Research And Implementation Of The Virtual Camera’s Simulation

The virtual camera is an important part of the animation, movies, games and otherfields. Through the virtual camera, by rendering techniques generate images, the vividimage will been shown to the audience, and so the audience (user) get true feelings.This paper is based on the virtual test system of unmanned vehicle as the platform.In the virtual environment, the virtual camera simulates the real camerawork, forthe real images and virtual images matching and it is used in application of lanerecognition, traffic lights, obstacle detection and extraction variousalgorithms, simulation of virtual images and real images. The premise is to verifythe similarity conditions of virtual image and real image, to study thevarious experiments of the unmanned vehicle in virtual scene. So that it is veryimportant to study the virtual camera.The establishment of the virtual environment is created by Director3D, thecalculation and the rendering is finished in GPU. The depth and color information ofthe objects in the scene is gotten in the imaging module. As is known to all, Object outof the depth of field has a certain degree of ambiguity and the degree ofambiguity relates the distance to DOF. There are many fuzzy methods, but the methodsdon’t consider the distance of the edge of the object, because the distances ofthe pixel around edges are different. The pixels around the edges have differentcontribution to edge pixels’ fuzzy. So this paper is proposed to identify the pixelsinvolved weights according to the depth of information, to realizethe effective simulation for camera depth of field. The paper has finished the analysisof parameters of the virtual camera, the analysis and design of the three modules, and the realization of the basic functions of the virtual camera. In the last, the present workis summarized, and the future work is prospected.