Get started with object detection using a Jetson Orin Nano
Introduction
Set up your Jetson Orin Nano
Launch the image classification Docker container
Detect objects in video and images
Review
Next Steps

  Log an issue

  Fork and edit

  Discuss on Discord

Detect objects in video and images

Note

The first time you run a new model it will take some time to start up. Subsequent launches will be faster.

Change directory to prepare for image detection:

    

        
        
            cd build/aarch64/bin

DetectNet for object detection on a video feed

To start DetectNet using your camera:

    

        
        
            ./detectnet csi://0

If you want to increase or decrease the sensitivity, add a threshold parameter. By default it is set to 0.5.

    

        
        
            ./detectnet csi://0 --threshold=0.25

Image Alt Text:showing object labels in real time

DetectNet for object detection on image files

The following command is using one of the sample images provided in the docker image.

You can replace “peds_0.jpg” with your own image, and change output.jpg to an appropriate name.

    

        
        
            # --network=ssd-mobilenet-v2 is an optional parameter
./detectnet --network=ssd-mobilenet-v2 images/peds_0.jpg images/test/output.jpg

If needed, use the docker cp command examples to add your own images.

Image Alt Text:cat lying in between a keyboard and monitor

Image Alt Text:a dog on a bed

Image Alt Text:Robert Wolff in the Innovation Coffee studio

You are now able to perform object detection using the camera or your own images copied into the Docker container.

Back
Next