| Title |
A Realtime Stereo Vision System Design using Disparity Calibration |
| Authors |
서기덕(Gi-Duk Suh) ; 임준범(Jun-Beom Lim) ; 정승하(Seung-Ha Jeong) ; 박태근(Tae-Geun Park) |
| DOI |
https://doi.org/10.5370/KIEE.2023.72.7.855 |
| Keywords |
stereo vision; realtime system; adaptive support-weight; disparity calibration; VLSI design |
| Abstract |
In this paper, the error rate was reduced by using a calibration algorithm that can correct errors that occur in the initial disparity calculation stage using adaptive support-weights. Performance can be further improved by overlapping the initial disparity calculation and the disparity calibration stages. Because the adaptive support-weight algorithm includes nonlinear functions, it is impossible to reuse already processed data. In this study, it is designed to reuse data by separating rows and columns, and as a result, performance is improved and real-time processing is possible. To optimize the delay path, the exponential function is approximated with a step function, and a 5-stage pipeline is applied to the division operation. The proposed architecture was designed using VerilogHDL and synthesized with DBhitek 110nm standard cell library. The synthesized system used a maximum frequency of 666MHz (57fps) with 33K gates, and a memory of 4.9KB, and showed an error rate of 1.88% for tsukuba image. |