3-D Object Projection Method of Multi-Camera Revolve Batch Rendering

Abstract

3-D object recognition was different from 2-D one. In order to improve recognition rate, it was necessary to establish a full features base. With object as center, the paper offered a view space projection method of Multi-Camera Revolve Batch Rendering. Firstly, constructed three dimensional models for object, set positions for model and camera of top view. Then by the max panel stick to one of axis angle, copied camera with equal interval. Set up the moving path for hemisphere, and constraint to corresponding camera. In the last, established batch script with output parameters, automatic preservation random projection images for three dimensional objects.