Full HD Real-time Depth Estimation Algorithm and Hardware Implementation

2014 
This paper presents a novel implementation for realtime depth estimation which can support full HD resolution for 3D video stream. We propose a package of solution to address the major difficulties in depth estimation, such as expensive computation, depth accuracy, real-time implementation and high definition progressing. Different from the typical single stereo vision methods, we take fusion strategy which can reduce errors caused by individual measure. The FPGA-based design is selected for real-time implementation which can achieve the maximum processing speed of 125fps, with maximum disparity search range of 240 pixels, full HD (1920 × 1080p) resolution image. Multi-resolution method which includes interpolation method and synthesizers are used for improve the efficiency and stability. The implementation can be included in video system for live 3DTV application and can be used as an independent hardware module in low-power application. KeywordsDepth estimation; full HD; hardware implementation; Real-time
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    8
    References
    0
    Citations
    NaN
    KQI
    []