javacv | Java interface to OpenCV FFmpeg | Computer Vision library

The program tries to get the value of the first parameter but it does not check if the latter is provided.

You need to call the program with a videoFileName as a first parameter.

There seems to be no in-built function, but this could be just done like this:

Yes, you can just exclude what you don't need, like this:

Thanks to Samuel Audet, i figured out that The pixel format of the frames returned by FFmpeg is RGBA, not ARGB.

We can easily change the FFMpegFrameGrabber's pixel format by calling setPixelFormat() before start().

There is a list of every avaible pixel format here: list

In my case, this is the final solution:

After some research I've found that this is easily done using the "platform plugin of Gradle JavaCPP":

http://bytedeco.org/builds/

https://github.com/bytedeco/gradle-javacpp#the-platform-plugin

Using said plugin you can specify the target platform with : gradle -PjavacppPlatform=linux-armhf and only the needed platform dependencies will be used.

You also can look at the example:

https://github.com/bytedeco/gradle-javacpp/tree/master/samples/javacv-demo

I've tried building the example and a uber-jar of "only" around 100Mb is generated.

I think you could also add some exclusions to eliminate unneeded libs and further reduce the size of the jar.

I have okhttp3 and logging-interceptor 4.7.2 versions. I have java.lang.NoSuchMethodError: okhttp3.internal.Platform.log

It is a Dependency issue, You are using :

1. okhttp-4.7.2
2. logging-interceptor-4.7.2
3. okhttp-digest-2.0

And Runtime dependency of okhttp-digest-2.0 is okhttp-4.0.1

Maven Repo: https://mvnrepository.com/artifact/com.burgstaller/okhttp-digest/2.0

So now you have two different versions of same library

1.okhttp-4.0.1 from okhttp-digest-2.0

2.okhttp-4.7.2

There was a change in log function(Which is what error is saying NoSuchMethodError:okhttp3.internal.Platform.log) we can see in the image

Solutions:

1. You Can Update version of okhttp-digest which can work with okhttp-4.7.2
2. You could exclude the okhttp dependency from okhttl-digest if you're not using the features relying on it.

The solution is simple and also really silly IMHO.

In src/test/resources/application.yml, previously it was

It was a licensing issue that made impossible to include ffmpeg to

This question java.lang.UnsatisfiedLinkError: org.opencv.core.Mat.n_Mat(III)J? answers why you are getting that UnsatisfiedLinkError; the library for OpenCV is not loaded by your Java VM and therefore the relevant native calls cannot be made.

This question How to use opencv in using Gradle? should help you with configuring OpenCV using Gradle.

Per your comment question "I have a requirement for the application to work on Windows 10, Linux and MacOS." I don't know how to solve that using gradle, however in Java I would inject a different system property per o/s and use that in a static block in your class to load your library: