StanfordVL / GibsonEnv

Gibson Environments: Real-World Perception for Embodied Agents
http://gibsonenv.stanford.edu/
MIT License
866 stars 146 forks source link

how to deal raw data(mist/depth image) #139

Open songWell opened 1 year ago

songWell commented 1 year ago

hi, Thank you for providing the dataset. Now, I rectify equirectangular panoramic images, but I think the value in depth is wrong(I'm new to panoramic images and don't know how to deal with them), so should I do something like multiply by a ratio or use an intrinsic matrix? For example, in ("Tiny" Partition) Tolstoy scene, the depth value in the red circle is 1800. I think it should be 3600, so how to rectify?

point_p000001_view_equirectangular_domain_rgb

songWell commented 1 year ago

rectify rgb image rgb_num_0_0 3849237285596583_0 4739360750561077

songWell commented 1 year ago

I found it, is 1/512m? https://github.com/alexsax/2D-3D-Semantics

Depth: Depth images are stored as 16-bit PNGs and have a maximum depth of 128m and a sensitivity of 1/512m. Missing values are encoded with the value 2^16 - 1. Note that while depth is defined relative to the plane of the camera in the data (z-depth), it is defined as the distance from the point-center of the camera in the panoramics.