Release Notes

• What's New

  The ZED SDK 4.2 update introduces compatibility with the new ZED One camera and HDR modes for ZED X One 4k cameras, along with improved memory handling, gravity estimation, and depth mode speed. It also adds a Health Check module for camera status monitoring and enables custom ONNX file inference and semantic mask ingestion. The Fusion module is now compatible with Object Detection, improving data synchronization and positional tracking.

• 4.2.2

  Nov 19, 2024

  SDK

  • Fixed a memory leak that occurred when closing GMSL cameras.
  • Corrected the serialization of `initParameters` in JSON, addressing a previous misinterpretation of the `InputType`.
  • Adjusted Neural depth confidence to prevent significant degradation of point cloud density in specific use cases with very low confidence thresholds. Confidence levels now align with those from version 4.1.X, with an improvement for very low values that previously removed too many reliable floor points in the far range.
  • Resolved an issue where ZED X One timestamps were set incorrectly, causing issues with camera startup.

  Tools

  • Enhanced ZED Explorer’s handling of the ZED X One: it no longer reports invalid calibration and now supports enabling HDR on ZED X One 4K.
  • Enabled ZED Media Server to stream ZED X One 4K in 3600x1800 HDR mode (QHD+).
  • Fixed issues in ZED Sensor Viewer when opening certain non-standard ZED X One configurations.

  Compatibility

  • Added support for Ubuntu 24.04 LTS.
  • Added support for Jetpack 6.1.
  • Updated compatibility with TensorRT 10 in the YOLO samples.

  Wrappers

  • ROS 2
    • Introduced custom object detection from ONNX files in the ROS 2 Wrapper
    • Introduced support for ZED X One camera in the ZED ROS 2 Wrapper as a new ROS 2 Component
    • Full changelog here
    • The zedxone-ros2 repository is now obsolete and will no longer be maintained.

• 4.2.1

  Oct 8, 2024

  SDK

  • Added verbosity in case of self-calibration issue (a specific error code will be added in the future minor update).
  • Removed some GNSS-related verbose.

  Bug fixes

  • Fixed Positional tracking initial position scale for tracking GEN_2.
  • Fixed SVO recording regression leading to oversized files and high CPU usage.

• 4.2.0

  Sep 30, 2024

  SDK

  • Allow the ZED SDK to stream or record SVO2 files at a different framerate than the one of the depth computation. It can be enabled with the InitParameters::async_image_retrieval parameter.
  • Improved initial gravity estimation.
  • Improved the speed of the NEURAL depth mode, especially when running on several cameras at the same time.
  • Add ZED One compatibility with ZED SDK. ZED One can now be created with sl::CameraOne objects and the API is the same than with other ZED cameras. The available modules are Capture, Recording and Streaming. Samples are available too.
  • Release HDR modes for ZED X One 4k cameras for two resolutions : 1920x1200 and 3200x1800. You can enable HDR with the boolean sl::InitParametersOne::enable_hdr, or within ZED Media Server.
  • Add a Health Check module: the status of the camera can now be retrieved with sl::Camera::getHealthStatus. The status will detect and report issues if the camera is down, if the image looks occulted or corrupted, depending on the parameter set in sl::InitParameters::enable_image_validity_check.
  • Fix the mixing of the cameras when using multiple ZED X cameras and a unified driver (on jetpack 6.0).- Improve ZED X One stability.
  • Allow to easily load and save ZED SDK parameters.
  • Allow users to provide an ONNX file directly to the ZED SDK, without further coding. The ZED SDK will take care of running the inference. The Custom Object Detection option if still there for users who needs the flexibility.
  • Allow users to ingest instance semantic masks instead of custom object detections with the Camera::ingestCustomMaskObjects function.
  • Improved the Positional Tracking GEN2 initialization with the IMU data.
  • Safer memory handling of sl::Mat.

  
  Fusion

  • The Fusion is now compatible with the Object Detection module. It can be enabled with Fusion::enableObjectDetection and objects are retrieved with Fusion::retrieveObjects. A fused_objects_group_name at sender level can be set to group the objects from different detection models.
  • Improve the Fusion data synchronization quality when the sender have low or irregular framerates.
  • Region of interest are now correctly applied within the Fusion module.
  • Fix the retrieved position in rare case where the IMU orientation are corrupted.
  • The Fusion module now returns a FusedPositionalTrackingStatus object when retrieving the position. This new object contains information the status of the different modules acting in the fusion of the positional tracking.
  • Update the FUSION_ERROR_CODE to fit the ZED SDK standard: negative values are a warning, positive are an error.

  Tools

  • Add ZED One compatibility with ZED Explorer.
  • Add ZED One compatibility with ZED Sensor Viewer.
  • Fixed ZED Sensor Viewer IMU recording.
  • Improved ZED Depth Viewer opening reliability.

  Wrappers

  • Add ZED One compatibility with Python.
  • Fixed the Fusion implementation of the C# wrapper.