Video input in real time object detection with 10 lines of code

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hi in the video of real time object detection with 10 lines of code I want to use a prerecorded video of mp4 format and want to input that instead of using the camera in detect.py code can someone please explain how can I do that?

2

Hi @rahulsinghrox701, I have been working on adding this functionality to the dev branch of jetson-inference. You can test it out by first installing the dev branch:

git clone --recursive --branch dev https://github.com/dusty-nv/jetson-inference jetson-inference-dev
cd jetson-inference-dev
mkdir build
cd build
cmake ../
make -j4
sudo make install

Then, first test that you can view your video ok:

cd jetson-inference-dev/build/aarch64/bin
python3 video-viewer.py my_video.mp4

Then to run object detection on your video:

python3 detectnet.py my_video.mp4
3

hey @dusty_nv i can confirm that your code is working with pre recorded video.Detection is slightly worse on car video but the fps is high.thank you