求免费.NET(MVC)扫描+OCR提取指定字段存入数据库的类库

Hey there! Let's break down what you need—scanning documents from a scanner, running OCR to pull text, and extracting specific fields (like name and insurance ID) to save to a database—using free .NET MVC-compatible tools. Here's a practical, developer-focused breakdown:

If you're building a desktop .NET MVC app, the built-in Windows Image Acquisition (WIA) API is a solid free option. It lets you interact directly with scanners without extra paid libraries. You'll just need to add a reference to the WIA interop assembly (usually Interop.WIA.dll) or leverage scanning utilities in the System.Windows.Forms namespace (Windows-only, but free).

For ASP.NET MVC (web apps): Browsers can't access scanners directly, so you'll need users to scan documents locally first (via a desktop tool or free client-side JS scanning tools) then upload the scanned image to your MVC backend for processing.

Tesseract is a powerful open-source OCR engine, and its .NET wrapper Tesseract.NET is perfect for this use case. It's free, NuGet-installable, and works with both desktop and web .NET projects.

To get started:

• Install the Tesseract NuGet package in your MVC project.
• Download the Tesseract language data pack (English is default) and place it in a tessdata folder in your project.

Once you have the raw OCR text, regex is your best bet for pulling specific fields—assuming your insurance documents have consistent formatting (e.g., Name: John Doe or Insurance ID: INS-12345).

For example, if your document has lines like:

Full Name: Jane Smith
Insurance Policy ID: INS-98765

You can craft regex patterns to match these lines and extract the values cleanly.

Here's a simplified snippet that ties scanning, OCR, field extraction, and database saving together (using EF Core for database operations):

// Step 1: Scan document (WIA example for desktop)
var scanner = new WIA.DeviceManager();
var scannerDevice = scanner.DeviceInfos.Cast<WIA.DeviceInfo>()
    .First(d => d.Type == WIA.WiaDeviceType.ScannerDeviceType)
    .Connect();

var scanItem = scannerDevice.Items.Cast<WIA.Item>().First();
var imageFile = (WIA.ImageFile)scanItem.Transfer(WIA.FormatID.wiaFormatPNG);
imageFile.SaveFile(@"C:\ScannedDocs\insurance_scan.png");

// Step 2: Run OCR with Tesseract
using var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default);
using var img = Pix.LoadFromFile(@"C:\ScannedDocs\insurance_scan.png");
using var page = engine.Process(img);
var ocrText = page.GetText();

// Step 3: Extract Name and Insurance ID with regex
var nameMatch = Regex.Match(ocrText, @"Full Name:\s*(.*)", RegexOptions.IgnoreCase);
var insuranceIdMatch = Regex.Match(ocrText, @"Insurance Policy ID:\s*(.*)", RegexOptions.IgnoreCase);

string extractedName = nameMatch.Success ? nameMatch.Groups[1].Value.Trim() : string.Empty;
string extractedId = insuranceIdMatch.Success ? insuranceIdMatch.Groups[1].Value.Trim() : string.Empty;

// Step 4: Save to database (EF Core example)
using var dbContext = new YourInsuranceDbContext();
dbContext.InsuranceRecords.Add(new InsuranceRecord
{
    FullName = extractedName,
    PolicyId = extractedId,
    ScannedDate = DateTime.Now
});
await dbContext.SaveChangesAsync();

• WIA is Windows-only—if you need cross-platform support, free scanning options are limited, but you can pair cross-platform image processing libraries (like ImageSharp) with user-uploaded scans.
• Tesseract's accuracy improves with high-resolution, clean scans. Preprocess images (resize, adjust contrast) using free .NET libraries to boost OCR results.
• Refine your regex patterns with sample scans—document formatting can vary, so test with real-world examples to ensure reliable extraction.

内容的提问来源于stack exchange，提问作者lorraine chen