search

arclab-hku / fast-dodging

dodging fast small moving objects with RGBD sensor
23 stars 1 forks source link

Datasets #1

Open yinloonga opened 1 year ago

yinloonga commented 1 year ago

Good job! I would like to know if you saved the rosbag datasets when you did the experiment. I would like to see how the code runs in the simulation environment first. If so, would you consider making it publicly available for people to test and use. Thank you!

MichaelLu-hku commented 1 year ago

Thank you for your interest in our work. I promise will prepare it as soon as possible. Meanwhile, we provide the tennis ball detection model. If you have a tennis ball and a depth camera you can try to run the code in real time as well.

yinloonga commented 1 year ago

Okay, thank you very much. Unfortunately, I have a broken D435i on my hand. so I'm going to see the results on the bag first. Looking forward to your work.

ZYZ18870128671 commented 1 year ago

You are so great！ I would like to know wheather the open source code include that UAV aviod the ball？Now i have d435i and NUC,how do i run the program to see the effect about detect the tennis ball ？ Thank you！

yinloonga commented 1 year ago

just run the code following the README!

ZYZ18870128671 commented 1 year ago

Do I need to install NCNN？

yinloonga commented 1 year ago

Hey, bro. Have you seen the README???

refer to https://github.com/Tencent/ncnn, copy the contents below ncnn/build/install/ to yolo_fastest_ros/ncnn/

MichaelLu-hku commented 1 year ago

Just follow README is OK. The obstacle avoidance part of this work hasn't been open source yet, but will be open souce in the future. Thank you for your interest.

ZYZ18870128671 commented 1 year ago

NCNN

But In https://github.com/Tencent/ncnn,I don't see the document below ncnn/build/install/

ZYZ18870128671 commented 1 year ago

Just follow README is OK. The obstacle avoidance part of this work hasn't been open source yet, but will be open souce in the future. Thank you for your interest.

Looking forward to your open source！Thank you

MichaelLu-hku commented 1 year ago

NCNN

But In https://github.com/Tencent/ncnn,I don't see the document below ncnn/build/install/

Maybe you have to see how to build this package.

ZYZ18870128671 commented 1 year ago

ok，Thank you very much

ZYZ18870128671 commented 1 year ago

Hello！ I use an intel NUC computer, and I found that it is easy to get stuck when the program is running, especially when tennis balls are detected. Is this a problem of insufficient computing power? Can the program obtain the frame rate of the video stream? Thanks for answering

MichaelLu-hku commented 1 year ago

Thank you for your question.

Try to set the parameter if_gpu false. NUC cpu can run the detection higher than 30hz.

By the way, if you have any question, I suggest that you open a new issued with the corresponding title, which is better for the manage and search for the issues, both for other readers and author. Thank you very much.

ZYZ18870128671 commented 1 year ago

thank you for your reply.

ZYZ18870128671 commented 1 year ago

Hello！I want to know why can't I see the movement track after I add a marker and select the corresponding topic after running the code