Implementation of Real-Time Post-Processing for High-Quality Stereo Vision

다국어 초록 (Multilingual Abstract)

We propose a novel post-processing algorithm and its very-large-scale integration architecture that simultaneously uses the passive and active stereo vision information to improve the reliability of the three-dimensional disparity in a hybrid stereo vision system. The proposed architecture consists of four steps — left-right consistency checking, semi-2D hole filling, a tiny adaptive variance checking, and a 2D weighted median filter. The experimental results show that the error rate of the proposed algorithm (5.77%) is less than that of a raw disparity (10.12%) for a real-world camera image having a 1,280 × 720 resolution and maximum disparity of 256. Moreover, for the famous Middlebury stereo image sets, the proposed algorithm’s error rate (8.30%) is also less than that of the raw disparity (13.7%). The proposed architecture is implemented on a single commercial field-programmable gate array using only 13.01% of slice resources, which achieves a rate of 60 fps for 1,280 × 720 stereo images with a disparity range of 256.

참고문헌 (Reference)

1 L. Yin, "Weighted Median Filters: A Tutorial" 43 (43): 157-192, 1996

2 J.I. Woodfill, "Tyzx DeepSea High Speed Stereo Vision System" 41-46, 2004

3 D.I. Han, "The Design of HD Image Rectification Architecture Using Floating Point IP" 6 : 39-44, 2014

4 L. Wang, "Stereoscopic Inpainting: Joint Color and Depth Completion from Stereo Images" 1-8, 2008

5 H. Jiang, "Research of Stereo Matching Based on Improved Median Filter" 4 : 479-486, 2011

6 C. Banz, "Real-Time Stereo Vision System Using Semiglobal Matching Disparity Estimation: Architecture and FPGAImplementation" 93-101, 2010

7 S.M. Choi, "Post-Processing Algorithms for Real-Time Active Stereo Vision" 1-2, 2014

8 S.A. Fahmy, "Novel FPGA-Based Implementation of Median and Weighted Median Filters for Image Processing" 142-147, 2005

9 R. Zabih, "Non-parametric Local Transforms for Computing Visual Correspondence" 151-158, 1994

10 정재찬, "High-Quality Stereo Depth Map Generation Using Infrared Pattern Projection" 한국전자통신연구원 35 (35): 1011-1020, 2013

11 J. H. Chang, "High-Quality Stereo Depth Map Generation Using Infrared Pattern Projection" 2014

12 D. Scharstein, "High-Accuracy Stereo Depth Maps Using Structured Light" 195-202, 2003

13 S.B. Kang, "Handling Occlusions in Dense Multi-view Stereo" 103-110, 2001

14 S. Jin, "FPGA Design and Implementation of a Real-Time Stereo Vision System" 20 (20): 15-26, 2010

15 C. Çigla, "Efficient Edge-Preserving Stereo Matching" (6–13) : 696-699, 2011

16 C.C. Pham, "Domain Transformation-Based Efficient Cost Aggregation for Local Stereo Matching" 23 (23): 1119-1130, 2012

17 F. Tombari, "Classification and Evaluation of Cost Aggregation Methods for Stereo Correspondence" 1-8, 2008

18 C. Tomasi, "Bilateral Filtering for Gray and Color Images" 839-846, 1998

19 A. Hosni, "Accuracy-Efficiency Evaluation of Adaptive Support Weight Techniques for Local Stereo Matching" 7476 : 337-346, 2012

20 D. Scharstein, "A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms" 47 (47): 7-42, 2002

21 Y. Jia, "A Miniature Stereo Vision Machine (MSVM-III)for Dense Disparity Mapping" 728-731, 2004

22 J. Heikkila, "A Four-Step Camera Calibration Procedure with Implicit Image Correction" 1106-1112, 1997

23 S.M. Choi, "A FPGA Based Real-Time Post-Processing Architecture for Active Stereo Vision" 1-2, 2014