DepthAI Docs
latest
  • DepthAI’s Documentation
    • DepthAI Viewer
    • Example Use Cases
    • Tools & API Examples
    • Ecosystem
      • First steps with DepthAI
        • Installing DepthAI
        • Device setup
        • DepthAI Viewer
        • Default model
        • Next steps
      • Spatial AI
        • 1. Neural inference fused with depth map
        • 2. Semantic depth
        • 3. Stereo neural inference
      • AI / ML / NN
        • Converting model to MyriadX blob
        • Deploying Custom Models
        • Use a Pre-trained OpenVINO model
        • Custom training
        • AI vision tasks
        • Model Performance
      • Depth perception
        • Passive stereo depth perception
        • Active stereo depth perception
        • Time-of-Flight depth perception
      • Computer Vision
        • Run your own CV functions on-device
      • On-device programming
        • Run your own CV functions on-device
        • On-device Pointcloud NN model
        • Using Script node
        • Creating custom NN models
        • Creating custom OpenCL kernel
      • C++/Python API
      • DepthAI SDK
      • Hardware Products
      • FAQs & How-To
        • Why Does DepthAI Exist?
        • What is DepthAI?
        • What is SpatialAI? What is 3D Object Localization?
        • How is DepthAI Used? In What Industries is it Used?
        • What Distinguishes OAK-D From Other Cameras?
        • How Does DepthAI Provide Spatial AI Results?
        • What is the Gen2 Pipeline Builder?
        • What is megaAI?
        • Which Model Should I Order?
        • How hard is it to get DepthAI running from scratch? What Platforms are Supported?
        • Is OAK camera easy to use with Raspberry Pi?
        • Can all the models be used with the Raspberry Pi?
        • Does DepthAI Work on the Nvidia Jetson Series?
        • Can I Use Multiple DepthAI With One Host?
        • Is DepthAI OpenVINO Compatible?
        • Can I Train My Own Models for DepthAI?
        • Do I Need Depth Data to Train My Own Custom Model for DepthAI?
        • If I train my own network, which Neural Operations are supported by DepthAI?
        • What network backbones are supported on DepthAI?
        • My Model Requires Pre-Processing (normalization, for example). How do I do that in DepthAI?
        • Can I Run Multiple Neural Models in Parallel or in Series (or Both)?
        • Can DepthAI do Arbitrary Crop, Resize, Thumbnail, etc.?
        • Can DepthAI Run Custom CV Code? Say CV Code From PyTorch?
        • How do I Integrate DepthAI into Our Product?
        • What Hardware-Accelerated Capabilities Exist in DepthAI and/or megaAI?
        • Are CAD Files Available?
        • How to enable depthai to perceive closer distances
        • What are the Minimum Depths Visible by DepthAI?
        • What Are The Maximum Depths Visible by DepthAI?
        • How Does DepthAI Calculate Disparity Depth?
        • How Do I Calculate Depth from Disparity?
        • How Do I Display Multiple Streams?
        • How do I Synchronize Streams and/or Meta Data (Neural Inference Results)
        • How do I Record (or Encode) Video with DepthAI?
        • What are the Capabilities of the Video Encoder on DepthAI?
        • What Is The Stream Latency?
        • How To Do a Letterboxing (Thumbnailing) on the Color Camera?
        • Is it Possible to Use the RGB Camera and/or the Stereo Pair as a Regular UVC Camera?
        • How Do I Force USB2 Mode?
        • What is “NCS2 Mode”?
        • What Information is Stored on the OAK cameras
        • Dual-Homography vs. Single-Homography Calibration
        • How Do I Get Different Field of View or Lenses for DepthAI and megaAI?
        • What are the Highest Resolutions and Recording FPS Possible with OAK cameras?
        • What are the theoretical maximum transmission rate for USB3 Gen1 and Gen2?
        • What is the best way to get FullHD in good quality?
        • How to run OAK-D as video device
        • How Much Compute Is Available? How Much Neural Compute is Available?
        • How are resources allocated? How do I see allocation?
        • What Auto-Focus Modes Are Supported? Is it Possible to Control Auto-Focus From the Host?
        • What is the Hyperfocal Distance of the Auto-Focus Color Camera?
        • Is it Possible to Control the Exposure and White Balance and Auto-Focus (3A) Settings of the RGB Camera From the Host?
        • Is it possible to control exposure and ISO with separate cameras?
        • Am I able to attach alternate lenses to the camera? What sort of mounting system? S mount? C mount?
        • Can I Power DepthAI Completely from USB?
        • What is the Screw Mount Specification on OAK-1 and OAK-D?
        • How to use DepthAI under VirtualBox
        • What are the SHAVES?
        • How to increase SHAVES parameter?
        • Can I Use DepthAI with the New Raspberry Pi HQ Camera?
        • Can I use DepthAI with Raspberry Pi Zero?
        • How Much Power Does the DepthAI Raspberry Pi CME Consume?
        • A strange noise pattern appears on the OAK-D Lite (RGB), how do I resolve this?
        • How To Unbind and Bind a Device?
        • How Do I Get Shorter or Longer Flexible Flat Cables (FFC)?
        • What are CSS MSS UPA and DSS Returned By meta_d2h?
        • Where are the Github repositories? Is DepthAI Open Source?
        • How Do I Build the C++ API?
        • Can I Use an IMU With DepthAI?
        • Can I Use Microphones with DepthAI?
        • Where are Product Brochures and/or Datasheets?
        • How Much Does OAK Devices Weight?
        • How Can I Cite Luxonis Products in Publications?
        • Where can I find your Logo?
        • How Do I Talk to an Engineer?
      • OAK as a webcam
        • Using UVC
        • Webcam workarounds
      • Troubleshooting
        • DepthAI can’t connect to an OAK camera
        • Reporting firmware crash dump
        • Ping was missed, closing the device connection
        • ImportError: No module named ‘depthai’
        • Why is the Camera Calibration running slow?
        • Permission denied error
        • DepthAI does not show up under /dev/video* like web cameras do. Why?
        • Intermittent Connectivity with Long (2 meter) USB3 Cables
        • Forcing USB2 Communication
        • Output from DepthAI keeps freezing
        • DepthAI freezes after a few frames
        • Udev rules on Linux
        • CTRL-C Is Not Stopping It!
        • Nothing happening when running a DepthAI script
        • “DLL load failed while importing cv2” on Windows
        • python3 depthai_demo.py returns Illegal instruction
        • Neural network blob compiled with incompatible openvino version
        • “realloc(): invalid pointern Aborted” on RPi
        • [error] Attempted to start camera - NOT detected!
        • [error] input tensor exceeds available data range
        • Converting YUV420 to CV2 frame
      • SLAM with OAK
        • On-device SuperPoint for localization and SLAM
        • RAE on-device VIO & SLAM
        • Syncing frames and IMU messages
      • OAK on drones
        • Drone on-device NN-based localization
      • OAK for Education
        • Cortic AI Toolkit
        • Looking for more?
      • Support
        • Requesting support
        • Refunds and returns policy

