Giters

FanglinBao / HADAR

This is an LWIR stereo-hyperspectral database to develop HADAR algorithms for thermal navigation. Based on this database, one can develop algorithms for TeX decomposition to generate TeX vision. One can also develop algorithms about object detection, semantic or scene segmentation, optical or scene flow, stereo depth etc. based on TeX vision instead of traditional RGB or thermal vision.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Heat cube generation

Bera97 opened this issue · comments

commented

Hi Fanglin,

first of all thank you for this great project. I have a few questions.

  1. Is the heat cube a standardised output of the FLIR A325sc thermal camera you used? If it is not, what steps did you have to make in order to construct the heat cube?
  2. Did you try stereo depth or you only did mono depth using GCNDepth model? If you tried stereo depth did you use any neural network for improved disparity estimation and which one?

Thank you very much.

commented

Hi Bera97,

Thanks for your interest in our work.

  1. The heat cubes in this HADAR database are mainly synthesized through Blender. Please check the readme file of the database as well as the Monte Carlo Path Tracing section in Methods of the HADAR paper for a broad description. None of the heat cubes is from FLIR A325sc. Scene11 was collected by a pushbroom hyperspectral imager.
  2. For Scene11, the real-world experiment, we used GCNDepth since we only have one view. For Scenes1-10, we used DeepPruner (pre-trained on the SceneFlow dataset) for binocular stereovision, please check Fig.S19 in the Supple. Info. of the HADAR paper.
commented

Thank you very much for the information!