Open Data Cam Is a Computer Vision Traffic Counter That You Can Build Yourself with an NVIDIA Jetson TX2

Despite its important in urban planning, traffic counting is a tricky and labor-intensive prospect. In many cases, people are paid to stand on a street corner and literally count cars. In our modern world of technology, nobody should have to perform such a monotonous task—and now they don’t, thanks to moovel lab’s Open Data Cam.

Open Data Cam is a device built on NVIDIA’s Jetson TX2 that monitors a video feed and interprets what it sees. To illustrate the capability of Open Data Cam, the team created a traffic counting program. A user can place the camera to point towards traffic, and access an interface from their phone to setup the data collection parameters. Multiple parts of the video frame can be used to count cars, trucks, and pedestrians, and then the collected data can be exported for analysis.

The system works with computer vision backed by a library of objects that it can detect. In the traffic counting example, those objects are vehicles and people. But, Open Data Cam has potential for a lot of other uses, like capturing wildlife or tracking patrons of a retail store. You’re really only limited by your imagination and what can be recognized by computer vision.

If you want to build Open Data Cam yourself, moovel lab has got you covered. After you’ve purchased the off-the-shelf hardware, you can use the provided printable templates to cut out the enclosure and assemble the device. Then simply install the open source software, and you’re ready to start counting cars.