Details
Depth Measuring / Image Recognition & Location / Safety & Security Monitoring
Intelligent Driving / Robotics / Object Tracking
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-1.jpg)
OAK is an embedded, high-performance, 3D AI+CV platform consisting of an open-source hardware, firmware, software ecosystem that provides complete and ready-to-use embedded 3D AI + hardware-accelerated computer vision.
The OAK-D-S2 is an compact version of the OAK-D. Its functions, performance, and lenses are the same as OAK-D, but smaller in size and lighter in weight, and can be used in scenes where space is limited and requires high quality. In addition, OAK-D-S2 removes the 5V power port and has a larger depth measurement range than OAK-D in performance.
The OAK-D S2 is the series 2 of the OAK-D, which puts the VPU (Myriad X) directly on the board instead of on the SoM, making the products smaller and lighter. OAK-D-S2 combines depth perception, object detection (neural reasoning) and object tracking, and can help you achieve these functions with a simple and easy-to-use Python API. Onboard three cameras (a 4k/30fps RGB camera and two mono binocular cameras), BNO086 IMU and USB3 port, so you can access OAK via USB on PC, Raspberry Pi and other popular embedded hosts.
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-3.jpg)
DEPTH MEASURING RANGE | 0.2 ~ 35 m |
---|---|
DEPTH CAMERA | 1MP, global shutter |
RGB CAMERA | 12MP |
MAX FRAMERATE | 60fps (RGB), 120fps (Depth) |
AI CHIP | Intel Myriad X 4TOPS computing performance |
VIDEO PLUG FLOW | 4K 30 fps H.265 plug flow |
CONNECTOR | USB3.0 Type-c |
EXPANSION INTERFACES | GPIO, SPI, UART |
NN PLATFORM SUPPORT | all platforms |
POWER CONSUMPTION | 2W ~ 5.5W |
DEVELOPMENT LANGUAGE | Python, C++ |
ECLOSURE | aluminum enclosure |
WEIGHT | 91g |
The OAK Provides Up To 4TOPS Computing Performance
10 Times Of OpenMV, 40 Times Of Raspberry Pi 4B
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-5.jpg)
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-7.jpg)
Parameter | RGB Camera | Binocular Camera |
---|---|---|
Sensor | IMX378 | OV9282 |
FOV | 81°D / 69°H / 55°V | 89°D / 80°H / 55°V |
Resolution | 12MP (4032×3040) | 1MP (1280×800) |
Max FPS | 60 FPS | 120 FPS |
Focal Length | 4.81 | 2.35 |
Aperture | 2.0 | 2.2 |
Focus Range | AF: 8cm ~ ∞ | FF: 19.6cm ~ ∞ |
Lens Size | 1/2.3″ | 1/4″ |
Distortion | < 1.0% | < 1.5% |
Pixel Size | 1.55µm × 1.55µm | 3µm × 3µm |
Shutter | Rolling | Global |
Perceive the real 3D world and give robots a pair of eyes
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-9.jpg)
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-11.png)
Real-time data processing on the chip, low performance requirements on the host
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-13.png)
Edge Detection
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-15.png)
Feature Tracking
Run neural network models directly on the device and support your own models to OAK
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-17.png)
Object Detection
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-19.png)
Semantic Segmentation
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-21.png)
OCR
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-23.png)
Attitude detection
Supports almost all common platforms, programming languages, and neural network frameworks
(The OpenCV AI Kit Official plans to support Java and in progress)
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-25.jpg)
The core chip with 4 TOPS computing performance has built-in functions such as depth measurement, neural network inference acceleration, and OpenCV image processing acceleration. The development environment can be installed with one click, and it features open-source hardware similar to Raspberry Pi, massive application cases and free industrial-grade neural network pre-training models, so you can quickly design your own products.
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-27.jpg)
Lightly Loaded CPU
The processing of artificial intelligence and depth vision on OAK can be completed on the chip without the involvement of the host.
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-29.jpg)
Easy to Develop
It takes less than 30 seconds to run the 3D AI function of OAK.
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-31.jpg)
Open Source Ecology
The hardware and software of OAK are open-source, allowing users to prototype and produce solutions with low risk.
![](/img/devkit/accBoard/OAK-D-S2/OAK-D-S2-details-size.jpg)
![](/img/devkit/LCD/common/res_banner.png)
* Resources for different product may vary, please check the wiki page to confirm the actually provided resources.
Model | OAK-D-Lite | OAK-D | OAK-D-S2 | OAK-D-Pro | OAK-D-POE | OAK-D-S2-POE |
---|---|---|---|---|---|---|
IMU | - | support | ||||
RGB CAMERA | 13MP (4208×3120) | 12MP (4032×3040) | ||||
Binocular Depth Camera | 480P (640×480) | 1MP (1280×800) | ||||
FOV (RGB) | 81°D / 69°H /54°V | 81°D/ 69°H / 55°V | ||||
FOV (Depth) | 86° D/ 73° H/58°V | 89°D/ 80° H/ 55°V | 82°D/ 72° H/ 50°V | 89°D/ 80° H/ 55°V | ||
IR sensitive | - | support | - | |||
Depth technology | Binocular vision | Structured light + binocular vision | Binocular vision | |||
Depth Accuracy | Low | Medium | High | Medium | ||
CONNECTOR | USB Type C | PoE | ||||
waterproof rating | - | IP67 | IP65 |