diff options
Diffstat (limited to 'src/main/java/info/selflearner/ocr_telegram/util')
-rw-r--r-- | src/main/java/info/selflearner/ocr_telegram/util/FileContentReader.java | 22 | ||||
-rw-r--r-- | src/main/java/info/selflearner/ocr_telegram/util/OCR.java | 120 |
2 files changed, 142 insertions, 0 deletions
diff --git a/src/main/java/info/selflearner/ocr_telegram/util/FileContentReader.java b/src/main/java/info/selflearner/ocr_telegram/util/FileContentReader.java new file mode 100644 index 0000000..34123c5 --- /dev/null +++ b/src/main/java/info/selflearner/ocr_telegram/util/FileContentReader.java @@ -0,0 +1,22 @@ +package info.selflearner.ocr_telegram.util; + +import java.io.File; +import java.io.FileReader; +import java.io.IOException; + +public class FileContentReader { + + public static String read(File file) throws IOException { + try(FileReader fileReader = new FileReader(file)) { + StringBuilder content = new StringBuilder(); + int nextChar; + + while ((nextChar = fileReader.read()) != -1) { + content.append((char) nextChar); + } + + return String.valueOf(content); + } + } + +} diff --git a/src/main/java/info/selflearner/ocr_telegram/util/OCR.java b/src/main/java/info/selflearner/ocr_telegram/util/OCR.java new file mode 100644 index 0000000..55e577f --- /dev/null +++ b/src/main/java/info/selflearner/ocr_telegram/util/OCR.java @@ -0,0 +1,120 @@ +package info.selflearner.ocr_telegram.util; + +import info.selflearner.ocr_telegram.exception.PassportNotFoundException; +import info.selflearner.ocr_telegram.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(); + + 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, PassportNotFoundException { + 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 PassportNotFoundException("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; + } + +} + |