System information (version)

• OpenCV => 4.3
• Operating System / Platform => Linux
• Compiler => Python3.6.10

Detailed description

Please add the Stitcher::composePanorama(InputArrayOfArrays images, OutputArray pano) function to the python bindings.
This would allow reusing the estimated transforms multiple times (same, fixed, camera setup).
Currently you can only call composePanorama(OutputArray pano), so you have to estimate the transform every time (sometimes features in other image pairs can not be detected, so manual stitching is needed if I can not reuse the old transform).

Found the same request as a question here: https://answers.opencv.org/question/145886

Steps to reproduce

does not apply

Issue submission checklist

• [x ] I report the issue, it's not a question
• [x ] I checked the problem with documentation, FAQ, open issues,
• [x ] I updated to latest OpenCV version and the issue is still there
• [? ] There is reproducer code and related data files: videos, images, onnx, etc