如何在 C# 中对映像执行 DCT

Question

public Rgb Jasnosc()
{
Byte rob =(byte)(0.299*this.r + 0.587*this.g + 0.114*this.b);
Rgb ret;
ret.r = rob;
ret.g = rob;
ret.b = rob;
return ret;
}

所以我有这种结构，我正在将加载的图像更改为灰度。 我的问题是我想使用类似的方法对图像执行DCT，但我完全不知道该怎么做。这是我用来执行它的代码：

public static Bitmap Jasnosc(Bitmap bitmapaWe)
{
int wysokosc = bitmapaWe.Height;
int szerokosc = bitmapaWe.Width;
Bitmap bitmapaWy = new Bitmap(szerokosc, wysokosc, PixelFormat.Format24bppRgb);
BitmapData bmWeData = bitmapaWe.LockBits(new Rectangle(0, 0, wysokosc, wysokosc), ImageLockMode.ReadWrite, PixelFormat.Format24bppRgb);
BitmapData bmWyData = bitmapaWy.LockBits(new Rectangle(0, 0, wysokosc, wysokosc), ImageLockMode.ReadWrite, PixelFormat.Format24bppRgb);
int strideWe = bmWeData.Stride;
int strideWy = bmWeData.Stride;
IntPtr scanWe = bmWeData.Scan0;
IntPtr scanWy = bmWyData.Scan0;
unsafe
{
for (int y = 0; y < wysokosc; y++)
{
byte* pWe = (byte*)(void*)scanWe + y * strideWe;
byte* pWy = (byte*)(void*)scanWy + y * strideWy;
for (int x = 0; x < szerokosc; x++)
{
((Rgb*)pWy)[x] = ((Rgb*)pWe)[x].Jasnosc();
((Rgb*)pWy)[y] = ((Rgb*)pWe)[y].Jasnosc();
}
}
}
bitmapaWy.UnlockBits(bmWyData);
bitmapaWe.UnlockBits(bmWeData);
return bitmapaWy;

Answer 1

我建议你使用"emguCV"("OpenCV"(进行此类操作。 这个库在性能方面设计得很好，并公开了许多有用的计算机视觉功能。

可以在 Git 存储库中找到DCT的示例。

/// <summary>
/// performs forward or inverse transform of 2D floating-point image
/// </summary>
/// <param name="type">Transformation flags</param>
/// <returns>The result of DCT</returns>
[ExposableMethod(Exposable = true, Category = "Discrete Transforms")]
public Image<TColor, Single> DCT(CvEnum.CV_DCT_TYPE type)
{
Image<TColor, Single> res = new Image<TColor, float>(Width, Height);
CvInvoke.cvDCT(Ptr, res.Ptr, type);
return res;
}

EMGUCV的DCT功能声明为：

public static void Dct(
IInputArray src,
IOutputArray dst,
DctType flags
)