diff options
Diffstat (limited to 'src/main/java/info/selflearner/ocr/util')
| -rw-r--r-- | src/main/java/info/selflearner/ocr/util/Excel.java | 56 | ||||
| -rw-r--r-- | src/main/java/info/selflearner/ocr/util/OCR.java | 167 |
2 files changed, 223 insertions, 0 deletions
diff --git a/src/main/java/info/selflearner/ocr/util/Excel.java b/src/main/java/info/selflearner/ocr/util/Excel.java new file mode 100644 index 0000000..d924aaa --- /dev/null +++ b/src/main/java/info/selflearner/ocr/util/Excel.java @@ -0,0 +1,56 @@ +package info.selflearner.ocr.util; + +import info.selflearner.ocr.model.Passport; +import org.dhatim.fastexcel.Workbook; +import org.dhatim.fastexcel.Worksheet; + +import java.io.IOException; +import java.io.OutputStream; +import java.nio.file.Files; +import java.nio.file.Paths; +import java.time.format.DateTimeFormatter; +import java.util.List; +import java.util.Optional; + +public class Excel { + public static void write(List<Passport> passportList) throws IOException { + String filePath = System.getProperty("java.io.tmpdir") + "/Dokumentlar.xlsx"; + + try (OutputStream os = Files.newOutputStream(Paths.get(filePath)); Workbook wb = new Workbook(os, "OCR web", "1.0")) { + Worksheet worksheet = wb.newWorksheet("Dokumentlar"); + worksheet.width(0, 25); + worksheet.width(1, 15); + + worksheet.range(0, 0, 0, 12).style().bold().set(); + worksheet.value(0, 0, "1-inchi turi"); + worksheet.value(0, 1, "2-inchi turi"); + worksheet.value(0, 2, "Bergan davlat/firma"); + worksheet.value(0, 3, "Familiyasi"); + worksheet.value(0, 4, "Ismi(-lari)"); + worksheet.value(0, 5, "Pasport raqami"); + worksheet.value(0, 6, "Fuqaroligi"); + worksheet.value(0, 7, "Tug'ilgan sanasi"); + worksheet.value(0, 8, "Jinsi"); + worksheet.value(0, 9, "Amal qilish muddati"); + worksheet.value(0, 10, "JSHSHIR"); + worksheet.value(0, 11, "MRZ 1-inchi qatorining ishonchlilik darajasi"); + worksheet.value(0, 12, "MRZ 2-inchi qatorining ishonchlilik darajasi"); + + for (int i = 0; i < passportList.size(); i++) { + worksheet.value(i + 1, 0, String.valueOf(passportList.get(i).type)); + worksheet.value(i + 1, 1, String.valueOf(Optional.ofNullable(passportList.get(i).subtype).orElse(' '))); + worksheet.value(i + 1, 2, passportList.get(i).issuer); + worksheet.value(i + 1, 3, passportList.get(i).surname); + worksheet.value(i + 1, 4, passportList.get(i).givenNames); + worksheet.value(i + 1, 5, passportList.get(i).number); + worksheet.value(i + 1, 6, passportList.get(i).nationality); + worksheet.value(i + 1, 7, DateTimeFormatter.ofPattern("dd.MM.yyyy").format(passportList.get(i).dateOfBirth)); + worksheet.value(i + 1, 8, String.valueOf(passportList.get(i).sex)); + worksheet.value(i + 1, 9, DateTimeFormatter.ofPattern("dd.MM.yyyy").format(passportList.get(i).expirationDate)); + worksheet.value(i + 1, 10, passportList.get(i).personalNumber); + worksheet.value(i + 1, 11, passportList.get(i).mrzLinesConfidence[0] + "%"); + worksheet.value(i + 1, 12, passportList.get(i).mrzLinesConfidence[1] + "%"); + } + } + } +} diff --git a/src/main/java/info/selflearner/ocr/util/OCR.java b/src/main/java/info/selflearner/ocr/util/OCR.java new file mode 100644 index 0000000..1999cd9 --- /dev/null +++ b/src/main/java/info/selflearner/ocr/util/OCR.java @@ -0,0 +1,167 @@ +package info.selflearner.ocr.util; + +import cn.easyproject.easyocr.EasyOCR; +import info.selflearner.ocr.model.Passport; +import net.sourceforge.tess4j.*; +import net.sourceforge.tess4j.util.LoadLibs; +import org.json.JSONArray; +import org.json.JSONObject; + +import javax.imageio.ImageIO; +import java.awt.*; +import java.awt.geom.AffineTransform; +import java.awt.image.AffineTransformOp; +import java.awt.image.BufferedImage; +import java.io.File; +import java.io.IOException; +import java.nio.file.Files; +import java.nio.file.Path; + +public class OCR { + private static final ITesseract TESSERACT = new Tesseract(); + private static final EasyOCR EASY_OCR = new EasyOCR(); + + static { + TESSERACT.setDatapath("src/main/resources/tessdata/"); + TESSERACT.setVariable("user_defined_dpi", "300"); + System.setProperty("java.library.path", LoadLibs.extractTessResources("win32-x86-64").getPath()); + } + + public static String perform(File file, String languages) throws TesseractException { + TESSERACT.setLanguage(languages); + String ocrResult = TESSERACT.doOCR(file); + file.delete(); + return ocrResult; + } + + public static Passport performPassport(File file) throws IOException, InterruptedException { + for (int imageRotateAngle = 0; imageRotateAngle <= 270; imageRotateAngle += 90) { + if (imageRotateAngle == 90) { + System.out.println("Rotating image to 90 degrees..."); + rotateImageVertically(file, true); + } else if (imageRotateAngle == 180) { + System.out.println("Rotating image to 180 degrees..."); + rotateImageHorizontally(file); + } else if (imageRotateAngle == 270) { + System.out.println("Rotating image to 270 degrees..."); + rotateImageVertically(file, false); + } + + Runtime.getRuntime().exec(new String[]{"surya_ocr", file.getPath(), "--results_dir", System.getProperty("java.io.tmpdir"), "--langs=en"}).waitFor(); + JSONArray jsonTextLines = new JSONObject(Files.readAllLines(Path.of(System.getProperty("java.io.tmpdir"), "/", file.getName().split("\\.")