Added edge input feature to fast_line_detector

[tsukada-cs]

Fixed issue #2667 in opencv/opencv_contrib

Added the is_edge option to the detect method in fast_line_detector.

The Fast Line Detector used to use the canny method for edge detection internally, but is_edge=true makes it possible to input the edge image directly. Parameters for the canny method are ignored. The default value is false, so it won't break anyone's code.

This feature will provide users with additional flexibility and versatility in FastLineDetector.

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

• I agree to contribute to the project under Apache 2 License.
• To the best of my knowledge, the proposed patch is not based on a code under GPL or other license that is incompatible with OpenCV
• The PR is proposed to proper branch
• There is reference to original bug report and related work
• There is accuracy test, performance test and test data in opencv_extra repository, if applicable
  Patch to opencv_extra has the same branch name.
• The feature is well documented and sample code can be built with the project CMake

[alalek]

detect() is not a right place for algorithm parameters tuning. See createFastLineDetector()

Can we check some canny parameters to handle that without touching API?
For example, canny_aperture_size == 0

[tsukada-cs]

@alalek
Thanks for the comment.
I tried changing the internal assert statement and then canny_aperture_size = 0, canny_th1=0, canny_th2=0, etc., but it didn't work as expected.
So I removed the is_edge option of detect() and added the input_edge option to createFastLineDetector().

[alalek]

but it didn't work as expected

Try to split this assertion check first.

BTW, There is no sense to define canny_* parameters if we don't use that.

[tsukada-cs]

@alalek

Try to split this assertion check first.

I failed to state the situation correctly.
I split the assertion check and tried canny_appature_size=0, but the canny method was applied as a result (The edges of the input image were detected using the canny method).
Also, the documentation says that the ksize of Sobel() must be 1, 3, 5 or 7, so we are not supposed to enter canny_appature_size = 0 (ksize=0).

BTW, There is no sense to define canny_* parameters if we don't use that.

I agree.
Should we change the implementation?

[alalek]

and tried canny_appature_size=0, but the canny method was applied as a result

You should add this check before Canny() call to bypass it.

[tsukada-cs]

@alalek

Thanks for the kind comments.
I changed the implementation to use the input image as an edge image when canny_aperture_size=0.

[alalek]

Consider using dedicated lines for if/else bodies.
It is necessary to see real picture from "code coverage" line-based reports (example of weekly report).

[sturkmen72]

will be performed

[tsukada-cs]

Thanks for pointing that out. Fixed it.

[tsukada-cs]

@alalek

Consider using dedicated lines for if/else bodies.
It is necessary to see real picture from "code coverage" line-based reports (example of weekly report).

I tried, but is this what you mean?

[alalek]

Well done! Thank you 👍

[tsukada-cs]

Thank you for accepting it. Thanks for all the kind comments.