Zum Inhalt springen

WSAL21086 – OAK-D-POE/en

Aus JW Tech Wiki
Version vom 6. September 2026, 00:30 Uhr von JWTechPublisher (Diskussion | Beiträge) (Freigegebene technische Dokumentation veröffentlicht; Text-only und noindex)
(Unterschied) ← Nächstältere Version | Aktuelle Version (Unterschied) | Nächstjüngere Version → (Unterschied)

Documentation notice: This technical documentation was independently rewritten from an archived supplier source. Check technical details and commands against the product version and manufacturer documentation before use.

Summary

The OAK-D-POE is a compact AI vision system featuring an Intel Myriad X VPU, a 4K RGB camera, and binocular depth cameras. It is powered and connected via a single Cat5e cable using Power over Ethernet (802.3af), features an IP67-rated enclosure, and is suitable for demanding environments.

Related products at JW Tech

Technical overview

  • Processes image data through an integrated Intel Myriad X processor with approximately 4 TOPS of computing performance.
  • Features a 12 MP color camera (IMX378, rolling shutter, autofocus from 8 cm) plus two 1 MP monochrome depth cameras (OV9282, global shutter, fixed focus from 50 cm or 19.6 cm).
  • Supports H.265 video streaming at 4K resolution and 30 fps.
  • Average power consumption is approximately 2.5 W.
  • The PoE connection requires an 802.3af-compatible switch or router; the cable range extends up to 100 m.
  • The IP67 enclosure protects against dust and temporary water exposure.
  • Development is possible in Python and C++ via the depthai library.
  • Depth is reliably measured in the 0.7 to 12 m range; theoretically up to 35 m is possible, with an error of about 6–10% at 30 m.
  • Suitable for applications such as intelligent driving, robotics, security monitoring, and computer-vision education.

Commands and examples – verify before use: 

git clone https://gitee.com/oakchina/depthai.git
git clone https://gitee.com/oakchina/depthai-python.git
echo 'SUBSYSTEM=="usb", ATTRS{idVendor}=="03e7", MODE="0666"' | sudo tee /etc/udev/rules.d/80-movidius.rules
sudo udevadm control --reload-rules && sudo udevadm trigger
python3 -m pip install -r depthai/requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
python3 depthai/depthai_demo.py

Notes and version constraints

  • The exact compatibility requirements for the PoE switch (e.g., minimum current, voltage profile) are not fully specified in the source; consult the manufacturer's datasheet.
  • For first-time setup, the device must be on the same LAN as the host computer, otherwise the software will not detect it.
  • Some installation commands reference mirror sites (e.g., gitee.com); accessibility of these sources may be regionally restricted.
  • 3D positioning accuracy depends heavily on object texture; accuracy decreases on smooth surfaces.
  • When using Jetson, do not run the dependency script directly to avoid OpenCV overwrites; configure the Jetson environment properly first.
  • The stated depth values (0.7–12 m, theoretically 35 m) are indicative; actual measurement range depends on lighting and object characteristics.

Source and revision record

  • OAK - Waveshare Wiki
  • Source revision: 89738
  • Captured: 2026-09-03T02:34:38+00:00
  • SHA-256: 6318c6207197ca6799e3e1a214cdb380b41ec9aae9d441fed88c1defbf95f1a8