Installing OpenCV w/ GPU in Windows: Difference between revisions

Revision as of 13:17, 21 September 2021

The OpenCV installing via pip does not include GPU support and will significantly reduce performance. This decreases FPS from around 30FPS to under 2FPS in darknet. To enable the GPU, you must compile opencv with CUDA support.

Create a DEV folder at c:\dev\
Install Cuda from here
Install Visual Studio from here
Install cmake
Download OpenCV from here and extract to dev folder as opencv
Download OpenCV Contrib from here (note that versions should match) and extract to dev folder as opencv-contrib
Run cmake-gui and point the sources folder to c:\dev\opencv\sources
Set the build directory to c:\dev\opencv\build
Adjust the following flags:
- CMAKE_BUILD_TYPE=RELEASE \
- CMAKE_INSTALL_PREFIX=/usr/local \
- WITH_TBB=ON \
- ENABLE_FAST_MATH=1 \
- CUDA_FAST_MATH=1 \
- WITH_CUBLAS=1 \
- WITH_CUDA=ON \
- BUILD_opencv_cudacodec=OFF \
- WITH_CUDNN=ON \
- OPENCV_DNN_CUDA=ON \
- CUDA_ARCH_BIN=7.5 \
- WITH_V4L=ON \
- WITH_QT=OFF \
- WITH_OPENGL=ON \
- WITH_GSTREAMER=ON \
- OPENCV_GENERATE_PKGCONFIG=ON \
- OPENCV_PC_FILE_NAME=opencv.pc \
- OPENCV_ENABLE_NONFREE=ON \
- OPENCV_PYTHON3_INSTALL_PATH=~/.virtualenvs/cv/lib/python3.8/site-packages \
- PYTHON_EXECUTABLE=~/.virtualenvs/cv/bin/python \
- OPENCV_EXTRA_MODULES_PATH=~/Downloads/opencv_contrib-4.5.2/modules \
- INSTALL_PYTHON_EXAMPLES=OFF \
- INSTALL_C_EXAMPLES=OFF \
- BUILD_EXAMPLES=OFF
Click Configure and then Generate. When prompted, select the correct version of Visual Studio
From the build directory, open the INSTALL.vcxproj project in Visual Studio

@@ Line 3: / Line 3: @@
 # Create a DEV folder at c:\dev\
 # Install Cuda from [https://developer.nvidia.com/cuda-downloads here]
-# Install Visual Studio
+# Install Visual Studio from [https://visualstudio.microsoft.com/downloads/ here]
 # Install cmake
 # Download OpenCV from [https://opencv.org/releases/ here] and extract to dev folder as opencv