手动识别和处理表单- Java

进口java.io.File;进口java.io.IOException;进口java.nio.file.Files;进口java.nio.file.Paths;进口java.util。*;进口leadtools。*;进口leadtools.codecs。*;进口leadtools.document.writer。*;进口leadtools.forms.common。*;进口leadtools.forms.processing。*;进口leadtools.forms.recognition。*;进口leadtools.forms.recognition.ocr。*;进口leadtools.ocr。*;

公共静态无效main (String [] args)｛新ManuallyRecognizeAndProcessAFormTutorial () .run (args);｝

私人无效运行(String [] args)｛试一试｛Platform.setLibPath (“C: \ \ LEADTOOLS22 \ \ Bin \ \ CDLL \ \ x64”）;Platform.loadLibrary (LTLibrary.LEADTOOLS);Platform.loadLibrary (LTLibrary.CODECS);Platform.loadLibrary (LTLibrary.OCR);SetLicense ();InitFormsEngines ();CreateMasterFormAttributes ();RecognizeForm (formToRecognize);｝抓(异常交货){System.err.println (ex.getMessage ());ex.printStackTrace ();｝最后｛如果(ocrEngine ! =零) {ocrEngine.shutdown ();ocrEngine.dispose ();｝如果(编解码器! =零）codecs.dispose ();｝｝

私人OcrEngine =零；私人RasterCodecs codecs =零；私人FormRecognitionEngine =零；私人FormProcessingEngine processingEngine =零；私人字符串masterformDir =“C: \ \ LEADTOOLS22 \ \资源\ \图片\ \ \ \ \ \ OCR MasterForm集形式”；私人字符串formToRecognize =“C:\\LEADTOOLS22\\资源\\图像\\表单\\表单待识别\\OCR\\ w9_ocr_filling .tif”；

私人无效InitFormsEngines () {试一试｛System.out.println (“初始化引擎……”）;编解码器=新RasterCodecs ();.getLoad .getRasterizeDocument codecs.getOptions () () () .setResolution (300);recognitionEngine =新FormRecognitionEngine ();processingEngine =新FormProcessingEngine ();ocrEngine = OcrEngineManager.createEngine(ocrengintype . lead);ocrEngine.startup(编解码器,新DocumentWriter (),零，零）;OcrObjectsManager =新OcrObjectsManager (ocrEngine);recognitionEngine.getObjectsManagers阀门()(ocrObjectsManager);processingEngine.setOcrEngine (ocrEngine);System.out.println (“引擎初始化成功。”）;｝抓(例外ex) {System.err.println (ex.getMessage ());ex.printStackTrace ();｝｝

私人无效CreateMasterFormAttributes () {System.out.println (“处理MasterForm……”）;//遍历文件夹中所有的。tif图像文件masterformOcrFolder =新文件(masterformDir);为masterformFile: masterformOcrFolder.listFiles((filename) -> . listfiles.endsWith .toLowerCase filename.toString () () (“.tif”))) {字符串masterformfileName = masterformFile.toString();字符串masterformName = masterformfileName.substring(masterformfileName.lastIndexOf (File.separator) + 1,masterformfileName.lastIndexOf (“。”））;RasterImage image =零；试一试｛图像=编解码器。load(masterformfileName, 0, CodecsLoadByteOrder.)Bgr_or_gray, 1， -1);masterFormAttributes = recognitionEngine。createMasterForm (masterformName UUID.randomUUID (),零）;//获取每个页面的表单属性为（int我= 0;我< image.getPageCount ();我+ +){image.setPage (i + 1);recognitionEngine。addMasterFormPage (masterFormAttributes形象,零）;｝recognitionEngine.closeMasterForm (masterFormAttributes);//将属性写入文件Files.write(路径。get (masterformName +“。斌”)、masterFormAttributes.getData ());｝抓(例外ex) {System.err.println (ex.getMessage ());ex.printStackTrace ();｝最后｛如果(图片! =零）image.dispose ();如果(编解码器! =零）codecs.dispose ();｝｝System.out.println (“主窗体处理完成。”）;｝

私人无效识别表单(String unidentifiedform) {System.out.println (“认识形式……”）;字符串projectDirectory = System.getProperty(“user.dir”）;RasterImage image =零；试一试｛图像=编解码器。load(unidentifiedform, 0, CodecsLoadByteOrder.)Bgr_or_gray, 1， -1);filledFormAttributes = recognitionEngine.createForm(零）;为（int我= 0;我< image.getPageCount ();我+ +){image.setPage (i + 1);recognitionEngine。addFormPage (filledFormAttributes形象,零）;｝recognitionEngine.closeForm (filledFormAttributes);布尔发现=假；文件夹=新文件(projectDirectory);File[] folderBinFiles = folder.listFiles((filename) -> filename. tostring ().toLowerCase().endsWith(“。斌”））;为(文件masterformBinFile: folderBinFiles){masterformBinFileName = masterformBinFile.toString();masterformName = masterformBinFileName.substring(masterformBinFileName.lastIndexOf (File.separator) + 1,masterformBinFileName.lastIndexOf (“。”））;masterformXmlFilename = masterformDir +"\\" + masterformName + ". xml”;processingEngine.loadFields (masterformXmlFilename);//比较填充表单属性和主表单属性masterformAttributes =新FormRecognitionAttributes ();masterformAttributes.setData (Files.readAllBytes (masterformBinFile.toPath ()));FormRecognitionResult = recognitionEngine.compareForm(masterformAttributes, filledFormAttributes，零）;//如果信心>=80，那么我们已经找到了主窗体如果(recogntionresult . getconfidence () >= 80) {列表 alignments =新ArrayList < PageAlignment > ();为(PageRecognitionResult pageResult: recognitionResult.getPageResults())alignments.add (pageResult.getAlignment ());System.out.println ("此表格已被公认为"+ masterformName": \ n "）;ProcessForm(图片、平面图);发现=真正的；打破；｝｝如果(!)System.out.println (“这种形式无法被识别。”）;｝抓(例外ex) {System.err.println (ex.getMessage ());ex.printStackTrace ();｝最后｛如果(图片! =零）image.dispose ();｝｝

私人无效ProcessForm(RasterImage图像，列表< pagealign > alignments) {字符串resultsMessage =＂＂；//在所有页面的所有表单字段中显示处理后的文本值processingEngine。过程(图片、排列);为(FormPage: processingEngine.getPages())为(FormField: formPage)如果(formField ! =零）//为了便于演示，假设所有字段都是文本resultsMessage += String.format(“%s at %s = %s\n”，formField.getName (),.toString formField.getBounds () (),((TextFormFieldResult) formField.getResult ()) .getText ());如果(resultsMessage = =零|| resultsMessage ==＂＂）System.out.println (“没有字段被处理。”）;其他的System.out.println (resultsMessage);｝