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

About real-world dataset and matlab test data

TooZE23 opened this issue · comments

commented

Hello Dr. Bao,
Thank you for doing such a remarkable job. I've spent the last weeks studying the paper and the code and have now reproduced the experimental results. There are two issues that are currently bothering me:

  1. How was the dataset for the 11th real-world scene obtained? Is it related to the matlab test data test.bsq?
  2. I want to test more data in matlab, and how to use heatcube_001.mat-heatcube_004.mat for matlab testing? I tried to use wavenumber.mat, matlib_full.mat and Huelib_full.mat for this but the experimental result as shown in the picture is not good.
    图片
commented

Hello,
Thanks for your interest in our work.
Scene 11 was generated by interpolating test.bsq. This is to make it consistent with the synthetic scenes for machine learning.
If you check the demonstration videos, you can find more TeX vision frames generated by the Matlab code (TeX-SGD). Properly setting the environmental radiance is important. You may want to double check this in your implementation.

commented

Hello Dr. Bao,
Regarding the results of heatcube_001, we still have questions. The following figures are the emap obtained by the TeX algorithm and the gt, respectively, and there is still a large difference between the two. Have you processed the data of scene 11 with matlab code and I wonder is this a normal result?
image
image

commented

Hi,
Yes, I checked scene 11 before we uploaded it. Our result of eMap was not like yours. Again, the environmental radiance matters.
I have multiple deadlines to meet in the coming month. Unfortunately, my further response will be delayed. Let's keep this case open. I will catch up with you soon.
Thanks.

commented

Hi, Yes, I checked scene 11 before we uploaded it. Our result of eMap was not like yours. Again, the environmental radiance matters. I have multiple deadlines to meet in the coming month. Unfortunately, my further response will be delayed. Let's keep this case open. I will catch up with you soon. Thanks.

Okay. Thank you for the information.

commented

Hello Dr. Bao,
Sorry to bother you. After you have finished your work, could you please tell me exactly how to set the environment radiance correctly?
We import the mat data using the following codes. Is there something wrong?
HSI0=load('heatcube_0001.mat');
HSI=HSI0.HSI;
S_EnvObj0=load('S_EnvObj_0001.mat');
S_EnvObj=S_EnvObj0.S_EnvObj;
matLib0=load('matLib_FullDatabase.mat');
matLib=matLib0.matLib;

commented

Hello Dr. Bao, Sorry to bother you. After you have finished your work, could you please tell me exactly how to set the environment radiance correctly? We import the mat data using the following codes. Is there something wrong? HSI0=load('heatcube_0001.mat'); HSI=HSI0.HSI; S_EnvObj0=load('S_EnvObj_0001.mat'); S_EnvObj=S_EnvObj0.S_EnvObj; matLib0=load('matLib_FullDatabase.mat'); matLib=matLib0.matLib;

Besides, we convert the 'test.bsq' data to 'mat' data format and adopt the above codes to import data, the results are correct. We use the same codes to process 'heatcube_0001.mat' and the results are wrong, as shown in the figure above.