Research And Implementation Of Key Technologies For3DTV Mobile Terminal Based On Clond

With3DTV or Three-dimensional Television, observers can get the stereo feeling.3DTV based on mobile terminal is utilizing mobile devices to process and present the3D image. In order to achieve this, problems like3D video sources, limitation of mobile devices’hardware resources should be solved, and a higher efficiency encode and decode algorithm, also the image rendering method should be applied.Cloud or Cloud Computing is a hot topic nowadays, and has changed the way of acquiring and processing data. The strong computing ability and storage capability of cloud can ease the limitation on mobile terminal and enable the realization of mobile3DTV system. The3D video acquisition, encode transforming and transmission can all be done on cloud. The mobile terminal only need to receive, decode and render the3D video from the cloud, can then display stereo image on the terminal. Achievements are as follows:(1) A platform of3DTV mobile terminal based on cloud and the framework of the cloud end and terminal end is proposed. Based on this platform, the data format, data flow and data storage method is presented. This paper also discusses the resource scheduling method for the3D video service on this platform.(2) Based on H.264standard and X264encoder technique, a dynamic encoding method is designed to fit for the cloud platform. Also an enhanced intraframe encoding method is proposed to increase the encode efficiency of SBS (side-by-side) format3D video, which can achieve a higher compression efficiency while maintaining the image quality. Besides, this paper realizes the3D video decode on the terminal platform based on the Davinci technology.(3) Different rendering method should be designed for different terminal display requirement. This paper realizes a multi-view image rendering method based on slanted lenticular auto-stereoscopic display. The principle of lenticular lens is presented and the RGB sub-pixel arrangement algorithm is proposed. Finally this paper realizes the rendering and multi-view interlacing on FPGA, designed the hardware architecture and logic blocks. Compared with the software design outcome, the hardware implementation experimental results represent higher process efficiency while maintaining the image quality, and can be improved to adapt to future SoC design.