Add ANSALPR country format

.claude/settings.local.json

@@ -145,7 +145,11 @@
       "Bash(grep -nE \"^\\\\}$|warmupModel\\\\\\(|#if 0|#endif|ANSONNXCL_legacy_Init\" C:/Projects/CLionProjects/ANSCORE/modules/ANSODEngine/ANSONNXCL.cpp)",
       "Bash(git -C C:/Projects/CLionProjects/ANSCORE restore modules/ANSODEngine/ANSONNXCL.h modules/ANSODEngine/ANSONNXCL.cpp)",
       "Bash(git -C C:/Projects/CLionProjects/ANSCORE status --short modules/)",
-      "Bash(git -C C:/Projects/CLionProjects/ANSCORE diff --stat HEAD modules/ANSODEngine/ANSONNXCL.cpp modules/ANSODEngine/ANSONNXOBB.cpp modules/ANSODEngine/ANSONNXPOSE.cpp modules/ANSODEngine/ANSONNXSEG.cpp modules/ANSODEngine/ANSYOLO12OD.cpp engines/ONNXEngine/ONNXEngine.cpp engines/ONNXEngine/ONNXSAM3.cpp)"
+      "Bash(git -C C:/Projects/CLionProjects/ANSCORE diff --stat HEAD modules/ANSODEngine/ANSONNXCL.cpp modules/ANSODEngine/ANSONNXOBB.cpp modules/ANSODEngine/ANSONNXPOSE.cpp modules/ANSODEngine/ANSONNXSEG.cpp modules/ANSODEngine/ANSYOLO12OD.cpp engines/ONNXEngine/ONNXEngine.cpp engines/ONNXEngine/ONNXSAM3.cpp)",
+      "Bash(awk -F'[:\\(\\)]' '{gsub\\(/^ +| +$/,\"\",$2\\); split\\($2,a,\" \"\\); ms=a[1]; if \\(ms+0 >= 500\\) print $0}')",
+      "Bash(awk -F'[:\\(\\)]' '{split\\($2,a,\" \"\\); ms=a[1]; if \\(ms+0 >= 100\\) print $0}')",
+      "Bash(grep -cE \"^\\\\s*Plate: [^ ]+$\" \"C:/Users/nghia/Downloads/Log4.txt\")",
+      "Bash(grep -oE \"Plate: [^ ]+$\" \"C:/Users/nghia/Downloads/Log5.txt\")"
     ]
   }
 }

modules/ANSLPR/ANSLPR_OCR.cpp

@@ -194,6 +194,12 @@ namespace ANSCENTER
 				else                         _country = Country::JAPAN; // Default for OCR mode
 			}
 
+			// Install the Vietnam-only plate-format defaults now that the
+			// country is known.  Idempotent — clears `_plateFormats` first.
+			// SetCountry() also calls this so a runtime country switch picks
+			// up the right format set without an Initialize / restart.
+			ApplyCountryDefaultPlateFormats();
+
 			// Store the original model zip path — the OCR models (ansocrdec.onnx,
 			// ansocrcls.onnx, ansocrrec.onnx, dict_ch.txt) are bundled inside the
 			// same ALPR model zip, so we reuse it for ANSONNXOCR initialization.
@@ -929,6 +935,140 @@ namespace ANSCENTER
 		return stripped;
 	}
 
