Visual Intelligence · Orbit Vision Edge
Vision AI that understands camera video in real time
We connect camera input, AI analysis and the screen people use. Orbit Eye describes a single live view with a VLM. Orbit Vision Edge scales that to 32 channels, running segmentation, object detection and a VLM together.
Orbit Vision Edge: 32 channels, one screen
Up to 32 live channels on one screen, each running segmentation (SAM) and object detection. The VLM describes any scene you select.
Know what's happening on site without watching every screen.
See live video from across the site on one screen, with people and objects detected and segmented on every channel. Pick a scene and the VLM describes it. Operators start with what the AI flags instead of watching every feed.
From video input to information on screen
Live streams run through three AI models at once, and the results appear on the matching video in a monitoring view. What gets analyzed and how it's shown depends on your product.
Video input
Connect stream input and decoding, and pass channel previews and processing status to the app.
Concurrent AI
Attach segmentation and detection results to each channel, and the VLM's description to the selected scene.
Monitoring view
Build a screen for reviewing video with results and choosing channels and scenes.
Analyze on site, send only what matters
Video can be analyzed on a device near the cameras, sending mostly events and summaries to servers or control rooms. Useful when raw video must stay on site.
Where it fits
Vision Edge for monitoring large sites. Orbit Eye for devices where people make decisions from a live view.
Monitoring large sites
In facilities, work sites and data centers, see detections on every channel and get descriptions of the scenes that matter.
Devices with a live view
Use Orbit Eye on devices where people need to make sense of the people, objects and surroundings on screen.
A setup for your product
Set the channel count, device, analysis and display needs, then check hardware resources, models and runtime fit.
From camera BSP to Vision AI algorithms
We develop camera BSPs and drivers, customize ISP processing for the sensor and capture conditions, and build video analysis algorithms and monitoring apps. We connect the full flow from camera input to on-screen AI results for your product.
Camera BSP, drivers and ISP customization
We develop camera BSPs and drivers for Android and embedded Linux, and tune ISP processing, including HDR and noise reduction, for the sensor and capture conditions.
Edge Vision AI algorithms and app integration
We optimize video analysis algorithms, including segmentation, object detection and VLMs, for the device and integrate them into apps. Orbit Vision Edge runs video processing and AI analysis for up to 32 channels, together with the monitoring app, in one edge box.
Tell us what you want to solve with video.
Tell us how many cameras you have, where they are, what you need to catch and where results should appear. We'll review the scope with you.