Conversion Technology And Implementation Of Real-time 2D-3D Based On FPGA

With the development of hardware technology and intelligent TV system, and more and more families hope to remain within doors at home watching 3D movies, manufacturersraced to launch 3D TV. Algorithm of converting 2D to 3D is particularly important in the 3D TV. After traditional 2D disc sources converts into 3D format, it can greatly enhance sense of perception. A good converting algorithm can guarantee good converting effect and the efficiency, meanwhile it brings great profits to the manufacturer. Thus, research on real-time HD video converting system of 2D to 3D has gained more and more attention.This paper adopts a video converting system of 2D to 3D designed by Altera company’s FPGA development board. The system gathers 2D camera image information from the external D5M, then goes through the FPGA image preprocessing, motion detection, depth extracting,3D synthesis algorithm, finally is displayed by DVI. This paper implements Verilog language design and achieves the key algorithms in the system, including color space converting module, median filtering module, morphological processing module, motion detection module, depth extracting module, cavity filling module etc..Statement of the papers in each module will first explain the principle of algorithm, and list several common algorithms and other algorithm by using FPGA, to describe the detail of the hardware implementation on the use of the algorithm in this paper, at the same time, give the effect picture. The most crucial and difficult part in converting system is the implementation of motion detection, depth extraction and cavity filling algorithm. This paper will separately describes these three parts by three chapters in detail. Using the block matching method to implement the motion detection and depth extraction algorithm. For the cavity filling algorithm, this paper will implement the fill of big and small cavities, and restore the matching error.