[0], "/results.json")).getFirst()) + .getJSONArray(file.getName().split("\\.")[0]).getJSONObject(0).getJSONArray("text_lines"); + int mrzStartingIndex = -1; + + for (int arrayIndex = 0; arrayIndex < jsonTextLines.length() && mrzStartingIndex == -1; arrayIndex++) { + if (jsonTextLines.getJSONObject(arrayIndex).getString("text").indexOf("P<") == 0) { + mrzStartingIndex = arrayIndex; + } + } + + if (mrzStartingIndex != -1) { + System.out.println("MRZ1: " + jsonTextLines.getJSONObject(mrzStartingIndex).getString("text")); + System.out.println("MRZ2: " + jsonTextLines.getJSONObject(mrzStartingIndex + 1).getString("text")); + try { + return new Passport(new String[]{jsonTextLines.getJSONObject(mrzStartingIndex).getString("text"), jsonTextLines.getJSONObject(mrzStartingIndex + 1).getString("text")}, + new int[]{(int) (jsonTextLines.getJSONObject(mrzStartingIndex).getDouble("confidence") * 100), (int) (jsonTextLines.getJSONObject(mrzStartingIndex + 1).getDouble("confidence") * 100)}); + } catch (Exception _) { + + } + } + } + throw new IOException("Passport not found"); + } + + private static void rotateImageVertically(File image, boolean counterClockwise) throws IOException { + BufferedImage bufferedImage = ImageIO.read(image); + BufferedImage output = new BufferedImage(bufferedImage.getHeight(), bufferedImage.getWidth(), bufferedImage.getType()); + + AffineTransform affineTransform = getAffineTransform(counterClockwise, bufferedImage); + + AffineTransformOp affineTransformOp = new AffineTransformOp(affineTransform, AffineTransformOp.TYPE_BILINEAR); + affineTransformOp.filter(bufferedImage, output); + + ImageIO.write(output, "jpg", image); + } + + public static void rotateImageHorizontally(File image) throws IOException { + BufferedImage imageToRotate = ImageIO.read(image); + + int widthOfImage = imageToRotate.getWidth(); + int heightOfImage = imageToRotate.getHeight(); + int typeOfImage = imageToRotate.getType(); + + BufferedImage output = new BufferedImage(widthOfImage, heightOfImage, typeOfImage); + + Graphics2D graphics2D = output.createGraphics(); + + graphics2D.rotate(Math.toRadians(180), widthOfImage / 2.0, heightOfImage / 2.0); + graphics2D.drawImage(imageToRotate, null, 0, 0); + + ImageIO.write(output, "jpg", image); + } + + private static AffineTransform getAffineTransform(boolean counterClockwise, BufferedImage bufferedImage) { + int imageWidth = bufferedImage.getWidth(); + int imageHeight = bufferedImage.getHeight(); + + AffineTransform affineTransform = new AffineTransform(); + affineTransform.rotate((!counterClockwise ? Math.PI : -Math.PI) / 2, imageWidth / 2.0, imageHeight / 2.0); + + double offset = (imageWidth - imageHeight) / 2.0; + if (!counterClockwise) { + affineTransform.translate(offset, offset); + } else { + affineTransform.translate(-offset, -offset); + } + return affineTransform; + } + +// public static Passport performPassport(File file) throws TesseractException, IOException { +// TESSERACT.setLanguage("eng"); +// +// for (int i = 0; i < 4; i++) { +// String ocrResult = null; +// +// switch (i) { +// case 0 -> ocrResult = TESSERACT.doOCR(file); +// case 1 -> ocrResult = EASY_OCR.discern(file); +// case 2 -> ocrResult = EASY_OCR.discernAndAutoCleanImage(file, ImageType.CAPTCHA_INTERFERENCE_LINE); +// } +// +// for (double imageWidthRatio = 0.8; imageWidthRatio <= 1.6; imageWidthRatio += 0.2) { +// for (double imageHeightRatio = 0.8; imageHeightRatio <= 1.6; imageHeightRatio += 0.2) { +// if (i == 3) { +// ocrResult = EASY_OCR.discernAndAutoCleanImage(file, ImageType.CAPTCHA_NORMAL, imageWidthRatio, imageHeightRatio); +// System.out.println("imageWidthRatio: " + imageWidthRatio + ", imageHeightRatio: " + imageHeightRatio); +// } +// +// int mrzStartingIndex = ocrResult.indexOf("P<"); +// +// if (mrzStartingIndex != -1 && mrzStartingIndex + 89 <= ocrResult.length()) { +// String[] mrz = ocrResult.substring(mrzStartingIndex, mrzStartingIndex + 89).split("\n"); +// +// if (mrz[0].length() >= 44 && mrz[1].length() >= 44) { +// System.out.println("mrz: " + Arrays.toString(mrz)); +// Passport passport; +// +// try { +// passport = new Passport(mrz); +// } catch (StringIndexOutOfBoundsException | NumberFormatException exception) { +// continue; +// } +// +// return passport; +// } +// } +// +// if (i != 3) break; +// } +// if (i != 3) break; +// } +// } +// +// throw new IOException("Passport not found"); +// } + +} |