A Reconfigurable Disparity Engine for Stereovision in Advanced Driver Assistance Systems

Depth extraction in stereovision applications is very time-consuming and requires hardware acceleration in real-time context. A large number of methods have been proposed to handle this task. Each method answers more or less to real-time constraints, depending on the applicative context and user’s needs. Thus, flexibility is a strong requirement for a generic hardware acceleration solution, particularly when ASIC implementation is targeted. This paper presents REEFS, a reconfigurable architecture for embedded real-time stereovision applications. This architecture is composed of three reconfigurable modules that enable flexibility at each step of depth extraction, from correlation window size to the matching method. It generates VGA depth maps with 64 disparity levels at almost 87 frames per second, answering hard real-time requirements, like in Advanced Driver Assistance Systems.

[1]  Pieter P. Jonker,et al.  Real-time Stereo with Dense Output by a SIMD-computed Dynamic Programming Algorithm , 2002, PDPTA.

[2]  Pascal Fua,et al.  A parallel stereo algorithm that produces dense depth maps and preserves image features , 1993, Machine Vision and Applications.

[3]  Andreas Steininger,et al.  Flexible Hardware-Based Stereo Matching , 2008, EURASIP J. Embed. Syst..

[4]  Geovany de Araújo Borges,et al.  VoC: a reconfigurable matrix for stereo vision processing , 2006, Proceedings 20th IEEE International Parallel & Distributed Processing Symposium.

[5]  Jonathan M. Garibaldi,et al.  Real-Time Correlation-Based Stereo Vision with Reduced Border Errors , 2002, International Journal of Computer Vision.

[6]  J. Lazaro,et al.  Real-Time Stereo Vision Processing System in a FPGA , 2006, IECON 2006 - 32nd Annual Conference on IEEE Industrial Electronics.

[7]  Reinhard Männer,et al.  Calculating Dense Disparity Maps from Color Stereo Images, an Efficient Implementation , 2004, International Journal of Computer Vision.

[8]  John Iselin Woodfill,et al.  Tyzx DeepSea High Speed Stereo Vision System , 2004, 2004 Conference on Computer Vision and Pattern Recognition Workshop.

[9]  D. Scharstein,et al.  A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms , 2001, Proceedings IEEE Workshop on Stereo and Multi-Baseline Vision (SMBV 2001).