Camera/dfrobot imx662

[DickHorner]

This pull request adds support for the DFRobot IMX662 USB Night Camera combined with the raspberry pi 25 mm lens across all platforms and introduces a camera profile system to enable automatic selection of plate-solving parameters (such as field of view and database) based on the connected camera's USB VID/PID. The engine now dynamically loads the appropriate solver database and field-of-view estimate for supported cameras, improving plate-solving accuracy and flexibility. The web server and related components have been updated to propagate the correct FOV to the frontend and computations.

Camera support and identity propagation:

• Added the DFRobot IMX662 USB Night Camera (VID: 1BCF, PID: 2D4F) to the built-in camera allowlists for Linux, macOS, and Windows (camera_server.c, CameraAllowlist.swift). [1] [2] Fd088409L200R200)
• The camera server now reports both the camera model and its USB VID/PID as camera_model and camera_id in the handshake JSON on all platforms. [1] [2] [3]

Camera profile system and solver improvements:

• Introduced a CameraProfile system in python/evf/camera/profile.py to map known USB cameras to their plate-solving parameters (database file, FOV, error margin), with a default fallback for unknown cameras.
• The engine now selects and applies the correct camera profile at startup, passing the relevant database and FOV parameters to the plate solver. [1] [2]
• Refactored PlateSolver to accept FOV parameters and database filename at initialization, enabling per-camera customization. [1] [2] [3]

Web server and frontend propagation:

• The web server now receives the camera FOV from the engine and uses it for computations such as pixel origin, ensuring frontend alignment with the actual camera profile. [1] [2] [3] [4]

Codebase and initialization changes:

• Refactored engine and main startup flow to ensure the solver is only started after the camera (and thus the camera profile) is known. [1] [2] [3]
• Updated tetra3rs_database_path to accept a filename, supporting multiple solver databases for different camera profiles.

These changes collectively enable automatic, profile-driven plate-solving for different supported cameras, improving accuracy and ease of use.