How does REBOTNIX solve these problems?

RB-Compress reduces energy consumption by compressing image based content faster, more securely and more cost-effectively over terrestrial or orbital communication paths to low-power terminals. RB-Compress is able to recognise and segment objects individually. Highest priority is given to objects that are contextual and relevant. These objects are automatically recognised by an AI model and the objects that are important for the contextual information are cut out of the image. The individual objects can then be individually compressed.

How it works?

Each visual object can then be given its own compression layer. Objects such as text and others that are relevant to the transmission of information, such as traffic signs or license plates, are encoded at a higher bit rate than objects that are not relevant, such as small objects in the background. Unlike a motion vector method, the object segmentation single encoding method achieves a higher compression rate.

In the default mode, the encoder works with a GPU or CPU. In the highest AI mode, a GPU is mandatory for encoding, but for decoding, we still can use CPU only for realtime applications.

What economic added value do we achieve with this?

• Low energy consumption for the transmission of visual information
• More capacity on devices with little storage capacity
• Higher runtimes for data acquisition
• Higher quality in the recordings
• Works on low-power computers between 5-15 watt
• Reduces the load on cloud and internet infrastructures as well as mobile networks