Based on point cloud and texture data generation algorithm of two-dimensional images
Ke Xiao, Wang Lijing · 2016
Optical Character Recognition (OCR) are widely utilized in two-dimensional (2D) image, but can not be utilized in the three-dimensional (3D) model. The direct method gets the screenshot of 3D text model and save it as a BMP image. But this way will make the image has serious distortion. In this paper, based on texture mapping principle, using geometric knowledge to convert a 3D model into 2D images. Experimental result shows that the OCR rate is high, efficiently and accurately restore the original 3D model.