【已解决】安装cv2时Building wheel for opencv-python终端卡死

1 问题背景

环境：

• Ubuntu20.04
• Python3.6

现象：终端输出类似下面的命令，并卡住不动

Building wheel for opencv-python (PEP 517) ... 
1

2 问题探索

首先，OpenCV较旧的Python版本不需要进行滚轮编译，因此造成编译在

Building wheel for opencv-python (PEP 517)
1

处停止。为了防止因为pip版本较旧引起干扰，首先运行

pip install --upgrade pip
1

更新pip版本，接着再次安装

pip install opencv-python -i https://pypi.tuna.tsinghua.edu.cn/simple
1

接着大概率能解决Building wheel for opencv-python (PEP 517)，但是遇到新的现象，终端运行到下面指令时卡住

building wheel for opencv(pyproject.toml)  ...
1

我们希望看看后端到底是什么情况，用--verbose运行安装命令跟踪安装过程，也可以检查安装的百分比

[ 48%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/src/layers/fast_convolution/fast_convolution.avx2.cpp.o
[ 48%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/layers/layers_common.avx2.cpp.o
[ 48%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/int8layers/layers_common.avx2.cpp.o
[ 48%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/layers/layers_common.avx512_skx.cpp.o
[ 48%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/int8layers/layers_common.avx512_skx.cpp.o
[ 48%] Linking CXX static library ../../lib/libopencv_dnn.a
[ 48%] Built target opencv_dnn
[ 48%] Processing OpenCL kernels (features2d)
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/affine_feature.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/agast.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/agast_score.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/akaze.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/bagofwords.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/blobdetector.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/brisk.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/draw.cpp.o
[ 49%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/dynamic.cpp.o
[ 50%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/evaluation.cpp.o
[ 50%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/fast.cpp.o
[ 50%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/fast_score.cpp.o
[ 50%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/feature2d.cpp.o
[ 50%] Building CXX object modules/features2d/CMakeFiles/opencv_features2d.dir/src/gftt.cpp.o
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

发现后端确实一直在运行，由于OpenCV体积较大，编译时间较长，导致不显示安装进度的化，前端出现假死现象，静静等待即可。

3 问题解决

更新pip版本

pip install --upgrade pip
1

显示安装进程

pip install opencv-python -i https://pypi.tuna.tsinghua.edu.cn/simple --verbose
1

等待安装完成

文章最后，推荐一个研究深度学习都会碰到的bug解析
【已解决】探究CUDA out of memory背后原因，如何释放GPU显存？

4 告别Bug

本文收录于《告别Bug》专栏，该专栏记录人工智能领域中各类Bug以备复查，文章形式为：问题背景 + 问题探索 + 问题解决，订阅专栏+关注博主后可通过下方名片联系我进入AI技术交流群帮忙解决问题

---