Feature: Add Libcamera backend

Summary:

This pull request introduces a new backend for the cv::VideoCapture module, providing support for cameras using the Libcamera library on Linux (and potentially other supported platforms).

You can find the complete implementation here.
Detailed usage instructions and sample testing program can be found here.

Motivation / Reason for this change:

Libcamera is the modern standard API for camera control on many recent Linux systems, including popular development boards like the Raspberry Pi. It offers a more robust and feature-rich interface compared to the legacy V4L2 API for many newer camera modules. Adding official Libcamera support to OpenCV will allow users on these platforms to fully utilize their cameras with cv::VideoCapture, replacing or complementing the existing V4L2 backend where applicable.

Implementation Details:

• A new backend implementation cap_libcamera.cpp has been added to the modules/videoio/src/ directory.
• The new backend registers itself under the API ID cv::VideoCaptureAPIs::CAP_LIBCAMERA.
• The implementation utilizes the Libcamera C++ API to enumerate cameras, configure streams, capture frames, and set/get basic properties.
• Reference to LCCV: The design and implementation concepts are significantly inspired by and borrow structure from the excellent LCCV project.

Implemented Features:

• grabFrame, retrieveFrame for frame capture.
• Setting and getting basic frame properties (width, height, frame rate - CAP_PROP_FRAME_WIDTH, CAP_PROP_FRAME_HEIGHT, CAP_PROP_FPS eg).
• Multithreaded Frame Reading and Writing

Issues of multi-cameras

• multiple USB cameras (can recognize and read)
• one USB camera and one MIPI camera (index 0 will be the first recognized camera)
• multiple MIPI cameras (can recognized but only index 0 can read)[WIP]