Tutorials

  • First steps with DepthAI
    • Installing DepthAI
    • Device setup
    • DepthAI Viewer
    • Default model
    • Next steps
  • Spatial AI
    • 1. Neural inference fused with depth map
      • 3D Object Localization
      • 3D Landmark Localization
    • 2. Semantic depth
    • 3. Stereo neural inference
  • AI / ML / NN
    • Converting model to MyriadX blob
      • Local OpenVINO Model Conversion
        • What is OpenVINO?
        • What is the Open Model Zoo?
        • Install OpenVINO
        • Download the face-detection-retail-0004 model
        • Compile the model
        • Run and display the model output
        • Reviewing the flow
      • Model Optimizer
        • FP16 Data Type
        • Mean and Scale parameters
        • Model layout parameter
        • Color order
      • Compile Tool
      • Converting and compiling models
        • 1. Using online blobconverter
        • 2. Using blobconverter package
        • 3. Local compilation
      • Troubleshooting
        • Supported layers
        • Unsupported layer type “layer_type”
        • Incorrect data types
    • Deploying Custom Models
      • 1. Face mask recognition model
        • Converting to .blob
        • Deploying the model
        • End result
      • 2. QR code detector
        • Converting QR code detector to OpenVINO
        • Using Inference Engine (IE) to evaluate the model
        • Decoding QR code detector
        • Testing accuracy degradation due to FP16 quantization
        • Integrating QR code detector into DepthAI
        • QR Code end result
    • Use a Pre-trained OpenVINO model
      • Run DepthAI Default Model
      • Run model
      • Trying Other Models
      • Spatial AI - Augmenting the Model with 3D Position
    • Custom training
      • Overview
      • The Tutorials
      • Supporting Notebooks
    • AI vision tasks
    • Model Performance
  • Depth perception
    • Passive stereo depth perception
    • Active stereo depth perception
    • Time-of-Flight depth perception
  • Computer Vision
    • Run your own CV functions on-device
      • Create a custom model with PyTorch
      • Kornia
  • On-device programming
    • Run your own CV functions on-device
      • Create a custom model with PyTorch
      • Kornia
    • On-device Pointcloud NN model
      • Depth to NN model
      • Optimizing the Pointcloud model
      • On-device Pointcloud demo
    • Using Script node
    • Creating custom NN models
    • Creating custom OpenCL kernel

