我无法使用opencv2和接收到的ImportError:libgl.so.1无法打开共享对象文件没有这样的文件或目录

Question

**环境：*ubuntu16.04PythonPython3.7.8 cuda10.0 gcc5.5

命令：

conda activate myenv
python
import cv2

错误： 回溯（最近一次呼叫最后一次）： 文件“”，第1行，在 文件“/home/.conda/envs/myenv/lib/python3.7/site packages/cv2/_init__.py”，第5行，在 从cv2进口* ImportError:libGL.so.1:无法打开共享对象文件：没有此类文件或目录

我试过：

RUN apt install libgl1-mesa-glx -y
    RUN apt-get install 'ffmpeg'\
        'libsm6'\
        'libxext6'  -y

但此版本已安装且为最新版本（libgl1-mesa-glx18.0.5-0ubuntu0~16.04.1）。 然后我试着：

sudo apt-get install --reinstall libgl1-mesa-glx

它不起作用。 最后，我尝试删除该包：

sudo apt-get --purge remove libgl1-mesa-glx

发生另一个错误：

Reading package list... Done
Analyzing the dependency tree of the package
Reading status information... Done
Some packages cannot be installed. If you are using an unstable distribution, this may be
Because the system cannot reach the state you requested. There may be some software you need in this version
The packages have not been created yet or they have been moved out of the Incoming directory.
The following information may be helpful in solving the problem:

The following packages have unmet dependencies:
 libqt5multimedia5-plugins: Dependency: libqgsttools-p1 (>= 5.5.1) but it will not be installed
E: Error, pkgProblemResolver::Resolve failed. This may be due to a software package being required to maintain the status quo.

任何帮助都会非常有帮助。提前感谢。

康达列表：

# packages in environment at /home/lwy/.conda/envs/mmdet1:
#
# Name                    Version                   Build  Channel
_libgcc_mutex             0.1                 conda_forge    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
_openmp_mutex             4.5                       1_gnu    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
addict                    2.3.0                     <pip>
albumentations            0.5.1                     <pip>
appdirs                   1.4.4                     <pip>
asynctest                 0.13.0                    <pip>
attrs                     20.2.0                    <pip>
ca-certificates           2020.6.20            hecda079_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
certifi                   2020.6.20        py37he5f6b98_2    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
chardet                   3.0.4                     <pip>
cityscapesScripts         2.1.7                     <pip>
codecov                   2.1.10                    <pip>
coloredlogs               14.0                      <pip>
coverage                  5.3                       <pip>
cycler                    0.10.0                    <pip>
Cython                    0.29.21                   <pip>
decorator                 4.4.2                     <pip>
flake8                    3.8.4                     <pip>
future                    0.18.2                    <pip>
humanfriendly             8.2                       <pip>
idna                      2.10                      <pip>
imagecorruptions          1.1.0                     <pip>
imageio                   2.9.0                     <pip>
imgaug                    0.4.0                     <pip>
importlib-metadata        2.0.0                     <pip>
iniconfig                 1.1.1                     <pip>
isort                     5.6.4                     <pip>
kiwisolver                1.3.1                     <pip>
kwarray                   0.5.10                    <pip>
ld_impl_linux-64          2.35                 h769bd43_9    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
libffi                    3.2.1                         1    https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
libgcc-ng                 9.3.0               h5dbcf3e_17    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
libgomp                   9.3.0               h5dbcf3e_17    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
libstdcxx-ng              9.3.0               h2ae2ef3_17    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
matplotlib                3.3.2                     <pip>
mccabe                    0.6.1                     <pip>
mmcv                      1.1.6                     <pip>
mmdet                     1.2.0+unknown             <pip>
ncurses                   6.2                  he1b5a44_2    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
networkx                  2.5                       <pip>
numpy                     1.19.4                    <pip>
opencv-python             4.4.0.46                  <pip>
openssl                   1.1.1h               h516909a_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
ordered-set               4.0.2                     <pip>
packaging                 20.4                      <pip>
Pillow                    6.2.2                     <pip>
pip                       20.2.4                     py_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
pluggy                    0.13.1                    <pip>
py                        1.9.0                     <pip>
pycocotools               2.0                       <pip>
pycodestyle               2.6.0                     <pip>
pyflakes                  2.2.0                     <pip>
pyparsing                 2.4.7                     <pip>
pyquaternion              0.9.9                     <pip>
pytest                    6.1.2                     <pip>
pytest-cov                2.10.1                    <pip>
pytest-runner             5.2                       <pip>
python                    3.7.8           h6f2ec95_1_cpython    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
python-dateutil           2.8.1                     <pip>
python_abi                3.7                     1_cp37m    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
PyWavelets                1.1.1                     <pip>
PyYAML                    5.3.1                     <pip>
readline                  8.0                  he28a2e2_2    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
requests                  2.24.0                    <pip>
scikit-image              0.17.2                    <pip>
scipy                     1.5.3                     <pip>
setuptools                49.6.0           py37he5f6b98_2    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
Shapely                   1.7.1                     <pip>
six                       1.15.0                    <pip>
sqlite                    3.33.0               h4cf870e_1    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
tifffile                  2020.10.1                 <pip>
tk                        8.6.10               hed695b0_1    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
toml                      0.10.2                    <pip>
torch                     1.5.0+cu92                <pip>
torchvision               0.6.0+cu92                <pip>
tqdm                      4.51.0                    <pip>
typing                    3.7.4.3                   <pip>
ubelt                     0.9.3                     <pip>
urllib3                   1.25.11                   <pip>
wheel                     0.35.1             pyh9f0ad1d_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
xdoctest                  0.15.0                    <pip>
xz                        5.2.5                h516909a_1    https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
yapf                      0.30.0                    <pip>
zipp                      3.4.0                     <pip>
zlib                      1.2.11                        0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free

Answer 1

通常，这些包是作为系统包安装的，而不仅仅是Python包。因此，很多时候，即使在成功安装像opencv、cmake、dlib这样的包之后，它们也无法工作

最好的安装方法是使用

sudo apt-get install python3-opencv

这是根据官方Opencv Docs.在Ubuntu上成功安装opencv的首选方法

Answer 2

我已经解决了这个问题！ 首先，查找文件：

find /usr -name libgl*

我找到了/usr/lib/x86_64-linux-gnu/mesa/libGL.so.1 然后，我创建了一个软链接：

ln -s /usr/lib/x86_64-linux-gnu/mesa/libGL.so.1 /usr/lib/libGL.so.1

最后，我验证了它的有效性：

python
import cv2