https://github.com/uzh-rpg/rpg_open_remode 源码位置
单摄像头稠密实时重建
文章2014 ICRA: http://rpg.ifi.uzh.ch/docs/ICRA14_Pizzoli.pdf
Abstract— In this paper, we solve the problem of estimatingdense and accurate depth maps from a single moving camera.A probabilistic depth measurement is carried out in real timeon a per-pixel basis and the computed uncertainty is used toreject erroneous estimations and provide live feedback on thereconstruction progress. Our contribution is a novel approachto depth map computation that combines Bayesian estimationand recent development on convex optimization for imageprocessing. We demonstrate that our method outperforms stateof-the-arttechniques in terms of accuracy, while exhibitinghigh efficiency in memory usage and computing power. Wecall our approach REMODE (REgularized MOnocular DepthEstimation) and the CUDA-based implementation runs at 30Hzon a laptop computer.
大家研究这方面的,可以测试之后,留言给后来者个参考。谢谢!