Modules:

  • C++/Python API
  • DepthAI SDK
  • Hardware Products

Contents

  • FAQs & How-To
    • Why Does DepthAI Exist?
    • What is DepthAI?
    • What is SpatialAI? What is 3D Object Localization?
    • How is DepthAI Used? In What Industries is it Used?
    • What Distinguishes OAK-D From Other Cameras?
    • How Does DepthAI Provide Spatial AI Results?
      • Monocular Neural Inference fused with Stereo Depth
      • What is the Max Stereo Disparity Depth Resolution?
      • Notes
    • What is the Gen2 Pipeline Builder?
    • What is megaAI?
    • Which Model Should I Order?
      • System on Modules
    • How hard is it to get DepthAI running from scratch? What Platforms are Supported?
    • Is OAK camera easy to use with Raspberry Pi?
    • Can all the models be used with the Raspberry Pi?
    • Does DepthAI Work on the Nvidia Jetson Series?
    • Can I Use Multiple DepthAI With One Host?
    • Is DepthAI OpenVINO Compatible?
    • Can I Train My Own Models for DepthAI?
    • Do I Need Depth Data to Train My Own Custom Model for DepthAI?
    • If I train my own network, which Neural Operations are supported by DepthAI?
    • What network backbones are supported on DepthAI?
    • My Model Requires Pre-Processing (normalization, for example). How do I do that in DepthAI?
    • Can I Run Multiple Neural Models in Parallel or in Series (or Both)?
    • Can DepthAI do Arbitrary Crop, Resize, Thumbnail, etc.?
    • Can DepthAI Run Custom CV Code? Say CV Code From PyTorch?
    • How do I Integrate DepthAI into Our Product?
      • Use-Case 1: DepthAI are a co-processor to a processor running Linux, MacOS, or Windows.
      • Use-Case 2: Using DepthAI with a MicroController like ESP32, ATTiny8, etc.
      • Use-Case 3: Using DepthAI as the Only Processor on a Device.
      • Hardware for Each Case:
      • Getting Started with Development
    • What Hardware-Accelerated Capabilities Exist in DepthAI and/or megaAI?
      • Available in DepthAI API Today:
      • On our Roadmap (Most are in development/integration)
      • Pipeline Builder Gen2
    • Are CAD Files Available?
    • How to enable depthai to perceive closer distances
    • What are the Minimum Depths Visible by DepthAI?
      • Monocular Neural Inference fused with Stereo Depth
      • Onboard Camera Minimum Depths
        • Monocular Neural Inference fused with Stereo Depth Mode
        • Stereo Neural Inference Mode
      • Modular Camera Minimum Depths:
        • Monocular Neural Inference fused with Stereo Depth Mode
      • Extended Disparity Depth Mode
      • Left-Right Check Depth Mode
    • What Are The Maximum Depths Visible by DepthAI?
      • Subpixel Disparity Depth Mode
    • How Does DepthAI Calculate Disparity Depth?
      • What Disparity Depth Modes are Supported?
    • How Do I Calculate Depth from Disparity?
    • How Do I Display Multiple Streams?
      • Is It Possible to Have Access to the Raw Stereo Pair Stream on the Host?
    • How do I Synchronize Streams and/or Meta Data (Neural Inference Results)
      • Reducing the Camera Frame Rate
      • Synchronizing on the Host
    • How do I Record (or Encode) Video with DepthAI?
    • What are the Capabilities of the Video Encoder on DepthAI?
    • What Is The Stream Latency?
    • How To Do a Letterboxing (Thumbnailing) on the Color Camera?
    • Is it Possible to Use the RGB Camera and/or the Stereo Pair as a Regular UVC Camera?
    • How Do I Force USB2 Mode?
    • What is “NCS2 Mode”?
    • What Information is Stored on the OAK cameras
    • Dual-Homography vs. Single-Homography Calibration
    • How Do I Get Different Field of View or Lenses for DepthAI and megaAI?
    • What are the Highest Resolutions and Recording FPS Possible with OAK cameras?
