OpenCV-SharedMemory

OpenCV-SharedMemory

Shares an OpenCV image between two programs using Shared Memory on Windows.
Useful for using with GigE (eg: Basler) cameras if you need the same video in multiple programs but lack multi-cast hardware.
Relevant code can be found in OpenCV-SharedMemory.cpp: server() and receiver().

Run Demo

• Clone respository
• Launch Release/Server.exe
• Launch Release/Receiver.exe
• Press any key inside video window to quit (or wait for video to finish)

Test code

• Clone respository
• Make sure OpenCV is installed. This Visual Studio Solution is configured for VS2017 and OpenCV 3.3.0.
• Copy video.mp4 (Courtesy https://www.youtube.com/watch?v=GiJ6RmjJt8g) to the folder with .cpp files
• Launch OpenCV-SharedMemory.sln
• Configure Project properties to current version of OpenCV.
• Run the program. Pressing "S" launches the server.
• Run another copy of the program. Press "R" to launch the receiver.
• Press any key inside video window to quit (or wait for video to finish)

Summary

• A mutex is created to lock the memory buffer during reads and writes.

• A shared buffer of size Channels * Height * Width bytes is created to store the image.

• Make sure the image isContinuous()

  • If not, use clone() to copy the image to a new cv::Mat where it is continuous.

• LockMutex

• Sender: The image is copied using the pointer to the first byte cv::Mat.ptr() to the shared memory.

• Receiver: A new mat is created to store an image of the same CvType as that in sender (also match step), with its data pointer pointing to the shared memory.

• ReleaseMutex