Release v3.11.0

Features

• Multi-device pipelines
  • A single Pipeline can now build and run nodes on multiple devices. Add a device with pipeline.addDevice(...) or create a node with an explicit device argument
  • Links between nodes on different devices are automatically relayed through the host
  • If one device disconnects, the other devices continue to run and the lost device reconnects automatically. Use pipeline.setDeviceStateCallback(...) to monitor the device state
  • A frame sync example and a device-to-device relay example are available
• [RVC4] IFE sharing across camera sensors
  • Allows for running up to 5 cameras at once.
  • Different stream configurations are possible:
    • five concurrent cameras at up to 2.3 Gpix/s aggregate processing
    • 2×48MP @ ≤6.5fps + three others (≤20MP*)
    • 1×48MP @ ≤13.5fps + four others (≤20MP*)
  • Requires OS >= 1.44.0
• [Beta] MultiDeviceCalibration node
  • Calculates the extrinsics between multiple devices from synchronized camera frames
  • The result can be saved to JSON or applied with pipeline.setMultiDeviceCalibration, which puts all devices in one shared coordinate system
  • A Python example and a C++ example are available
• [RVC4] ToFStereoFusion updates
  • Available in the Depth node as the TOF_STEREO_FUSION algorithm
  • cropToOverlap crops the depth and confidence outputs to the overlap of the ToF and stereo cameras
• [RVC4] VPP accepts depth input
  • VPP can use aligned depth (for example, from a ToF sensor) instead of disparity
  • A Python example and a C++ example are available
• [Beta] Calibrated panorama stitching
  • setUseInputCalibration(true) creates the panorama from the calibration of the input frames and skips feature matching
  • The output ImgTransformation now describes the virtual camera of the panorama. A new Cylindrical camera model is added for this
• New Odometry message extends TransformData with linear velocity. The Basalt and RTAB-Map VIO nodes now output it
• DetectionParser confidence and IoU thresholds can now be changed at runtime through inputConfig messages
• [RVC4] Added support for priority_hint argument to SNPE backend properties

Bug fixes

• Dynamic calibration
  • Results with a translation-direction change of 15° or more are rejected with an explanatory message
  • The perpendicular optical-axis constraint is no longer applied to modular and FFC boards
• [RVC4] Better ToF filtering to reduce sensor bleeding
• [RVC4] Reduced ToF processing overhead
• getStereoPairs now orders left and right cameras correctly, also for vertical and backward-facing pairs
• ImageAlign now uses the ImgTransformation of the input frames instead of the EEPROM calibration.
• InputQueue.send in Python now releases the GIL, which prevents a deadlock when the device queue is full
• Python exceptions in host nodes keep their original type, so KeyboardInterrupt exits cleanly
• Added missing Python bindings for MessageDemux.setRunOnHost and setOtherModelFormat, which now accepts bytes and NumPy arrays
• [RVC4] Resolved an issue with memory filling up in repeated reconnects to the device if the host dies first
• [RVC4] Resolved an issue in GPU ImgManip where it could reuse memory blocks before the operation would finish, resulting in missordered outputs
• [RVC4]
• [RVC2 PoE] Decreased network stack loads which can result in up to 20% lower CPU usage when streaming 1920x1080@30FPS BGR streams
• [RVC2] Device lifetime improvements should result in better stability

Misc

• Removed the obsolete DepthAI application packages (.dap) for RVC2 standalone mode
• A warning is shown when the Ethernet link speed is 100 Mbit/s or lower
• Release builds now include SPDX and CycloneDX SBOMs
• Message save and load examples are available in Python and C++
• Standerdized logging between RVC2 and RVC4
• Embedded visualizer updated to 4.1.6:
  • Improved stability and performance at large streaming resolutions
  • UX improvements and polishing

[RVC4] Luxonis OS compatibility

Integration tested with Luxonis OS 1.30.1, 1.35.0, and 1.40.0.