Dynamically adaptive real-time disparity estimation hardware using iterative refinement

Abstract

a b s t r a c tThe computational complexity of disparity estimation algorithms and the need of large size and bandwidth for the external and internal memory make the real-time processing of disparity estimation challenging, especially for High Resolution (HR) images. This paper proposes a hardware-oriented adaptive window size disparity estimation (AWDE) algorithm and its real-time reconfigurable hardware implementation that targets HR video with high quality disparity results. Moreover, an enhanced version of the AWDE implementation that uses iterative refinement (AWDE-IR) is presented. The AWDE and AWDE-IR algorithms dynamically adapt the window size considering the local texture of the image to increase the disparity estimation quality. The proposed reconfigurable hardware architectures of the AWDE and AWDE-IR algorithms enable handling 60 frames per second on a Virtex-5 FPGA at a 1024 Â 768 XGA video resolution for a 128 pixel disparity range.

A real time power line detect system based on stereo vision

“…Wang et al [3] transplant AD-Census algorithm into FPGA with semi-global optimization after cost aggregation, which is able to process high definition images of 1600*1200 pixels at 42 frames per second. Abdulkadir et al [5] proposed a new algorithm named AWDE, which is very suitable to be implement on hardware, they implemented their algorithm on Xilinx Vitex-5 and can achieve 60 f/ps at resolution of 1024*768 with 128 disparity levels. Zhang et al [6] combine the mini-census transform and cross-based cost aggregation in their structure, which achieves 60 frames/s at 1024×768 pixel stereo images.…”

Trinocular adaptive window size disparity estimation algorithm and its real-time hardware

“…Various hardware architectures presented in the literature provide real-time DE using one pair of cameras [4][5][6][7]. Some implemented hardware architectures only target CIF or VGA video [4][5].…”

“…Some implemented hardware architectures only target CIF or VGA video [4][5]. Several of these implementations succeed to reach real-time for HR [6][7]. Real-time DE for HR images offers some crucial advantages compared to low-resolution DE such as increasing the object definition or providing more precise depth measurements.…”

“…However, the hardware implementation of high-quality real-time HR stereo video is challenging due to high computational load of this process and its severe demand on off-chip and on-chip memory size and bandwidth. Recently, a hardware-oriented adaptive window size disparity estimation (AWDE) algorithm and its real-time reconfigurable hardware implementation that targets HR video with high-quality disparity results has been presented in [7]. The AWDE hardware processes 60 frames per second (fps) on a Virtex-5 FPGA at a 1024×768 XGA video resolution and 128 pixels disparity.…”

“…This paper presents hardware-oriented and high-quality trinocular adaptive window size disparity estimation (T-AWDE) algorithm and its real-time hardware. The T-AWDE algorithm and its hardware are the enhanced versions of the recently published binocular AWDE algorithm and its implementation [7]. The implemented T-AWDE hardware is the first hardware implementation that succeeds real-time trinocular DE for HR video.…”