search

chelicynly / A-Deep-Learning-Approach-to-Drone-Monitoring

A-Deep-Learning-Approach-to-Drone-Monitoring
https://github.com/chelicynly/A-Deep-Learning-Approach-to-Drone-Monitoring
32 stars 5 forks source link

Thermal labels #3

Closed seawee1 closed 5 years ago

seawee1 commented 5 years ago

Hi, I'd like to use the thermal usc drone dataset. The ''images and labels'' folder however only contains two drone sequences and the labels are missing. For the 10 videos that are provided I also couldn't find any labels.

seawee1 commented 5 years ago

I labeled the thermal part of the dataset by myself. If someone is interested, contact me.

MuhammadAsadJaved commented 3 years ago

@seawee1 Can you share the labels Please? asadjavedgujar@gmail.com

Thanks

seawee1 commented 3 years ago

Dear Asad,

here you go. Could take a day until Google Drive will not raise the quota error however. https://drive.google.com/file/d/1oc4xCwpsaeXUAh_EqWYb0hzC6rhIcqWe/view?usp=sharing

I don't think I labeled all of the TIR images, but it should be around a few thousand labels or so. I can't really remember. As daytime data only had keyframe annotations, I added some additional labels in between keyframes for better bounding box interpolation results on these images.

The pre-processed dataset is in ImageNet format and also includes 2 TIR drone sequences of the Linkping Thermal InfraRed (LTIR) dataset. If you need more information about how the pre-processed dataset was created, I can send you an excerpt of my masters thesis.

Greetings Sebastian Rietsch

Am Di., 15. Sept. 2020 um 04:58 Uhr schrieb Asad notifications@github.com:

@seawee1 https://github.com/seawee1 Can you share the labels Please? asadjavedgujar@gmail.com

Thanks

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/chelicynly/A-Deep-Learning-Approach-to-Drone-Monitoring/issues/3#issuecomment-692432107, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK3AKUZJS25NM2BSELPDQWLSF3J3PANCNFSM4IT4JTRQ .

MuhammadAsadJaved commented 3 years ago

Dear Sebastian Rietsch, Thank you so much for your cooperation. They also provide some annotations (only for detection results) for Thermal dataset in the Thermal_data/videos/Detection Results/ folder. But all the annotations are in one .txt format. I have converted them to voc/coc format (in .txt and .xml) format. if you can then let me know. I will share it with you. or I will only share the python program and you can do by yourself. it can also use for other sets. Thanks Sebastian Rietsch

On Wed, Sep 16, 2020 at 5:40 PM Sebastian Rietsch notifications@github.com wrote:

Dear Asad,

here you go. Could take a day until Google Drive will not raise the quota error however.

https://drive.google.com/file/d/1oc4xCwpsaeXUAh_EqWYb0hzC6rhIcqWe/view?usp=sharing

I don't think I labeled all of the TIR images, but it should be around a few thousand labels or so. I can't really remember. As daytime data only had keyframe annotations, I added some additional labels in between keyframes for better bounding box interpolation results on these images.

The pre-processed dataset is in ImageNet format and also includes 2 TIR drone sequences of the Linkping Thermal InfraRed (LTIR) dataset. If you need more information about how the pre-processed dataset was created, I can send you an excerpt of my masters thesis.

Greetings Sebastian Rietsch

Am Di., 15. Sept. 2020 um 04:58 Uhr schrieb Asad <notifications@github.com

:

@seawee1 https://github.com/seawee1 Can you share the labels Please? asadjavedgujar@gmail.com

Thanks

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub < https://github.com/chelicynly/A-Deep-Learning-Approach-to-Drone-Monitoring/issues/3#issuecomment-692432107 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/AK3AKUZJS25NM2BSELPDQWLSF3J3PANCNFSM4IT4JTRQ

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/chelicynly/A-Deep-Learning-Approach-to-Drone-Monitoring/issues/3#issuecomment-693293892, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG4GR5H4BC4Y43E3327AZW3SGCBXPANCNFSM4IT4JTRQ .