WindowsFormsTest/OCRTest/Program.cs

using OpenCvSharp;
using PaddleOCRSharp;
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Drawing;
using System.IO;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using Tesseract;

namespace OCRTest
{
    internal class Program
    {
        static void Main(string[] args)
        {
            Console.WriteLine("请选择测试模式：");
            Console.WriteLine("1. 测试OCR识别引擎");
            Console.WriteLine("2. 创建训练项目示例");
            Console.WriteLine("3. 预处理训练图片");
            Console.WriteLine("4. 验证训练数据集");
            Console.WriteLine("5. 分析字符频率");
            Console.Write("\n请输入选项（1-5）：");

            string choice = Console.ReadLine();

            switch (choice)
            {
                case "1":
                    TestOCREngine();
                    break;
                case "2":
                    CreateTrainingProjectExample();
                    break;
                case "3":
                    PreprocessTrainingImagesExample();
                    break;
                case "4":
                    ValidateDatasetExample();
                    break;
                case "5":
                    AnalyzeCharactersExample();
                    break;
                default:
                    Console.WriteLine("无效选项，默认执行OCR识别测试");
                    TestOCREngine();
                    break;
            }
        }

        /// <summary>
        /// 测试新的OCR引擎
        /// </summary>
        static void TestOCREngine()
        {
            Console.WriteLine("=== OCR 识别测试 ===\n");

            // 测试图片路径（请修改为实际路径）
            string imagePath = "D:/work/WindowsFormsTest/TestImage/wechat_2025-10-16_143143_308.png";

            if (!File.Exists(imagePath))
            {
                Console.WriteLine($"错误：找不到图片文件 {imagePath}");
                return;
            }

            try
            {
                // 方式1：使用单例引擎（推荐，性能最优）
                Console.WriteLine("【方式1】使用单例引擎识别...");
                var ocrEngine = OCREngine.Instance;
                
                Stopwatch sw = Stopwatch.StartNew();
                string result1 = ocrEngine.RecognizeText(imagePath);
                sw.Stop();
                
                Console.WriteLine($"识别结果：{result1}");
                Console.WriteLine($"耗时：{sw.ElapsedMilliseconds} ms\n");

                // 方式2：指定语言识别（中文+英文）
                Console.WriteLine("【方式2】中英文混合识别...");
                var ocrEngineChi = OCREngine.GetInstance("chi_sim+eng");
                
                sw.Restart();
                string result2 = ocrEngineChi.RecognizeText(imagePath);
                sw.Stop();
                
                Console.WriteLine($"识别结果：{result2}");
                Console.WriteLine($"耗时：{sw.ElapsedMilliseconds} ms\n");

                // 方式3：使用自定义配置
                Console.WriteLine("【方式3】使用自定义配置识别...");
                var config = new OCRConfig
                {
                    Language = "eng",
                    EngineMode = EngineMode.LstmOnly, // 仅使用LSTM，速度更快
                    Whitelist = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789 .,!?-",
                    EnablePreprocessing = true
                };
                
                var ocrEngineCustom = OCREngine.GetInstance(config);
                
                sw.Restart();
                string result3 = ocrEngineCustom.RecognizeText(imagePath);
                sw.Stop();
                
                Console.WriteLine($"识别结果：{result3}");
                Console.WriteLine($"耗时：{sw.ElapsedMilliseconds} ms\n");

                // 方式4：批量识别
                Console.WriteLine("【方式4】批量识别测试...");
                string[] imagePaths = new string[] { imagePath }; // 可以添加多个图片路径
                
                sw.Restart();
                var results = ocrEngine.RecognizeBatch(imagePaths);
                sw.Stop();
                
                for (int i = 0; i < results.Count; i++)
                {
                    Console.WriteLine($"图片{i + 1}：{results[i]}");
                }
                Console.WriteLine($"批量识别总耗时：{sw.ElapsedMilliseconds} ms\n");

