Improved real-time correlation-based FPGA stereo vision system

In this paper, we describe the design and implementation of Field Programmable Gate Array (FPGA) based stereo vision system. Implementation of correlation-based algorithm is well designed to exploit the parallelism in FPGA. The system is capable of running at more than 70 frames per second with 512×512 images. To improve the inherent weakness of simple correlation-based algorithm, two methods have been added. The multi-window approach improves disparity quality at object borders and cross checking reduces possible errors in general. The experimental comparison shows the improvement on final results.

[1]  Heiko Hirschm.  Improvements in Real-Time Correlation-Based Stereo Vision , 2001 .

[2]  John Woodfill,et al.  Real-time stereo vision on the PARTS reconfigurable computer , 1997, Proceedings. The 5th Annual IEEE Symposium on Field-Programmable Custom Computing Machines Cat. No.97TB100186).

[3]  Andreas Steininger,et al.  Hardware implementation of an SAD based stereo vision algorithm , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

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

[5]  Alfred Schmitt,et al.  Real-Time Stereo by using Dynamic Programming , 2003, 2004 Conference on Computer Vision and Pattern Recognition Workshop.

[6]  Ramin Zabih,et al.  Non-parametric Local Transforms for Computing Visual Correspondence , 1994, ECCV.

[7]  W. van der Mark,et al.  A comparative study of fast dense stereo vision algorithms , 2004, IEEE Intelligent Vehicles Symposium, 2004.

[8]  Geoffrey Egnal,et al.  Detecting binocular half-occlusions: empirical comparisons of four approaches , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).