v7.4.0

What's Changed

• Minor bug fixes to data processing by @mattalvarado in #254
• Increase the number of large buffers for incoming images to 32 by @mattalvarado in #255
• Add better rx buffer handling by @mattalvarado in #256
• Pypi release fixes by @mattalvarado in #257
• Add multi-channel synchronizer by @mattalvarado in #253
• Bump version to v7.4 by @mattalvarado in #258

Full Changelog: 7.3.0...7.4.0

v7.3.0

What's Changed

• Remove 20.04 from the release build pipeline by @mattalvarado in #237
• Add debug info by @mattalvarado in #240
• Support compressed images by @mattalvarado in #241
• Declare nlohmann_json public dependency by @ethanmusser in #243
• Remove unreachable code by @ethanmusser in #244
• Complete python examples by @mattalvarado in #246
• Improve legacy channel errors by @mattalvarado in #247
• Support the ST25 camera by @mattalvarado in #249
• Make some utils more consistent with their usage and fix names by @dougbalish1 in #250
• Return the error code from a invalid sendto rather than throw by @mattalvarado in #251
• Improved buffer management by @mattalvarado in #252
• Add aux depth computation utilites by @mattalvarado in #248

Full Changelog: 7.2.0...7.3.0

v7.2.0

LibMultiSense v7.2.0 Release

New release of the LibMultiSene with support for the updated API. The updated API is enabled via the -DBUILD_LEGACY_API=0FF CMake build flag. The legacy API is still natively supported, and is largely unmodified outside of the isolation of the wire protocol.

All artifacts in this release are built with using the updated LibMultiSense API.

Supported Operating Systems Include:

• Windows 11
• MacOS Sequoia
• Ubuntu 22.04, 24.04

The updated API includes the following major changes

New Features

• Fix bug when stopping Aux streams (#235)
• Rename power_supply temperature to CPU temperature (#233)
• Fixes to cmake installs (#232)
• Add frame metadata (#230)
• Add support for stereo histograms (#229)
• Stop streams before testing the MTU (#224)
• Fix bug with the connect on startup logic when specifying an interface (#223)
• Tweak imagecalutil so usage is printed if all req arguments are not supplied (#222)
• Don't stamp images/imu data with 0ns ptp times (#218)

Prebuilt Binaries

The attached release artifacts include PIP wheels and zipped installations for the various supported operating systems. These were built with support for JSON serialization.

Python PIP

PIP wheels (*.whl) files are included for the supported operating system targets. Wheels can be installed using the command pip install <path-to-wheel>

Each wheel use the following naming convention:

libmultisense-<version>-cp<python-version>-cp<python-version>-<operating-system>-<operating-system-architecture>.whl

For example a file named libmultisense-7.2.0-cp310-cp310-linux_x86_64.whl could be installed on x86 linux systems running python 3.10.

The versions of Python supported by wheels are driven by the default version of Python installed on the GitHub Actions runners for Ubuntu, MacOS, and Windows. Please open an issue if the version of Python you require is missing.

Python wheels are also hosted on on PyPi https://pypi.org/project/libmultisense, and can be installed via the following command:

pip install libmultisense

Install Directories

The installation directories for each build of LibMultiSense are included for each of the supported operating systems. When possible, Carnegie Robotics recommends installing LibMultiSense directly on the target system, rather than copy the output of these installation directories onto a target system. These are included for completeness, and for applications where it might not be possible to build LibMultiSense with the necessary.

The contents of each install directory output uses the following naming convention:

libmultisense-<operating-system>.zip

v7.1.0

LibMultiSense v7.1.0 Release

New release of the LibMultiSene with support for the updated API. The updated API is enabled via the -DBUILD_LEGACY_API=0FF CMake build flag. The legacy API is still natively supported, and is largely unmodified outside of the isolation of the wire protocol.

All artifacts in this release are built with using the updated LibMultiSense API.

Supported Operating Systems Include:

• Windows 11
• MacOS Sequoia
• Ubuntu 20.04, 22.04, 24.04

The updated API includes the following major changes

New Features

• Add the ChangeIpUtiity for the new API (#215)
• fix spelling errors (#213)
• Use cibuildwheel to generate PyPi releases (#210)
• Restrict network config to valid values (#211)
• Invalidate the main image config white balance controls for new cams (#207)
• Add ImageCalUtility for the new API (#208, #214)
• Generate PyPi release artifacts (#205)

Prebuilt Binaries

The attached release artifacts include PIP wheels and zipped installations for the various supported operating systems. These were built with support for JSON serialization.