                // 方式5：从Bitmap识别
                Console.WriteLine("【方式5】从Bitmap识别...");
                using (var bitmap = new Bitmap(imagePath))
                {
                    sw.Restart();
                    string result5 = ocrEngine.RecognizeFromBitmap(bitmap);
                    sw.Stop();
                    
                    Console.WriteLine($"识别结果：{result5}");
                    Console.WriteLine($"耗时：{sw.ElapsedMilliseconds} ms\n");
                }

                Console.WriteLine("测试完成！按任意键退出...");
                Console.ReadKey();
            }
            catch (Exception ex)
            {
                Console.WriteLine($"发生错误：{ex.Message}");
                Console.WriteLine($"堆栈跟踪：{ex.StackTrace}");
            }
        }

        /// <summary>
        /// 创建训练项目示例
        /// </summary>
        static void CreateTrainingProjectExample()
        {
            Console.WriteLine("\n=== 创建训练项目示例 ===\n");

            string projectFolder = @"D:\work\WindowsFormsTest\OCRTest\TrainingProject";
            
            try
            {
                OCRTrainingHelper.CreateSampleProject(projectFolder);
                
                Console.WriteLine("\n项目创建成功！");
                Console.WriteLine($"项目路径：{projectFolder}");
                Console.WriteLine("\n下一步：");
                Console.WriteLine("1. 将训练图片放入 raw_images 文件夹");
                Console.WriteLine("2. 按照 README.md 的说明进行训练");
            }
            catch (Exception ex)
            {
                Console.WriteLine($"创建失败：{ex.Message}");
            }

            Console.WriteLine("\n按任意键继续...");
            Console.ReadKey();
        }

        /// <summary>
        /// 预处理训练图片示例
        /// </summary>
        static void PreprocessTrainingImagesExample()
        {
            Console.WriteLine("\n=== 预处理训练图片 ===\n");

            string inputFolder = @"D:\work\WindowsFormsTest\OCRTest\TrainingProject\raw_images";
            string outputFolder = @"D:\work\WindowsFormsTest\OCRTest\TrainingProject\processed_images";

            if (!Directory.Exists(inputFolder))
            {
                Console.WriteLine($"输入文件夹不存在：{inputFolder}");
                Console.WriteLine("请先创建训练项目并放入原始图片");
                return;
            }

            try
            {
                OCRTrainingHelper.PreprocessTrainingImages(
                    inputFolder, 
                    outputFolder,
                    targetWidth: 800,
                    targetHeight: 0  // 自动计算高度
                );
            }
            catch (Exception ex)
            {
                Console.WriteLine($"处理失败：{ex.Message}");
            }

            Console.WriteLine("\n按任意键继续...");
            Console.ReadKey();
        }

        /// <summary>
        /// 验证数据集示例
        /// </summary>
        static void ValidateDatasetExample()
        {
            Console.WriteLine("\n=== 验证训练数据集 ===\n");

            string datasetFolder = @"D:\work\WindowsFormsTest\OCRTest\TrainingProject\processed_images";

            if (!Directory.Exists(datasetFolder))
            {
                Console.WriteLine($"数据集文件夹不存在：{datasetFolder}");
                return;
            }

            try
            {
                var missingFiles = OCRTrainingHelper.ValidateDataset(datasetFolder);
                
                if (missingFiles.Count == 0)
                {
                    Console.WriteLine("\n✓ 数据集完整，可以开始训练");
                }
                else
                {
                    Console.WriteLine($"\n✗ 发现 {missingFiles.Count} 个缺少标注文件的图片");
                    Console.WriteLine("请为这些图片创建对应的 .txt 标注文件");
                }
            }
            catch (Exception ex)
            {
                Console.WriteLine($"验证失败：{ex.Message}");
            }

