Unable to load shared library OpenCvSharpExtern, Ubuntu 18.04

[alexdr488]

Summary of your issue

When I run any application (with code that runs on Windows) with OpenCvSharp 4, I get an error "System.DllNotFoundException: Unable to load shared library "OpenCvSharpExtern". I tried to put different versions of packages, but the problem remained. How do I run a project on Ubuntu? I have not found a working solution for this operating system.

Environment

C#, Ubuntu 18.04.4-x64 Installed packages (nuget): OpenCvSharp4 OpenCvSharp4.runtime.ubuntu.18.04-x64

Example code:

class Program
    {
        static void Main(string[] args)
        {
            Mat src = new Mat("1.jpg", ImreadModes.Grayscale);
            Mat dst = new Mat();

            Cv2.Canny(src, dst, 50, 200);
            using (new Window("src image", src)) 
            using (new Window("dst image", dst)) 
            {
                Cv2.WaitKey();
            }
        }
    }

Output:

Unhandled exception. System.TypeInitializationException: The type initializer for 'OpenCvSharp.NativeMethods' threw an exception.
 ---> System.DllNotFoundException: Unable to load shared library 'OpenCvSharpExtern' or one of its dependencies. In order to help diagnose loading problems, consider setting the LD_DEBUG environment variable: libOpenCvSharpExtern: cannot open shared object file: No such file or directory
   at OpenCvSharp.NativeMethods.redirectError(CvErrorCallback errCallback, IntPtr userdata, IntPtr& prevUserdata)
   at OpenCvSharp.ExceptionHandler.RegisterExceptionCallback()
   at OpenCvSharp.NativeMethods.LoadLibraries(IEnumerable`1 additionalPaths)
   at OpenCvSharp.NativeMethods..cctor()
   --- End of inner exception stack trace ---

[shimat]

Please try using ldd to check dependencies of the libOpenCvSharpExtern.so and refer to the CI script for necessary packages. https://github.com/shimat/opencvsharp/blob/master/.github/workflows/ubuntu18.yml#L25

[SuryaKumarSE]

Hello @shimat

I have faced the same issue which was mentioned above “Unable to load shared library OpenCvSharpExtern, Ubuntu 18.04” and I have tried adding all the set of dependencies which you have referenced in the link below: https://github.com/shimat/opencvsharp/blob/master/.github/workflows/ubuntu18.yml#L25

But when I try checking the dependency of the libOpenCvSharpExtern.so using ldd command, I can able to still find some of the dependency is being missed as shown below:

libavcodec.so.57 => not found libavformat.so.57 => not found libavutil.so.55 => not found libswscale.so.4 => not found libjpeg.so.8 => not found

Note: I have installed and checking this assemblies in docker container which is being hosted in Google Kubernetes.

Environment : C#, Ubuntu 18.04.4-x64 Installed packages (nuget): OpenCvSharp4 OpenCvSharp4.runtime.ubuntu.18.04-x64

Kindly let me know your suggestions for the same.

[ffrankozz]

Hello @SuryaKumarSE Any chance of solution so far for your problem? I have bumped into the same problem after dockerizing and trying to find out a solution. It seems @shimat is not in a position to reply.

[makejack]

add export LD_LIBRARY_PATH="/lib:/usr/lib:/usr/local/lib" environment variables and restart the system

[arialblack14]

@makejack that did not work for me. Any update on this?

[marianmelnychuk]

@arialblack14 I have the same problem. Did you find solution?

[arialblack14]

@marianmelnychuk No, sorry :(

[makejack]

@arialblack14 I set the environment variable after installing OpenCv. you can use the "ldd" command to detect the dependency of libOpenCvSharpExtern.so, use the "apt-file" command to search for “not found” dependency , install the searched dependency .

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[nietras]

It would be really great if the Linux so file had all dependencies linked in statically like for windows or perhaps have it as a separate nuget package (OpenCvSharp4.runtime.ubuntu.18.04-x64-static) so it can be used without having to install lots of deps, hence so it works like on windows.

[mburumaxwell]

@nietras ,

I figured out a way to get the necessary bits into a docker image without having to build the entire image, with long C++ builds, every time. Maybe this can help you or anyone else on this thread. Unfortunately, I don't think it is possible to place them in a NuGet package like for Windows because it would require about 800MiB in size for ubuntu 18.04

In the Dockerfile add the following lines

RUN apt-get update && apt-get install -y --no-install-recommends libexpat1
COPY --from=shimat/ubuntu20-dotnet6-opencv4.5.3 /usr/lib /usr/lib
COPY --from=shimat/ubuntu20-dotnet6-opencv4.5.3 /lib /lib

A complete Dockerfile (based on the one generated by Visual Studio) would look like this:

#See https://aka.ms/containerfastmode to understand how Visual Studio uses this Dockerfile to build your images for faster debugging.

FROM mcr.microsoft.com/dotnet/aspnet:6.0-bullseye-slim AS base
WORKDIR /app
EXPOSE 80
EXPOSE 443
RUN apt-get update && apt-get install -y --no-install-recommends libexpat1
COPY --from=ubuntu20-dotnet6-opencv4.5.3 /usr/lib /usr/lib
COPY --from=ubuntu20-dotnet6-opencv4.5.3 /lib /lib

FROM mcr.microsoft.com/dotnet/sdk:6.0-bullseye-slim AS build
WORKDIR /src
COPY ["MyProject/MyProject.csproj", "src/MyProject/"]
RUN dotnet restore "src/MyProject/MyProject.csproj"
COPY . .
WORKDIR "/src/MyProject"
RUN dotnet build "MyProject.csproj" -c Release -o /app/build

FROM build AS publish
RUN dotnet publish "MyProject.csproj" -c Release -o /app/publish

FROM base AS final
WORKDIR /app
COPY --from=publish /app/publish .
ENTRYPOINT ["dotnet", "MyProject.dll"]

In my case, I built the 4.6 image into a private container registry hence

RUN apt-get update && apt-get install -y --no-install-recommends libexpat1
COPY --from=myregistry.azurecr.io/opencvsharp:4.6.0 /usr/lib /usr/lib
COPY --from=myregistry.azurecr.io/opencvsharp:4.6.0 /lib /lib

@shimat , maybe you could have this in the documentation as an alternative to having to build an image from source?

[shimat]

I actually run my company's systems that way. However, I do not guarantee the maintenance of my container images (shimat/ubuntu20...) in Docker Hub (I may remove them suddenly one day),

I recommend that users themselves register OpenCvSharp container images with a container registry such as Docker Hub, Amazon ECR, Google Container Registry, etc.