论文信息 - FPGA Implementation of Optical Flow Algorithm Based on Cost Aggregation

FPGA Implementation of Optical Flow Algorithm Based on Cost Aggregation

In this paper, we describe an FPGA implementation of an optical flow estimation algorithm based on cost aggregation. The cost aggregation algorithms are mainly used in stereo vision systems, and show good results. We extended a cost aggregation algorithm for the optical flow estimation, and implemented it on FPGA. The error rate by our system is significantly lower than previous hardware systems, and its processing speed is fast enough to generate a dense flow map of high resolution image in real-time.

Tsutomu Maruyama | Yu Tanabe

[1] Tsutomu Maruyama,et al. Fast and Accurate Stereo Vision System on FPGA , 2014, TRETS.

[2] Dah-Jye Lee,et al. A Comparison Study on Implementing Optical Flow and Digital Communications on FPGAs and GPUs , 2010, TRETS.

[3] Eduardo Ros,et al. A Comparison of FPGA and GPU for Real-Time Phase-Based Optical Flow, Stereo, and Local Image Features , 2012, IEEE Transactions on Computers.