            Console.WriteLine("\n按任意键继续...");
            Console.ReadKey();
        }

        /// <summary>
        /// 分析字符频率示例
        /// </summary>
        static void AnalyzeCharactersExample()
        {
            Console.WriteLine("\n=== 分析字符频率 ===\n");

            string labelFolder = @"D:\work\WindowsFormsTest\OCRTest\TrainingProject\labels";

            if (!Directory.Exists(labelFolder))
            {
                Console.WriteLine($"标注文件夹不存在：{labelFolder}");
                return;
            }

            try
            {
                var frequency = OCRTrainingHelper.AnalyzeCharacterFrequency(labelFolder);
                
                Console.WriteLine("\n提示：根据字符频率可以优化白名单配置");
                Console.WriteLine("高频字符应该优先保证识别准确率");
            }
            catch (Exception ex)
            {
                Console.WriteLine($"分析失败：{ex.Message}");
            }

            Console.WriteLine("\n按任意键继续...");
            Console.ReadKey();
        }

        public static void OCRTest1()
        {
            OCRModelConfig config = null;
            PaddleOCREngine engine = new PaddleOCREngine(config, "");
            bool Image1Finsh = false;
            bool Image2Finsh = false;
            bool Image3Finsh = false;
            Mat image1 = Cv2.ImRead("D:/work/WindowsFormsTest/TestImage/Date1_20260411170215701.jpg");
            OCRResult ocrResult1 = engine.DetectText(image1.ToBytes());
            Stopwatch sw = Stopwatch.StartNew();
            Task.Run(() =>
            {

                Mat image = Cv2.ImRead("D:/work/WindowsFormsTest/TestImage/Date1_20260411170215701.jpg");
                OCRResult ocrResult = engine.DetectText(image.ToBytes());
                foreach (var item in ocrResult.TextBlocks)
                {
                    Console.WriteLine("图像1结果：{0}", item.Text);
                }
                Image1Finsh = true;
                //if (Image1Finsh && Image2Finsh && Image3Finsh)
                //{
                    sw.Stop();
                    Console.WriteLine("耗时：" + sw.ElapsedMilliseconds);
                //}
            });
            //Task.Run(() =>
            //{
            //    Mat image = Cv2.ImRead("D:/work/WindowsFormsTest/TestImage/3_20250908104000075.jpg", ImreadModes.Color);
            //    OCRResult ocrResult = engine.DetectText(image.ToBytes());
            //    foreach (var item in ocrResult.TextBlocks)
            //    {
            //        Console.WriteLine("图像2结果：{0}", item.Text);
            //    }
            //    Image2Finsh = true;
            //    if (Image1Finsh && Image2Finsh && Image3Finsh)
            //    {
            //        sw.Stop();
            //        Console.WriteLine("耗时：" + sw.ElapsedMilliseconds);
            //    }
            //});
            //Task.Run(() =>
            //{
            //    Mat image = Cv2.ImRead("D:/work/WindowsFormsTest/TestImage/3_20250908104000075.jpg", ImreadModes.Color);
            //    OCRResult ocrResult = engine.DetectText(image.ToBytes());
            //    foreach (var item in ocrResult.TextBlocks)
            //    {
            //        Console.WriteLine("图像3结果：{0}", item.Text);
            //    }
            //    Image3Finsh = true;
            //    if (Image1Finsh && Image2Finsh && Image3Finsh)
            //    {
            //        sw.Stop();
            //        Console.WriteLine("耗时：" + sw.ElapsedMilliseconds);
            //    }
            //});
            Console.ReadKey();
        }

        public static string OCRTest2()
        {
            using (var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default))
            {
                // 只允许识别英文字母、数字和常见符号
                engine.SetVariable("tessedit_char_whitelist", "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789 .,!?-@#$%");

                using (var img = Pix.LoadFromFile("D:/work/WindowsFormsTest/TestImage/wechat_2025-10-16_143143_308.png"))
                {
                    using (var page = engine.Process(img))
                    {
                        var text = page.GetText();
                        Console.WriteLine(text);
                        return text;
                    }

                }
            }
        }
    }
}