OpenCV Was Born At Intel
OpenCV and Intel have a long history of fruitful collaboration. The OpenCV project was been started at Intel back in 1998, with the first public release in 2000. Even luminaries like Andrew Ng are part of OpenCV’s history. Intel Corporation still funds the core OpenCV development team, many of which are Intel employees, and continues to maintain the build farm. 20 years after OpenCV’s start, Intel shows their dedication to the global open source computer vision community every day.
OpenCV-compatible Devices from Intel
Intel® RealSense™ Cameras and SDK 2.0 are compatible with the OpenCV platform. Compatible Cameras Include the following models:
Intel® RealSense™ Cameras:
- D415, D435, D435i, D455 and L515
Intel® RealSense™ Modules:
- D400, D410, D415, D420, D30 and D450 D435, D435i, D455 and L515
OpenCV examples are available on the Intel RealSense website, located at: https://dev.intelrealsense.com/docs/opencv-wrapper
Intel and OpenCV Optimized Software
Due to OpenCV and Intel’s long history, and Intel’s place at the forefront of modern computing, OpenCV is well-optimized for Intel platforms, in particular:
- A subset of Intel IPP (https://software.intel.com/en-us/ipp) is included by default to OpenCV builds on Intel platforms. This subset of IPP, called IPPICV, is free for all OpenCV users. IPP accelerates a number of kernels in OpenCV, mostly the image processing.
- Intel MKL (https://software.intel.com/en-us/mkl) can optionally be used as well, to get some matrix operations accelerated.
- Intel TBB (https://software.intel.com/en-us/tbb) is another optional component that OpenCV can be linked with to get a better scalability on many-core Intel CPUs.
- Intel Media SDK (https://software.intel.com/en-us/media-sdk) is yet another library that can optionally be used within OpenCV to enable hardware-accelerated video decoding/encoding. Another option to get the same effect is to use gstreamer backend in OpenCV VideoIO module, as long as the proper gstreamer plugins are installed.
Sponsorship
We are thankful to Intel for its technology leadership in edge AI and for sponsoring the OpenCV Spatial AI Competition in 2020, and OpenCV AI Competition 2021. Thanks to their generosity, over 1200 OAK-Ds were shipped to competitors free of charge.
OpenVINO is open-source collection of tools and models to enable highly-efficient AI solutions across a variety of Intel platforms. The Intel distribution of OpenVINO includes a custom optimized build of OpenCV, as well as high-performance deep learning inference engine and myriad other tools. The Intel inference engine can be used as a standalone library or as an accelerator for OpenCV DNN module.
The users of OpenVINO can also enjoy a collection of pre-trained, high-quality, free deep learning models with the Open Model Zoo covering a broad and always-growing range of computer vision applications.
