Can we use this package with ROS (kinetic)

[hsaleem1]

Hi there,

Can anybody please guide me that can I use this package with ROS (kinetic) or is it necessary to install and run this package with ROS 2?

Thanks for the help.

[fgonzalezr1998]

Hi @hsaleem1 , if you want to use this package under ROS Kinetic, you only have to clone the 'melodic' branch. It will work fine. Remember cloning the messagess too

[hsaleem1]

@fgonzalezr1998, thanks for your response.

I have cloned with melodic branch, but getting the error. -- ~~ traversing 1 packages in topological order: -- ~~ - darknet_ros_3d (plain cmake) -- ~~~~~~~~~~~~~ CMake Error at /opt/ros/kinetic/share/catkin/cmake/catkin_workspace.cmake:95 (message): This workspace contains non-catkin packages in it, and catkin cannot build a non-homogeneous workspace without isolation. Try the 'catkin_make_isolated' command instead. Call Stack (most recent call first): CMakeLists.txt:67 (catkin_workspace)

-- Configuring incomplete, errors occurred! See also "/home/user/catkin_ws8/build/CMakeFiles/CMakeOutput.log". See also "/home/user/catkin_ws8/build/CMakeFiles/CMakeError.log". Invoking "cmake" failed

Please guide me.

Thanks.

[fgonzalezr1998]

@hsaleem1 Do you have other packages in your workspace? What I would do is cleaning up the workspace (removing build and devel folders) and then, type catkin_make.

Make sure that you are cloning the repo into 'src' folder with git clone -b melodic https://github.com/IntelligentRoboticsLabs/gb_visual_detection_3d

[hsaleem1]

@fgonzalezr1998, Your help is very much appreciated. Now, I can detect objects with their 3d coordinates.

Can you please guide me that how can I detect custom/other objects with this package?

[fgonzalezr1998]

Hi @hsaleem1 , You can detect the object that you wants. You only have to add them to the 'interested_classes' in the configuration file. However, this package get the 2D bounding boxes provided by darknet_ros (the neural network). It means that you will not be able to detect object that Darknet doesn't detect. If you inspect the darknet_ros package, you will see that there is a '.yaml' file that contains what classes it can detect. If you want to detect custom object that Darknet doesn't detect, you need to re-train the neural network. Here you have very detailed instructions about how to do it.

Hope that I have been helpful

[hsaleem1]

@fgonzalezr1998, your guidance is much appreciated.

I will see the link which you provided.

I am closing this issue.

Thank you once again.