使用IronOCR识别图片文字

最近实验课要做一个图片文字识别，需求是Winform中有一个图片，进行框选后，识别框选中的内容，老师提示说去找OCR的开源库，所以我就找到了IronOCR

IronOCR：The C# OCR Library : Iron OCR

IronOcr makes it easy to read text from images in your .net apps & websites.

1.Read text and barcodes from scanned images & PDFs

2.Supports multiple international languages

3.Output as plain text or structured data

框选

进行识别前，必须框选出识别区域，所以就要先去实现框选，思路就是通过MouseUp、MouseMove、MouseDown三个事件去绘制一个矩形，类似于拖选

    private bool _mouseIsDown = false;
    private Rectangle _selectArea = Rectangle.Empty;

    private void pictureBox_MouseUp(object sender, MouseEventArgs e)
    {
        // 修正Width和Height可能为负数的问题
        if (_selectArea.Width < 0)
        {
            _selectArea.Width = -_selectArea.Width;
            _selectArea.X -= _selectArea.Width;
        }

        if (_selectArea.Height < 0)
        {
            _selectArea.Height = -_selectArea.Height;
            _selectArea.Y -= _selectArea.Height;
        }

        Cursor.Clip = Rectangle.Empty;
        _mouseIsDown = false;
        DrawRectangle();

        _selectArea = Rectangle.Empty;
    }

    private void pictureBox_MouseMove(object sender, MouseEventArgs e)
    {
        if (_mouseIsDown)
            ResizeToRectangle(e.Location);
    }

    /// 
  
    /// 初始化选择框  
    ///   
    ///   
    private void DrawStart(Point startPoint)
    {
        //指定工作区为整个控件  
        Cursor.Clip = RectangleToScreen(new Rectangle(pictureBox.Location.X, pictureBox.Location.Y,
            pictureBox.Width, pictureBox.Height));
        _selectArea = new Rectangle(startPoint.X + pictureBox.Location.X,
            startPoint.Y + pictureBox.Location.Y,1,1);
    }
    /// 
  
    /// 在鼠标移动的时改变选择框的大小  
    ///   
    /// 鼠标的位置  
    private void ResizeToRectangle(Point p)
    {
        _selectArea.Width = p.X + pictureBox.Location.X - _selectArea.Left;
        _selectArea.Height = p.Y + pictureBox.Location.Y - _selectArea.Top;
        DrawRectangle();
    }
    /// 
  
    /// 绘制选择框  
    ///   
    private void DrawRectangle()
    {
        pictureBox.Refresh();
        var rect = RectangleToScreen(_selectArea);
        ControlPaint.DrawReversibleFrame(rect, Color.White, FrameStyle.Thick);
    }

图片文字识别

IronOCR的使用很简单，官网给出的教程：

    using IronOcr;
    //..
    var Ocr = new AutoOcr();
    var Result = Ocr.Read(@"C:\path\to\image.png");
    Console.WriteLine(Result.Text);

所以接下来只需要将框选区域进行截图，然后交给OCR识别即可

    private readonly AutoOcr _ocr;
    public MainForm()
    {
        InitializeComponent();
        _ocr = new AutoOcr();

        // 强制初始化OCR
        var img = new Bitmap(1,1);
        _ocr.Read(img);
    }

    // 框选的代码.......

    /// 

    /// 识别框选的单词/句子
    /// 
    /// 识别的结果
    private string SelectText()
    {
        var rect = RectangleToScreen(_selectArea);
        var img = new Bitmap(rect.Width, rect.Height);

        var graphics = Graphics.FromImage(img);
        graphics.CompositingQuality = CompositingQuality.HighQuality;
        graphics.CopyFromScreen(rect.Left, rect.Top, 0, 0,
            new Size(rect.Width, rect.Height));

        var result = _ocr.Read(img);

        Debug.WriteLine(result.Text);

        return result.Text;
    }

最后在MouseUp事件的时候调用SelectText即可，然后拿着识别出来的文字去干啥都行

项目源码：OCRDemo