我正在寻找一个简单和快速的函数，没有额外的库来做这项工作。我尝试了尼克斯解决方案，但我需要29.4秒来“提取”一个atlas文件的1195张图像。所以后来我用这种方法做了同样的工作，需要2.43秒。也许这个会有帮助。

// content of the Texture class
public class Texture
{
    //name of the texture
    public string name { get; set; }
    //x position of the texture in the atlas image
    public int x { get; set; }
    //y position of the texture in the atlas image
    public int y { get; set; }
    //width of the texture in the atlas image
    public int width { get; set; }
    //height of the texture in the atlas image
    public int height { get; set; }
}

Bitmap atlasImage = new Bitmap(@"C:\somepicture.png");
PixelFormat pixelFormat = atlasImage.PixelFormat;

foreach (Texture t in textureList)
{
     try
     {
           CroppedImage = new Bitmap(t.width, t.height, pixelFormat);
           // copy pixels over to avoid antialiasing or any other side effects of drawing
           // the subimages to the output image using Graphics
           for (int x = 0; x < t.width; x++)
               for (int y = 0; y < t.height; y++)
                   CroppedImage.SetPixel(x, y, atlasImage.GetPixel(t.x + x, t.y + y));
           CroppedImage.Save(Path.Combine(workingFolder, t.name + ".png"), ImageFormat.Png);
     }
     catch (Exception ex)
     {
          // handle the exception
     }
}

对于任何愿意使用“不安全”代码的人来说，您可以获得比标准System.Drawing.Graphics方法更好的性能，如果使用Bitmap.Clone()则更好。

请记住，32bpp是该方法支持的唯一格式。(其他格式可以工作，只要1像素存储为4字节)

我包含了两个版本，一个使用Span，它在裁剪到较小的图像时性能稍好。 如果裁剪到1000x1000的图像，它们的速度差不多。

如果感兴趣，基准在下面。

public static class BitmapExtension
{
    unsafe public static Bitmap Crop(this Bitmap bitmap, int left, int top, int width, int height)
    {
        Bitmap cropped = new Bitmap(width, height);
        BitmapData originalData = bitmap.LockBits(new System.Drawing.Rectangle(0, 0, bitmap.Width, bitmap.Height), ImageLockMode.ReadOnly, bitmap.PixelFormat);
        BitmapData croppedData = cropped.LockBits(new System.Drawing.Rectangle(0, 0, width, height), ImageLockMode.WriteOnly, bitmap.PixelFormat);

        int* srcPixel = (int*)originalData.Scan0 + (left + originalData.Width * top);
        int nextLine = originalData.Width - width;

        for (int y = 0, i = 0; y < height; y++, srcPixel += nextLine)
        {
            for (int x = 0; x < width; x++, i++, srcPixel++)
            {
                *((int*)croppedData.Scan0 + i) = *srcPixel;
            }
        }

        bitmap.UnlockBits(originalData);
        cropped.UnlockBits(croppedData);

        return cropped;
    }

    unsafe public static Bitmap CropSmall(this Bitmap bitmap, int left, int top, int width, int height)
    {
        Bitmap cropped = new Bitmap(width, height);
        BitmapData originalData = bitmap.LockBits(new System.Drawing.Rectangle(0, 0, bitmap.Width, bitmap.Height), ImageLockMode.ReadOnly, bitmap.PixelFormat);
        BitmapData croppedData = cropped.LockBits(new System.Drawing.Rectangle(0, 0, width, height), ImageLockMode.WriteOnly, bitmap.PixelFormat);

        Span<int> srcPixels = new Span<int>((void*)originalData.Scan0, originalData.Width * originalData.Height);

        int nextLine = originalData.Width - width;

        for (int y = 0, i = 0, s = left + originalData.Width * top; y < height; y++, s += nextLine)
        {
            for (int x = 0; x < width; x++, i++, s++)
            {
                *((int*)croppedData.Scan0 + i) = srcPixels[s];
            }
        }

        bitmap.UnlockBits(originalData);
        cropped.UnlockBits(croppedData);

        return cropped;
    }
}

裁剪3440x1440到1000x1000

裁剪3440x1440到256x256

裁剪3440x1440到1440x1440

这里有一个裁剪图像的简单例子

public Image Crop(string img, int width, int height, int x, int y)
{
    try
    {
        Image image = Image.FromFile(img);
        Bitmap bmp = new Bitmap(width, height, PixelFormat.Format24bppRgb);
        bmp.SetResolution(80, 60);

        Graphics gfx = Graphics.FromImage(bmp);
        gfx.SmoothingMode = SmoothingMode.AntiAlias;
        gfx.InterpolationMode = InterpolationMode.HighQualityBicubic;
        gfx.PixelOffsetMode = PixelOffsetMode.HighQuality;
        gfx.DrawImage(image, new Rectangle(0, 0, width, height), x, y, width, height, GraphicsUnit.Pixel);
        // Dispose to free up resources
        image.Dispose();
        bmp.Dispose();
        gfx.Dispose();

        return bmp;
    }
    catch (Exception ex)
    {
        MessageBox.Show(ex.Message);
        return null;
    }            
}

在c#中裁剪图像非常简单。然而，做的东西，你将如何管理裁剪你的图像将会有点困难。

下面的示例是如何在c#中裁剪图像的方法。

var filename = @"c:\personal\images\horizon.png";
var img = Image.FromFile(filename);
var rect = new Rectangle(new Point(0, 0), img.Size);
var cloned = new Bitmap(img).Clone(rect, img.PixelFormat);
var bitmap = new Bitmap(cloned, new Size(50, 50));
cloned.Dispose();

你可以使用[图形。使用DrawImage][1]将裁剪图像从位图绘制到图形对象上。

Rectangle cropRect = new Rectangle(...);
using (Bitmap src = Image.FromFile("") as Bitmap)
{
    using (Bitmap target = new Bitmap(cropRect.Width, cropRect.Height))
    {
        using (Graphics g = Graphics.FromImage(target))
        {
            g.DrawImage(src, new Rectangle(0, 0, target.Width, target.Height),
                cropRect,
                GraphicsUnit.Pixel);
        }
    }
}

比公认的答案更简单的是:

public static Bitmap cropAtRect(this Bitmap b, Rectangle r)
{
    using (var nb = new Bitmap(r.Width, r.Height))
    {
        using (Graphics g = Graphics.FromImage(nb))
        {
            g.DrawImage(b, -r.X, -r.Y);
            return nb;
        }
    }
}

并且它避免了最简单答案的“内存不足”异常风险。

注意，位图和图形是可分割的，因此使用了using子句。

编辑:我发现这是很好的png保存的位图。Save或Paint.exe，但失败的png保存，如Paint Shop Pro 6 -内容被取代。添加GraphicsUnit。Pixel给出了不同的错误结果。也许只是这些失败的png有问题。