+	// ────────────────────────────────────────────────────────────────────
+	//  Country-specific post-processing.  See header for full contract.
+	//  Mirrors the ANSALPR_OD pipeline so OD and OCR engines emit the
+	//  same canonical form for Vietnam / Indonesia / Australia / USA,
+	//  while leaving Japan / China text untouched (kana / kanji would
+	//  be destroyed by an ASCII alnum filter, so we pass-through).
+	// ────────────────────────────────────────────────────────────────────
+	std::string ANSALPR_OCR::AnalyseLicensePlateText(const std::string& ocrText) {
+		if (ocrText.empty()) return ocrText;
+
+		// Country gate — Vietnam-only.  Every other country (Japan, China,
+		// Indonesia, Australia, USA, …) returns the input unchanged so the
+		// upstream StripPlateSeparators output flows through verbatim and
+		// kana / kanji / region-specific punctuation is preserved.
+		if (_country != Country::VIETNAM) {
+			return ocrText;
+		}
+
+		try {
+			// Vietnam: keep only ASCII alphanumerics (StripPlateSeparators
+			// already removed the common separators; this is a defensive
+			// second pass that also catches any stray punctuation the OCR
+			// may have emitted).  Uppercase so the format-match literals
+			// 'M', 'D', 'N', 'G', 'Q', 'T', 'C', 'V' work against the
+			// uppercase plate output.
+			std::string analysedLP;
+			analysedLP.reserve(ocrText.size());
+			for (char c : ocrText) {
+				if (std::isalnum(static_cast<unsigned char>(c))) {
+					analysedLP += c;
+				}
+			}
+			std::transform(analysedLP.begin(), analysedLP.end(), analysedLP.begin(),
+				[](unsigned char ch) { return static_cast<char>(std::toupper(ch)); });
+
+			// Format validation: when `_plateFormats` is configured (which
+			// it is by default for Vietnam via ApplyCountryDefaultPlateFormats),
+			// reject plates that don't match any pattern.  Empty list = no
+			// validation (caller cleared it via SetPlateFormats), in which
+			// case everything is accepted just like ANSALPR_OD does.
+			if (!analysedLP.empty() && !_plateFormats.empty()
+			    && !MatchesPlateFormat(analysedLP)) {
+				ANS_DBG("ALPR_Format",
+					"REJECT plate='%s' (no _plateFormats match)",
+					analysedLP.c_str());
+				return "";
+			}
+			return analysedLP;
+		}
+		catch (const std::exception& e) {
+			this->_logger.LogFatal("ANSALPR_OCR::AnalyseLicensePlateText",
+				e.what(), __FILE__, __LINE__);
+			return "";
+		}
+	}
+
+	bool ANSALPR_OCR::MatchesPlateFormat(const std::string& plate) const {
+		if (_plateFormats.empty()) {
+			return true;  // No formats configured → accept all.
+		}
+		for (const auto& format : _plateFormats) {
+			if (plate.size() != format.size()) continue;
+			bool matches = true;
+			for (size_t i = 0; i < format.size(); ++i) {
+				char f = format[i];
+				char p = plate[i];
+				if (f == 'd') {
+					if (!std::isdigit(static_cast<unsigned char>(p))) { matches = false; break; }
+				}
+				else if (f == 'l') {
+					if (!std::isalpha(static_cast<unsigned char>(p))) { matches = false; break; }
+				}
+				else {
+					// Literal character — exact match required.
+					if (p != f) { matches = false; break; }
+				}
+			}
+			if (matches) return true;
+		}
+		return false;
+	}
+
+	void ANSALPR_OCR::ApplyCountryDefaultPlateFormats() {
+		// Reset first so repeated SetCountry calls don't accumulate stale
+		// patterns from previous countries.
+		_plateFormats.clear();
+
+		switch (_country) {
+			case Country::VIETNAM: {
+				// Same 11 patterns ANSALPR_OD installs for Vietnam.  Keeping
+				// these identical ensures the OD- and OCR-based pipelines
+				// accept exactly the same plate strings in production.
+				//   ddlddddd      — standard car (e.g. 29A12345)
+				//   ddldddd       — short variant
+				//   ddldddddd     — extended variant
+				//   ddllddddd     — two-letter series
+				//   ddMDdddddd    — military
+				//   dddddNGdd     — diplomatic NG
+				//   dddddQTdd     — diplomatic QT
+				//   dddddCVdd     — diplomatic CV
+				//   dddddNNdd     — diplomatic NN
+				//   lldddd        — special two-letter prefix
+				_plateFormats.push_back("ddlddddd");
+				_plateFormats.push_back("ddldddd");
+				_plateFormats.push_back("ddldddddd");
+				_plateFormats.push_back("ddllddddd");
+				_plateFormats.push_back("ddllddddd");
+				_plateFormats.push_back("ddMDdddddd");
+				_plateFormats.push_back("dddddNGdd");
+				_plateFormats.push_back("dddddQTdd");
+				_plateFormats.push_back("dddddCVdd");
+				_plateFormats.push_back("dddddNNdd");
+				_plateFormats.push_back("lldddd");
+				ANS_DBG("ALPR_Format",
+					"ApplyCountryDefaultPlateFormats: VIETNAM (%zu patterns)",
+					_plateFormats.size());
+				break;
+			}
+			case Country::CHINA:
+			case Country::AUSTRALIA:
+			case Country::USA:
+			case Country::INDONESIA:
+			case Country::JAPAN:
+			default:
+				// No default patterns configured for these countries.  The
+				// caller can still install custom formats via SetPlateFormat
+				// / SetPlateFormats; otherwise format validation accepts all.
+				ANS_DBG("ALPR_Format",
+					"ApplyCountryDefaultPlateFormats: country=%d (no default patterns)",
+					static_cast<int>(_country));
+				break;
+		}
+	}
+
 	std::string ANSALPR_OCR::RecoverKanaFromBottomHalf(
 		const cv::Mat& plateROI, int halfH) const
 	{
@@ -1641,6 +1781,13 @@ namespace ANSCENTER
 				combinedText = StripPlateSeparators(combinedText);
 				if (combinedText.empty()) continue;
 
+				// Vietnam-only: alnum filter + uppercase + format validation
+				// against `_plateFormats`.  No-op on every other country.
+				// Same canonical form ANSALPR_OD emits, so OD- and OCR-based
+				// pipelines can be A/B compared on the same input video.
+				combinedText = AnalyseLicensePlateText(combinedText);
+				if (combinedText.empty()) continue;
+
 				Object lprObject = lprOutput[info.origIndex];
 				lprObject.cameraId = cameraId;
 
