Ubuntu22.04编译带CUDA的OpenCV4.7，Python3绑定无法构建求助

在Ubuntu22.04系统上从源码编译支持CUDA11.6和cuDNN8.4的OpenCV4.7，已完成CUDA、cuDNN的安装以及OpenCV的编译安装。C++及GPU功能正常，但多次配置Python3（3.10）相关CMake参数后，仍无法构建Python3绑定。

CMake GUI设置的参数 #

BUILD_opencv_python_bindings_generator = ON
OPENCV_PYTHON_VERSIION = ON
PYTHON3_EXECUTABLE = /usr/bin/python3.10
PYTHON3_INCLUDE_DIR = /usr/include/python3.10
PYTHON3_LIBRARY = /usr/lib/x86_64-linux-gnu/libpython3.10.so
PYTHYON3_NUMPY_INCLUDE_DIRS = /home/USERNAME/.local/lib/python3.10/site-packages/numpy/core/include
PYTHON3_PACKAGES_PATH = /home/USERNAME/.local/lib/python3.10/site-packages

部分配置结果 #

Extra modules:
    Location (extra):            /home/USERNAME/Downloads/opencv/opencv_contrib-4.7.0/modules
    Version control (extra):     unknown

OpenCV modules:
    To be built:                 aruco barcode bgsegm bioinspired calib3d ccalib core cudaarithm
 cudabgsegm cudacodec cudafeatures2d cudafilters cudaimgproc cudalegacy cudaobjdetect cudaoptflow
 cudastereo cudawarping cudev datasets dnn ... 更多模块 ... xphoto
    Disabled:                    freetype
    Disabled by dependency:      -
    Unavailable:                 alphamat ... ovis python2 python3 sfm viz
    Applications:                tests perf_tests examples apps
    Documentation:               NO
    Non-free algorithms:         NO


NVIDIA CUDA:                   YES (ver 11.6, CUFFT CUBLAS)
    NVIDIA GPU arch:             80 86
    NVIDIA PTX archs:

  cuDNN:                         YES (ver 8.4.0)

  Python (for build):            /usr/bin/python3.10

排查建议 #

• 修正参数拼写错误：
  • 将OPENCV_PYTHON_VERSIION改为OPENCV_PYTHON_VERSION
  • 将PYTHYON3_NUMPY_INCLUDE_DIRS改为PYTHON3_NUMPY_INCLUDE_DIRS
• 安装Python3开发依赖：

  sudo apt-get install python3.10-dev python3-numpy
  

• 验证numpy路径准确性：
  执行以下命令获取numpy的正确头文件路径，确保与PYTHON3_NUMPY_INCLUDE_DIRS一致：

  python3.10 -c "import numpy; print(numpy.get_include())"
  

• 清理CMake缓存重新配置：
  删除编译目录下的CMakeCache.txt文件和CMakeFiles文件夹，再重新通过CMake GUI配置参数
• 显式开启Python3模块：
  检查并确保BUILD_opencv_python3参数设置为ON（部分OpenCV版本需要显式开启该参数）
• 补全编译依赖：
  安装可能缺失的基础编译依赖：

  sudo apt-get install build-essential cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
  

内容的提问来源于stack exchange，提问作者Kakazoo