Using CUDA sample codes in the jetson inference utils

I am seeking guidance how I can use the jetson util tools like the gstCamera that use CUDA and GStreamer in my Qt app for jetson nano. If I copy the header files into my Qt app will they work or are there build steps I need to take.

Please guide me on this, I’m a beginner. Thanks

Please check if you can apply above flags to your QT app. If you can apply it, it should work to run CUDA in the app.

If you are looking for a solution of running deep learning inference, you may consider to use DeepStream SDK. The package can be installed through SDKManager.

I’m not sure if you are looking for something like this

NOTE: It’s customized for (x86-64 architecture

I would recommend building/installing jetson-utils as per it’s CMakeLists.txt, as opposed to try and copying the headers into your own project. Also, if you build/install the top-level jetson-inference project, jetson-utils will be built/installed as a submodule.

Then see this example CMakeLists.txt as an example of an external application that links to jetson-inference / jetson-utils libraries using cmake find_package() command.

Thanks dusty_nv for the recommendation. I will try it. But I have one question. Is there an equivalent of find_package() in qmake or how would I link the libraries

Thanks too Collince, Very similar to what I need, let me look at it further and see if it can be useful to me.

QMake doesn’t appear to be compatible with the .cmake config files that get installed under /usr/local/share, however when you do the “sudo make install” step of jetson-inference, it installs the libraries to /usr/local/lib and the headers under /usr/local/include (both paths are included by the compiler/linker by default I believe). So you should be able to add jetson-inference and jetson-utils to you QMake LIBS variable.