Computer vision hardware
We prototype the physical capture layer together with the software that triggers, transfers, evaluates, and presents each image in its operating context.
Camera v2, rendered from its CAD model in the production finish.
Capture hardware
v1 proved the enclosed approach in a working kitchen. v2 is the camera we design and build today for Pizza Checker. Both combine the imaging and local-compute layers in one unit, and each installation is adapted to its optics, working distance, and lighting, so this is not a fixed public hardware catalogue.
v1 · Previous
A compact hardware layer combining the camera, local compute, wired connectivity, and service access in one maintainable package.
v2 · Current
The camera we design and build today for the packing station. A 4K autofocus camera, its own computer, and its cooling fit into a sealed aluminium body that mounts flat under the shelf.
Rendered from the CAD model in the production finish.
One operating system
Define the subject, optics, lighting, working distance, and trigger conditions before evaluating a model.
Run the capture service close to the camera and pass a structured image event into the wider application.
Connect results to the relevant order, branch, product, or process so a person can review and act on the evidence.
Reference implementation
The reference connects purpose-built capture hardware with image evaluation and a browser-based operations layer. The links separate the client, named contact, implementation evidence, and application surface.
Capture requirement