Small sensor. Plenty to build on.
Big ideas start with the right details. Explore Pico’s working distance, depth output and infrared capture modes. Check the USB and host requirements, then plan your next build.
The values below describe the current evaluation configuration documented in the Pico flyer. Final product specifications and release-software features are subject to change.
Evaluation specifications
| Feature | Current evaluation configuration |
|---|---|
| Working distance | 18 cm to 2 m |
| Depth data | Approximately 2,500 3D points per cloud |
| Depth-only output | 25 point clouds per second |
| Depth and infrared output | 12 point clouds and 12 flood-lit infrared JPEG images per second, alternating captures |
| Reported depth accuracy | 2% at 0.5 m, 4% at 1 m |
| Embedded controller | Raspberry Pi RP2354 |
| Connection | USB for power, control and data |
| Sensor module weight | 4.2 g for the sensor board and optical module, without enclosure |
| Viewer and integration | MkE Connect for Windows, Linux and macOS. Android 8.0 or later with USB OTG. MkE API and supported SDK clients for integration. |
| Calibration | Per-unit factory calibration stored on the sensor |
| Price | To be announced |
Choose the data for your next experiment
- Depth-only mode supplies point clouds. Combined mode alternates point clouds and flood-lit infrared JPEG images. These are separate exposures, not simultaneous RGB-D capture.
- Account for the time between frames when observing moving objects. Pico’s infrared images are not conventional color photographs.
What the numbers describe
- The 4.2 g weight covers the sensor board and optical module without an enclosure. Reported depth accuracy values are evaluation figures, not a guarantee for every scene.
- Point-cloud output rate is not end-to-end application response time. Host processing, recognition and control add their own work.
- The supplied evaluation figures do not establish system power consumption, field of view or end-to-end latency. Check the approved release specifications for those details.
Sensing on Pico. Application logic on your host.
Pico uses MagikEye’s Invertible Light Technology. An infrared dot pattern and a calibrated image-processing pipeline produce measured 3D points.
The Raspberry Pi RP2354 performs embedded depth reconstruction on Pico. A host is the computer or Android device connected to Pico by USB. It supplies USB power, receives sensing data and runs MkE Connect or your application.
The supplied documentation lists MkE Connect for Windows, macOS and Linux, plus Android 8.0 or later with USB OTG. Confirm exact device, OS and USB requirements for the offered release. Viewer support does not by itself confirm that an AI model or SDK will run on the same host.
Object classification, gesture analysis and application control require suitable software on that host. A separate Raspberry Pi computer is optional, and specific Raspberry Pi host support must be confirmed in the release documentation.
Additional capture ratios and event-driven mode changes are planned development directions. They are not included in the documented evaluation modes above.
Put your idea to the test
Have a desktop robot or touchless interaction in mind? Start with working distance, point coverage and infrared input. Test Pico in your own scene and find out what your idea needs next.
Getting startedRequest Early Access to PicoTaking your idea beyond the workbench? Tell MagikEye about your custom product’s sensing and integration requirements.
Inquire about a custom ILT sensor