A minor improvement regarding the CUDA support is required. Currently, haveCUDA() is equivalent to using the HAVE_CUDA define. However, existing code combines both checks, which implies that it is intended to do a runtime instead of a compile time check. For example:

Instead, we should use a runtime check in haveCUDA(), similar to haveVulkan(). Doing so, we can use CUDA by default whenever available by setting OPENCV_DNN_BACKEND_DEFAULT = DNN_BACKEND_CUDA, cf. #25112. The feature is implemented by simply checking for a valid CUDA device using cudaGetDevice() and saving its return value such that this function is only called once.

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 another license that is incompatible with OpenCV
• The PR is proposed to the proper branch
• There is a reference to the 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