of the entity images. It works over the color characteristic of the slice images, according to the characteristic of the pictures, improved the arithmetic of the follow edge, isolate the picture of red channel from the colorized picture, utilize the characteristic that picture goal edge of red channel and background color value differ greatly, adopts a threshold value changed to follow algorithms, searches for the edge to the picture, draws out the accurate image outline, realized the purpose of the picture's area segmentation.Finally, the article has accomplished 3D reconstruction of the images and manufacturing the entity model. After researches the 3D surface reconstruction method based on body draw, it choose CATIA software to finish the cloud building and disposal, and have realized manufacture the head entity's model through rapid prototype.The article's harvest has some theoretical and practice significance to the research of "the model of the visual human and the model of the dummy".