这里重点是摄像头运动估计和定位,不是地图或者路标(landmark)。以前大家都知道SLAM结合深度学习最多的是语义SLAM,比如语义分割和语义目标识别。而这里强调的是里程计和定位。另外,忽略一些特征提取和匹配的方法。
https://github.com/lmb-freiburg/demongithub.com
https://github.com/tinghuiz/SfMLearnergithub.com
https://sites.google.com/view/vid2depthsites.google.com
https://github.com/hlzz/DeepMatchVOgithub.com
ChiWeiHsiao/DeepVO-pytorchgithub.com
krrish94/DeepVOgithub.com
https://github.com/MightyChaos/LKVOLearnergithub.com
https://github.com/nianticlabs/monodepth2github.com
Huangying-Zhan/Depth-VO-Featgithub.com
https://github.com/JiawangBian/SC-SfMLearner-Releasegithub.com
https://github.com/yzcjtr/GeoNetgithub.com
guangmingw/DOPlearninggithub.com
Huangying-Zhan/DF-VOgithub.com
https://sites.google.com/view/struct2depthsites.google.com
saic-vul/odometrygithub.com
augustelalande/sfmgithub.com
https://github.com/yan99033/CNN-SVOgithub.com
https://github.com/lmb-freiburg/deeptamgithub.com
https://github.com/devendrachaplot/Neural-SLAMgithub.com