Eye(s) of the Jetson Nano

Exploring OpenCV 4.3.0 & AI with camera(s) & C++ on the Jetson Nano B01. This serves as a documentation & reference for the various resources available for the Jetson Nano & Xavier NX development board.

Organisation

The directory is split into 3 sections:

• AI_ML: AI & Machine Learning related project(s)
• cpp_proj: C++ related projects
• scripts: Installation & setup scripts

Camera quick test

$ gst-launch-1.0 nvarguscamerasrc sensor_mode=0 ! 'video/x-raw(memory:NVMM),width=3820, height=2464, framerate=21/1, format=NV12' ! nvvidconv flip-method=0 ! 'video/x-raw,width=960, height=616' ! nvvidconv ! nvegltransform ! nveglglessink -e

Nvidia lib location

• libargus/Argus (For image processing): /usr/src/jetson_multimedia_api/argus/
• Multimedia API: /usr/src/jetson_multimedia_api/
• VisionWorks: /usr/share/visionworks/sources

Scripts

Python & Shell Scripts are found in the scripts directory.

Update the script(s) permissions before running.

$ sudo chmod +x 

Note: Ensure that the CUDA compiler (nvcc) is added to ~/.bashrc

$ echo '# Append CUDA tools path' >> ~/.bashrc
$ echo 'export PATH=/usr/local/cuda-10.2/bin${PATH:+:${PATH}}'>> ~/.bashrc
$ echo 'export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}'>> ~/.bashrc
$ echo 'export CUDA_INCLUDE_DIRS=/usr/local/cuda-10.2/include'>> ~/.bashrc
$ echo 'export CUDA_ROOT_ENV=/usr/local/cuda-10.2'>> ~/.bashrc

CUDA Sample

The CUDA samples have to be compiled, run makeCUDAsample.sh to compile the samples.

The samples would be copied and compiled at ~/cuda_samples.

$ ./makeCUDAsample.sh

Heatsink fan

Only applicable if an external heatsink fan is attached to the Jetson Nano.

Run runFan.sh to turn on the heatsink fan.

$ ./runFan.sh

Testing the camera(s)

Make the runCam.sh script executable:

$ chmod + x runCam.sh

Test the camera connected via runCam.sh:

$ runCam.sh

To test 2 cameras, a -d flag would be added:

$ runCam.sh -d

Additional information on Gstreamer

Custom OpenCV Installation

Note: Compile the OpenCV source before installing ROS!

OpenCV 4.1.1 is installed in Jetson Nano by default, but without CUDA support. Run installOpenCV.sh to compile & install OpenCV 4.3.0 from source.

Install the source at /usr/local.

$ ./installOpenCV.sh

ROS Melodic Installation

Run getROS_melodic.sh to install ROS Melodic.

$ ./getROS_melodic.sh

OpenVSLAM Installation

Ensure that OpenCV is installed before installing openVSLAM. Run getOpenVSLAM.sh to install OpenVSLAM.

$ ./getOpenVSLAM.sh

Compiling the C++ project(s)

More information could be found here.

License

Licensed under the MIT License.