Machine Vision

USB Camera Support

IP cameras are not the only option! The USB ports in the Blazer are there for a reason. Two USB 3.0 Type-A ports, two USB 2.0 ports, and one MIPI CSI port make it possible to connect a special category of cameras commonly referred to as machine vision cameras.

Unfortunately, when it comes to IP cameras on our market, the best available option is typically Chinese 8 MP cameras that in reality deliver no more than about 3.5 effective megapixels. This is disappointing for many applications.

For example, industrial quality control tasks often require at least 20 real megapixels.

For inspecting soldering quality on electronic boards, at least 30 megapixels are required.

Even for automated vehicle inspection systems designed to detect hidden contraband, covering the entire underside of a vehicle in sufficient detail requires at least 12 megapixels.

For automated warehouse systems, the higher the camera resolution, the larger the area it can cover for simultaneous QR code tracking.

The range of applications for machine vision cameras is enormous — from medicine all the way to space technology.

But where can you actually connect a USB camera?
“A regular PC,” you might say — and technically, you would be right. But in practice, it is a highly unreliable solution. First of all, a PC itself — especially one running Windows — is not the best platform for stable, continuous monitoring tasks. Secondly, the USB port in a PC, historically evolving from the old COM port architecture, inherited many of its limitations. It simply was not designed to sustain heavy, high-load processes over long periods of time. Have you ever heard of large-scale USB video surveillance systems? There is a reason they are practically nonexistent: neither the cameras nor the PC can operate reliably for extended periods without constant reboots. Eventually, both the camera and the computer need restarting.

And really, why deal with all the complexity of PC-based systems?

Now there is VideoBlazer. Its somewhat smartphone-like architecture is perfectly suited for matrix cameras and was originally designed specifically for USB and MIPI CSI camera interfaces.

VideoBlazer is capable of both analyzing high-resolution video and encoding it using hardware H.264/H.265 codecs.

For static scenes, it supports resolutions up to 40 MP.
For dynamic scenes, up to 20 MP.

And all of this comes at a relatively low cost.

A full-scale computer, despite all the problems already mentioned, would still cost significantly more. Then multiply that price by two once you add an NVIDIA graphics card for neural network processing. On top of that, you need dedicated installation space, cooling systems, and ongoing maintenance.