C# 提取图片上的文字内容

一、平台:.net 6.0 控制台应用程序
二、插件安装


image.png

EPPlus 7.2.1
OpenCvSharp4.runtime.win 4.10.0.20240616
Sdcb.PaddleInference 2.5.0.1
Sdcb.PaddleInference.runtime.win64.mkl 2.6.1
Sdcb.PaddleOCR 2.7.0.3
Sdcb.PaddleOCR.Models.Local 2.7.0
System.Drawing.Common 8.0.7

三、实现代码

using OpenCvSharp;
using Sdcb.PaddleInference;
using Sdcb.PaddleOCR.Models.Local;
using Sdcb.PaddleOCR.Models;
using Sdcb.PaddleOCR;
using System.Diagnostics;
using System;
using System.Drawing;
using System.IO;
using OfficeOpenXml;

namespace ConsoleApp1
{
    internal class Program
    {
        static void Main(string[] args)
        {
            Console.WriteLine("【启动图片扫描程序】");
            Console.WriteLine("请指定一个文件根目录:");

            string rootFolder = Console.ReadLine();

            if (!Directory.Exists(rootFolder))
            {
                Console.WriteLine("【错误】路径不存在...");
                return;
            }

            try
            {
                Console.WriteLine("开始遍历文件夹...");
                string runDir = AppDomain.CurrentDomain.BaseDirectory;
                string savePath = $"{runDir}output.xlsx";
                ExcelPackage.LicenseContext = LicenseContext.NonCommercial;
                using (var package = new ExcelPackage())
                {
                    var sheet1 = package.Workbook.Worksheets.Add("结果表");
                    var sheet2 = package.Workbook.Worksheets.Add("错误表");

                    sheet1.SetValue(1, 1, "文件夹名称");
                    sheet1.SetValue(1, 2, "图片名称");
                    sheet1.SetValue(1, 3, "时间");
                    sheet1.SetValue(1, 4, "经度");
                    sheet1.SetValue(1, 5, "纬度");
                    sheet1.SetValue(1, 6, "地址");
                    sheet1.SetValue(1, 7, "路径");

                    sheet2.SetValue(1, 1, "文件夹名称");
                    sheet2.SetValue(1, 2, "图片名称");
                    sheet2.SetValue(1, 3, "路径");
                    sheet2.SetValue(1, 4, "识别错误原因");

                    TraverseDirectory(rootFolder, sheet1, sheet2, 2, 2);

                    package.SaveAs(savePath);
                }
                Console.WriteLine("识别完成,请查看输出文件!");
            }
            catch (Exception ex)
            {
                Console.WriteLine();
            }
        }

        static void TraverseDirectory(string path, ExcelWorksheet sheet1, ExcelWorksheet sheet2, int rowIndex1, int rowIndex2)
        {
            var rdir = new DirectoryInfo(path);
            Console.WriteLine($"***【文件夹{rdir.Name}】***");
            foreach (var fileInfo in rdir.GetFiles())
            {
                if (fileInfo.Extension == ".jpg" || fileInfo.Extension == ".png")
                {
                    Console.WriteLine($"正在识别文件:{fileInfo.Name}");
                    try
                    {
                        FullOcrModel model = LocalFullModels.ChineseV3;
                        using (PaddleOcrAll all = new PaddleOcrAll(model, PaddleDevice.Mkldnn())
                        {
                            AllowRotateDetection = false,
                            Enable180Classification = false,
                        })
                        {
                            using (Mat src = Cv2.ImRead(fileInfo.FullName))
                            {
                                PaddleOcrResult result = all.Run(src);
                                // 此处是我自定义将文字内容提取的方法,可以自己替换
                                var txt1 = result.Text.Replace("\r", "").Replace("\n", "").Replace(":", ":");

                                var index = txt1.IndexOf("间:");
                                if (index > -1)
                                {
                                    var txt2 = txt1.Substring(index);
                                    var i1 = txt2.IndexOf("度:");
                                    var i2 = txt2.LastIndexOf("度:");
                                    var i3 = txt2.IndexOf("址:");

                                    sheet1.SetValue(rowIndex1, 1, rdir.Name);
                                    sheet1.SetValue(rowIndex1, 2, fileInfo.Name);
                                    sheet1.SetValue(rowIndex1, 3, txt2.Substring(2, i1 - 2));
                                    sheet1.SetValue(rowIndex1, 4, txt2.Substring(i1 + 2, i2 - i1 - 2));
                                    sheet1.SetValue(rowIndex1, 5, txt2.Substring(i2 + 2, i3 - i2 - 2));
                                    sheet1.SetValue(rowIndex1, 6, txt2.Substring(i3 + 2));
                                    sheet1.SetValue(rowIndex1, 7, fileInfo.FullName);
                                    rowIndex1 += 1;

                                    Console.WriteLine($"识别成功...");
                                    //Console.WriteLine($"时间:{txt2.Substring(2, i1 - 2)}");
                                    //Console.WriteLine($"经度:{txt2.Substring(i1 + 2, i2 - i1 - 2)}");
                                    //Console.WriteLine($"纬度:{txt2.Substring(i2 + 2, i3 - i2 - 2)}");
                                    //Console.WriteLine($"地址:{txt2.Substring(i3 + 2)}");
                                }
                            }
                        }
                       
                    }
                    catch(Exception ex)
                    {
                        sheet2.SetValue(rowIndex2, 1, rdir.Name);
                        sheet2.SetValue(rowIndex2, 2, fileInfo.Name);
                        sheet2.SetValue(rowIndex2, 3, fileInfo.FullName);
                        sheet2.SetValue(rowIndex2, 4, ex.Message);
                        rowIndex2 += 1;
                        Console.WriteLine($"识别异常:{ex.Message}");
                    }
                }
            }

            foreach (string dir in Directory.GetDirectories(path))
            {
                TraverseDirectory(dir, sheet1, sheet2, rowIndex1, rowIndex2);
            }
        }
    }
}

效果:


image.png

原图:


image.png
©著作权归作者所有,转载或内容合作请联系作者
平台声明:文章内容(如有图片或视频亦包括在内)由作者上传并发布,文章内容仅代表作者本人观点,简书系信息发布平台,仅提供信息存储服务。

推荐阅读更多精彩内容