Python PIP

PIP wheels (*.whl) files are included for the supported operating system targets. Wheels can be installed using the command pip install <path-to-wheel>

Each wheel use the following naming convention:

libmultisense-<version>-cp<python-version>-cp<python-version>-<operating-system>-<operating-system-architecture>.whl

For example a file named libmultisense-7.1.0-cp310-cp310-linux_x86_64.whl could be installed on x86 linux systems running python 3.10.

The versions of Python supported by wheels are driven by the default version of Python installed on the GitHub Actions runners for Ubuntu, MacOS, and Windows. Please open an issue if the version of Python you require is missing.

Python wheels are also hosted on on PyPi https://pypi.org/project/libmultisense, and can be installed via the following command:

pip install libmultisese

Install Directories

The installation directories for each build of LibMultiSense are included for each of the supported operating systems. When possible, Carnegie Robotics recommends installing LibMultiSense directly on the target system, rather than copy the output of these installation directories onto a target system. These are included for completeness, and for applications where it might not be possible to build LibMultiSense with the necessary.

The contents of each install directory output uses the following naming convention:

libmultisense-<operating-system>.zip

v7.0.0

LibMultiSense v7.0.0 Release

Initial release of the LibMultiSene with support for the updated API. The updated API is enabled via the -DBUILD_LEGACY_API=0FF CMake build flag. The legacy API is still natively supported, and is largely unmodified outside of the isolation of the wire protocol.

All artifacts in this release are built with using the updated LibMultiSense API.

The updated API includes the following major changes

New Features

• Isolate the Legacy LibMultiSense wire protocol into a standalone library
• Refactor of LibMultiSense to use C++17 features (mainly smart pointers)
• Add optional Python bindings via pybind11 along with generate PIP wheels for installation
• Add optional JSON serialization via nlohmann-json
• Add optional basic OpenCV support
• Consolidate Configuration, Device Information, and Status Information into respective objects
• Add cross-platform build support for Ubuntu, MacOS, and Windows
• Remove support for EOL MultiSense cameras including the MultiSense SL
• Updated version of Utilities in both C++ and Python

Prebuilt Binaries

The attached release artifacts include PIP wheels and zipped installations for the various supported operating systems. These were built with support for JSON serialization.

Python PIP

PIP wheels (*.whl) files are included for the supported operating system targets. Wheels can be installed using the command pip install <path-to-wheel>

Each wheel use the following naming convention:

libmultisense-<version>-cp<python-version>-cp<python-version>-<operating-system>-<operating-system-architecture>.whl

For example a file named libmultisense-7.0.0-cp310-cp310-linux_x86_64.whl could be installed on x86 linux systems running python 3.10.

The versions of Python supported by wheels are driven by the default version of Python installed on the GitHub Actions runners for Ubuntu, MacOS, and Windows. Please open an issue if the version of Python you require is missing.

Install Directories

The installation directories for each build of LibMultiSense are included for each of the supported operating systems. When possible, Carnegie Robotics recommends installing LibMultiSense directly on the target system, rather than copy the output of these installation directories onto a target system. These are included for completeness, and for applications where it might not be possible to build LibMultiSense with the necessary.

The contents of each install directory output uses the following naming convention:

libmultisense-<operating-system>.zip

v6.2.0

• General bugfixes and performance improvements
• Remove c++11 features which broke builds for older compilers
• Add example features and utilities:
  • Color depth image utility
  • Feature detector
  • PTP status messages
  • New firmware update utility
  • Auto-MTU negotiation
• Add support for the following cameras:
  • KS21i
  • ST25
  • KS21 silver
• Add optional UDP rate limit
• Remove obsolete utilities

v6.1.0

• Add enums for the Tau2 ST21 imager
• Fix memory leak in the DepthCache

v6.0.0

• Add explicit aux camera controls
• Remove secondary exposure controls
• Remove support for directed streams
• Remove support for crop mode
• Remove SensorCalibration

v4.3.0

• Fix LED inversion serialization
• Improve internal buffer handling
• Add camera metadata to the SaveImageUtility
• Improve performance of the PointCloudUtility
• Add support for image sharpening controls
• Add support for rolling shutter LED flashing
• Allow binding to a specific network interface
• Fix ty scaling in the RectifiedFocalLengthUtility
• Add in controls for remote head camera synchronization
• Add support for generating pkg-config.pc files
• Resolve #107

v4.2.1

• Add utility to update a cameras rectified focal length
• Add support for ground surface output supported on newer 5.X firmware releases
• Add support for remote head cameras
• Add support for on-camera AprilTag detection outputs on newer 5.X firmware releases
• Fix support for building under Clang/MSVC
• Improve status message handling for disconnected cameras
• Fix network time sync to handle large time jumps
• Add support for inverting LEDs for newer 5.X firmware