Three-Dimensional Shape Recovery Based on Single Historical Image
Limei Song, Yuhua Wen · 2009
For unable to establish geometrical relationship of coordinates, conventional methods of 3D shape recovery can not be adopted to a great number of precious historical images to recover the 3D shape of the objects and environments in the images. The new method of 3D shape recovery based on single frame historical image is proposed in this paper. The core idea of this research is that, according to the change of the shade of gray of each pixel, the tilt angle and slant angle of the surface vertical vector of each object's pixel in the image is analyzed, then the depth of each pixel is calculated in response to the tilt angle and slant angle. Simulating human being's vision system and the process of thinking of brain, this method makes the process of man's 3D comprehension of single image computerize to obtain 3D digitized data. These 3D digitized data can be used in the fields of the 3D recovery of historical relics and the identification of 3D portrait and stereo literal.