PowerShell内嵌C#调用Win10内置OCR程序集加载异常求助

加载失败根因 #

你遇到的dll加载错误和路径定位无关，核心原因有两点：

• c:\Windows\System32\Windows.Media.Ocr.dll是WinRT原生系统组件，不是标准.NET程序集，直接通过Add-Type -ReferencedAssemblies指定dll路径加载的方式无法识别其格式，必然触发“尝试加载格式不正确的程序”错误
• 原测试代码存在语法缺陷：异步方法返回值应为Task<string>而非直接返回string，且缺少WinRT类型互操作、文件加载、识别结果处理的必要逻辑。

可直接运行的实现代码 #

该实现将OCR识别、BoundingRect坐标处理逻辑全部放在内嵌C#层执行，无跨语言调用开销，性能和纯C#实现一致，无需手动定位OCR相关dll路径：

# 加载WinRT互操作基础程序集
Add-Type -AssemblyName System.Runtime.WindowsRuntime
$null = [Windows.Storage.StorageFile, Windows.Storage, ContentType = WindowsRuntime]
$null = [Windows.Media.Ocr.OcrEngine, Windows.Foundation, ContentType = WindowsRuntime]
$null = [Windows.Globalization.Language, Windows.Globalization, ContentType = WindowsRuntime]
$null = [Windows.Graphics.Imaging.BitmapDecoder, Windows.Graphics.Imaging, ContentType = WindowsRuntime]

# 内嵌C#实现完整OCR逻辑
Add-Type -Language CSharp -ReferencedAssemblies System.Runtime, System.Runtime.WindowsRuntime, System.Threading.Tasks @'
using System;
using System.Collections.Generic;
using System.Threading.Tasks;
using Windows.Storage;
using Windows.Graphics.Imaging;
using Windows.Media.Ocr;
using Windows.Globalization;

public static class Win10OCR
{
    public class WordInfo
    {
        public string Text { get; set; }
        public double X { get; set; }
        public double Y { get; set; }
        public double Width { get; set; }
        public double Height { get; set; }
    }

    public class OcrOutput
    {
        public string FullText { get; set; }
        public List<WordInfo> Words { get; set; }
    }

    public async static Task<OcrOutput> Recognize(string imagePath, string langTag = "zh-CN")
    {
        var lang = new Language(langTag);
        var ocrEngine = OcrEngine.TryCreateFromLanguage(lang);
        if (ocrEngine == null) throw new InvalidOperationException($"系统未安装{langTag}对应的OCR语言包");

        var imgFile = await StorageFile.GetFileFromPathAsync(imagePath);
        using (var stream = await imgFile.OpenAsync(FileAccessMode.Read))
        {
            var decoder = await BitmapDecoder.CreateAsync(stream);
            var bitmap = await decoder.GetSoftwareBitmapAsync();
            var rawResult = await ocrEngine.RecognizeAsync(bitmap);

            // C#层直接处理BoundingRect，无跨语言性能损耗
            var wordList = new List<WordInfo>();
            foreach (var line in rawResult.Lines)
            {
                foreach (var word in line.Words)
                {
                    wordList.Add(new WordInfo
                    {
                        Text = word.Text,
                        X = word.BoundingRect.X,
                        Y = word.BoundingRect.Y,
                        Width = word.BoundingRect.Width,
                        Height = word.BoundingRect.Height
                    });
                }
            }

            return new OcrOutput
            {
                FullText = rawResult.Text,
                Words = wordList
            };
        }
    }
}
'@

# 调用示例
$targetImg = Join-Path $PSScriptRoot "20000111-p001.jpg"
# 英文识别传langTag为"en"，异步方法必须等待执行完成再取结果
$ocrResult = [Win10OCR]::Recognize($targetImg, "en").GetAwaiter().GetResult()

# 输出结果
Write-Host "识别全文："
$ocrResult.FullText
Write-Host "`n单词坐标列表："
$ocrResult.Words | Format-Table

使用注意事项 #

• 运行环境要求：PowerShell 5.1及以上版本，Windows 10 1809+或Windows 11系统
• 切换识别语言时，修改Recognize方法的langTag参数即可，前提是已在系统语言设置中安装对应语言的「光学字符识别」可选功能
• 禁止直接引用System32路径下的WinRT原生dll，所有WinRT类型通过ContentType = WindowsRuntime标记从系统元数据缓存加载即可
• 所有结果遍历、坐标处理逻辑均在C#层完成，不存在PowerShell与C#跨环境遍历集合的性能损耗，执行速度和纯C#实现一致

内容的提问来源于stack exchange，提问作者Darin