SR ANPR SDK (LPR / ANPR / License Plate Recognition)

Titan-ANPR is a native SDK focused primarily on Windows x64, with a Linux build also available. From an image (photo or frame) it finds plates, runs OCR, and returns country data. 32-bit or other binaries can be supplied on request. C-style API (DLL on Windows, shared library on Linux). The SDK package is not distributed from this website—email [email protected] to request it.

If you are new to ANPR SDKs

An SDK bundles libraries, documentation, and samples so you can add license plate recognition without training models yourself. The package includes a ready-to-run engine (under bin/ with the DLL or .so and dependencies), C/C++ headers, an import library where applicable, and C# and Python samples. Java integration samples (e.g. JNI/JNA) can be provided on request.

What the engine does

• License plate detection with four corner points per plate.
• OCR of plate text with confidence values.
• Per-detection country classification (numeric id, confidence, short label when available).
• Recommended unified API: TitanANPR_Init → TitanANPR_Detect → TitanANPR_Dispose (plus helpers such as TitanANPR_Clear and TitanANPR_GetSelectedEP).

What the SDK package includes

The package we provide on request bundles the production runtime and documentation. Typical layout (see docs/README.md inside the package):

• bin/ — engine and runtime dependencies (on Windows, Titan-ANPR.dll; on Linux, the shared library and dependencies as shipped). Typically includes ONNX Runtime, OpenVINO/TBB, OpenCV, etc., depending on the build.
• include/ — headers titan_anpr.h, titan_license.h.
• lib/ — Titan-ANPR.lib (Windows import library to link the DLL; on Linux as provided by the package).
• docs/ — user guide (README.md) and third-party notices (THIRD_PARTY_NOTICES.txt).
• At the package root: LICENSE-SDK.txt and, where applicable, LICENSE-SAMPLES.txt.
• samples/ — C# (.NET 8) and Python (ctypes) samples. Java samples (JNI/JNA or similar) on request.

Packages are typically Windows x64; a Linux build is also available (e.g. Titan-ANPR-Linux/). 32-bit or other targets can be built on request. To obtain the full SDK, email [email protected].

How to integrate (overview)

1. Extract the ZIP and keep bin/ intact (not only the main binary) so ONNX/OpenCV and the rest load correctly.
2. Windows: install the Microsoft Visual C++ 2015-2022 Redistributable (x64). Linux: install the runtime dependencies listed in your Linux package docs.
3. Licensing: the SDK is commercial; a 30-day trial is available. Obtain the hardware ID (TitanLicense_GetHardwareId or the sample tools), email it to [email protected] and activate with TitanLicense_Activate (C#: LicenseConsole project; Python: example_license.py).
4. In your code: TitanANPR_Init → for each image TitanANPR_Detect (multiple results per call) → TitanANPR_Dispose. Reuse the handle across frames; do not re-init per frame.
5. See docs/README.md inside the package and the README files under samples/csharp/ and samples/python/.

Additional API (same native binary)

Besides the unified pipeline, the native binary (e.g. the DLL on Windows) exposes optional lower-level APIs (PlateDetector_*, OcrDetector_*, CountryDetector_*). See include/titan_anpr.h and docs/README.md for exact signatures.

Common return codes

• 0 — success.
• Negative value: error.
• -10 — often means the license is not valid for runtime use.

Platform requirements

• Windows x64 (typical target for the package).
• Linux: a build is available; see the package documentation (e.g. Titan-ANPR-Linux/).
• 32-bit binaries: available on request (contact us).
• Microsoft Visual C++ 2015-2022 Redistributable (x64) — only for Windows deployments using the MSVC runtime above.
• For C# samples: .NET SDK 8. For Python: 3.10+ (3.11 or 3.12 recommended). Java: not in the default ZIP; samples on request.

The SDK is not downloaded from this site. Email [email protected] with your project or target environment (Windows/Linux, 32/64-bit) and we will reply with next steps.