Initial .NET scaffold: Core, Console, WPF projects

Introduced solution structure for AIFotoONLUS migration to .NET. Added Core library with YOLO-based detection/recognition engine using OpenCvSharp, Console batch runner, and WPF demo frontend with MVVM. Implemented model loading, directory processing, progress reporting, and preferences. Added README with build/run instructions.

src/README.md

@@ -0,0 +1,23 @@
+# AIFotoONLUS (.NET) — scaffold
+
+This workspace contains the initial scaffold for migrating the YOLO-based number recognition Python script to .NET 10.
+
+Projects:
+- `AIFotoONLUS.Core` — core inference library (OpenCvSharp)
+- `AIFotoONLUS.Console` — console batch runner
+- `AIFotoONLUS.WPF` — simple WPF demo frontend
+
+Quick build & run (Windows):
+1. Ensure .NET 10 SDK is installed.
+2. From `src` run:
+   - `dotnet restore`
+   - `dotnet build`
+3. Copy the `models/` folder (contains detection/recognition .cfg and .weights) to the `src` output folder or run the apps from repository root so `models/` is accessible.
+
+Console example:
+  dotnet run --project src\AIFotoONLUS.Console -- -d "..\images\onlus" -c result.csv
+
+Notes:
+- Core engine currently loads Darknet models and scaffolds the detection/recognition pipeline.
+- Parsing of network outputs (YOLO postprocessing) and recognition logic will be implemented next.
+- WPF app is a minimal demo (code-behind). We'll replace with MVVM as we iterate.