1 use testarea topology lat/lon to get a rawpano positon and yaw, all raw data in raw 2 use unique to get a unique line from rawpano.csv as pano_pose.csv (since there are some repetitive pano) 3 use pano_pose as a reference in the google map 4. delete avenue with vegetation 5. Final panopose as posepano.csv 6. use the python script streetview_gt.py to download all panorama and depth in the test area 7. notice our special way to register depth map
Illustrate 3D point of an example of street view panorama
The coarse matching bewteen google street view and its depth map
To rectilinear the panorama in a proper way
verify the depth registering way is corecct