In the peaceful time about Christmas, a new office resident was born: Armadillo. Although it is primarily a technological demonstrator and a toy, the time devoted to the Armadillo already gets its practical results.
The Armadillo needs to process the image from two cameras in real time for its activity. The Armadillo does not use just a classic chain "Capture video - compress - send - view", but he must be able to evaluate the image and avoid possible obstacles in the way. To achieve this the Armadillo uses a neural network.
We used our video processing experience to create a simple streaming system videos from cameras. The first practical deployment did not take too long:
Neural networks are not yet used in the video streamer. The video streamer is only used for display video from a camera that can provide an image to only one user at a time. However, the experience with Armadillo already shows us the way for further use: counting people, recognition faces, reading license plates, counting cars in the parking lot, monitoring 3D printing, visual quality control and the like.
Do you have a use for a video streamer? Write to us or give us a call!