@@ -1906,6 +2053,13 @@ namespace ANSCENTER
 				combined = StripPlateSeparators(combined);
 				if (combined.empty()) continue;
 
+				// Vietnam-only post-processing (alnum + uppercase + format
+				// validation).  Same gate as the full-frame path so OD and
+				// OCR engines emit identical strings on Vietnam input, and
+				// non-Vietnam countries pass through unchanged.
+				combined = AnalyseLicensePlateText(combined);
+				if (combined.empty()) continue;
+
 				Object out  = pm.lpObj;
 				out.className = combined;      // raw OCR — no ALPRChecker
 				out.cameraId  = cameraId;
@@ -2014,6 +2168,10 @@ namespace ANSCENTER
 		if (_ocrEngine) {
 			_ocrEngine->SetCountry(country);
 		}
+		// Re-install country-default plate formats so a runtime country
+		// switch picks up Vietnam's format list (or clears it when leaving
+		// Vietnam) without needing an Initialize / restart.
+		ApplyCountryDefaultPlateFormats();
 		// Log every SetCountry call so runtime country switches are
 		// visible and we can confirm the update landed on the right
 		// handle.  The recovery + rectification gates read _country on

modules/ANSLPR/ANSLPR_OCR.h

@@ -109,6 +109,41 @@ namespace ANSCENTER
         // --- Original model zip path (reused for ANSONNXOCR initialization) ---
         std::string     _modelZipFilePath;
 
+        // ────────────────────────────────────────────────────────────────
+        //  Country-specific plate-format processing — Vietnam-only.
+        //
+        //  When `_country == VIETNAM`, `AnalyseLicensePlateText` strips
+        //  any remaining non-alphanumeric characters, uppercases the
+        //  result, and validates it against `_plateFormats` (the same
+        //  11-pattern list ANSALPR_OD installs for Vietnam).  Plates
+        //  that don't match any configured format are dropped — the
+        //  return value is "" and the caller skips the plate.
+        //
+        //  Every other country (Japan, China, Indonesia, Australia, USA,
+        //  …) returns the input unchanged so the upstream
+        //  StripPlateSeparators output flows through verbatim, preserving
+        //  kana / kanji / region-specific punctuation.
+        //
+        //  Hooked in after StripPlateSeparators and before the ALPRChecker
+        //  voter, so locked text + skip-when-locked output are emitted in
+        //  the canonical Vietnam form just like ANSALPR_OD's path.
+        // ────────────────────────────────────────────────────────────────
+        [[nodiscard]] std::string AnalyseLicensePlateText(const std::string& ocrText);
+
+        // Format match helper — patterns: 'd' = any digit, 'l' = any
+        // letter, every other character is a literal that must match
+        // exactly.  Returns true when `_plateFormats` is empty (i.e.
+        // no formats configured → no validation, accept all).
+        [[nodiscard]] bool MatchesPlateFormat(const std::string& plate) const;
+
+        // Reset `_plateFormats` to the default list for the current
+        // `_country`.  Called at the end of Initialize (after country.txt
+        // is read) and from SetCountry so a runtime country switch picks
+        // up the right format set on the very next frame.  Currently
+        // populates Vietnam's 11 standard patterns; other countries leave
+        // the list empty (= no validation).
+        void ApplyCountryDefaultPlateFormats();
+
         // --- Colour detection helpers ---
         [[nodiscard]] std::string DetectLPColourDetector(const cv::Mat& lprROI, const std::string& cameraId);
         [[nodiscard]] std::string DetectLPColourCached(const cv::Mat& lprROI, const std::string& cameraId, const std::string& plateText);

tests/ANSLPR-UnitTest/ANSLPR-UnitTest.cpp

@@ -3953,9 +3953,14 @@ int ALPR_OCR_VideoTest() {
         return -1;
     }
 
-    // Step 2: Set country (JAPAN = 5 — adjust to match the dataset if needed)
+    // Step 2: Set country (VIETNAM = 0 — adjust to match the dataset if needed).
-    ANSALPR_SetCountry(&infHandle, 1);
+    // Country codes match country.txt parsing in ANSALPR_OCR::Initialize:
-    std::cout << "Country set to JAPAN" << std::endl;
+    //   0 = VIETNAM, 1 = CHINA, 2 = AUSTRALIA, 3 = USA, 4 = INDONESIA, 5 = JAPAN.
+    // Setting VIETNAM here also activates ANSALPR_OCR's Vietnam-only
+    // post-processing (alnum filter + uppercase + plate-format validation
+    // against the 11 standard Vietnam patterns).
+    ANSALPR_SetCountry(&infHandle, 0);
+    std::cout << "Country set to VIETNAM" << std::endl;
 
     // Step 3: Load engine
     auto engineStart = std::chrono::high_resolution_clock::now();