    • What are the theoretical maximum transmission rate for USB3 Gen1 and Gen2?
    • What is the best way to get FullHD in good quality?
    • How to run OAK-D as video device
    • How Much Compute Is Available? How Much Neural Compute is Available?
    • How are resources allocated? How do I see allocation?
    • What Auto-Focus Modes Are Supported? Is it Possible to Control Auto-Focus From the Host?
    • What is the Hyperfocal Distance of the Auto-Focus Color Camera?
    • Is it Possible to Control the Exposure and White Balance and Auto-Focus (3A) Settings of the RGB Camera From the Host?
      • Auto-Focus (AF)
      • Exposure (AE)
      • White Balance (AWB)
    • Is it possible to control exposure and ISO with separate cameras?
    • Am I able to attach alternate lenses to the camera? What sort of mounting system? S mount? C mount?
    • Can I Power DepthAI Completely from USB?
    • What is the Screw Mount Specification on OAK-1 and OAK-D?
    • How to use DepthAI under VirtualBox
    • What are the SHAVES?
    • How to increase SHAVES parameter?
    • Can I Use DepthAI with the New Raspberry Pi HQ Camera?
    • Can I use DepthAI with Raspberry Pi Zero?
    • How Much Power Does the DepthAI Raspberry Pi CME Consume?
    • A strange noise pattern appears on the OAK-D Lite (RGB), how do I resolve this?
    • How To Unbind and Bind a Device?
    • How Do I Get Shorter or Longer Flexible Flat Cables (FFC)?
    • What are CSS MSS UPA and DSS Returned By meta_d2h?
    • Where are the Github repositories? Is DepthAI Open Source?
      • Overall
      • Embedded Use Case
    • How Do I Build the C++ API?
    • Can I Use an IMU With DepthAI?
    • Can I Use Microphones with DepthAI?
    • Where are Product Brochures and/or Datasheets?
    • How Much Does OAK Devices Weight?
    • How Can I Cite Luxonis Products in Publications?
    • Where can I find your Logo?
    • How Do I Talk to an Engineer?
  • OAK as a webcam
    • Using UVC
    • Webcam workarounds
  • Troubleshooting
    • DepthAI can’t connect to an OAK camera
    • Reporting firmware crash dump
    • Ping was missed, closing the device connection
    • ImportError: No module named ‘depthai’
    • Why is the Camera Calibration running slow?
    • Permission denied error
    • DepthAI does not show up under /dev/video* like web cameras do. Why?
    • Intermittent Connectivity with Long (2 meter) USB3 Cables
    • Forcing USB2 Communication
    • Output from DepthAI keeps freezing
    • DepthAI freezes after a few frames
    • Udev rules on Linux
    • CTRL-C Is Not Stopping It!
    • Nothing happening when running a DepthAI script
    • “DLL load failed while importing cv2” on Windows
    • python3 depthai_demo.py returns Illegal instruction
    • Neural network blob compiled with incompatible openvino version
    • “realloc(): invalid pointern Aborted” on RPi
    • [error] Attempted to start camera - NOT detected!
    • [error] input tensor exceeds available data range
    • Converting YUV420 to CV2 frame
  • SLAM with OAK
    • On-device SuperPoint for localization and SLAM
    • RAE on-device VIO & SLAM
    • Syncing frames and IMU messages
  • OAK on drones
    • Drone on-device NN-based localization
      • OAK ArduPilot integration
      • Camera vibration
  • OAK for Education
    • Cortic AI Toolkit
    • Looking for more?
  • Support
    • Requesting support
      • DepthAI issue
      • Connectivity issue
      • Hardware issue
      • Image Quality issue
      • Calibration issue
      • Converting NN model issue
    • Refunds and returns policy
DepthAI Docs
  • »
  • Search

Revision 58324b8d.

Built with Sphinx using a theme provided by Read the Docs.