Delphi图像处理

    图像缩放是最常用的图像处理，在图像拉伸和取得图像略图中都要用到。图像缩放质量的好坏与图像像素插值方式有关，本文定义了常用的3种插值方式，即临近插值、线性插值和双立方插值方式：
type
// 插值方式: 缺省(线性插值)，临近，线性，双立方
TInterpolateMode = (imDefault, imNear, imBilinear, imBicubic);
    具体的缩放及其用到的插值过程代码如下：
过程定义：
// 设置双立方插值的斜率。缺省值为-0.75
procedure SetBicubicSlope(const Value: Single);
// 缩放图像，IpMode插值方式
procedure ImageScale(Dest: TImageData; const Source: TImageData;
IpMode: TInterpolateMode = imDefault); overload;
// Source分别按比例ScaleX和ScaleY缩放到Dest的(x,y)坐标，IpMode插值方式
procedure ImageScale(Dest: TImageData; x, y: Integer; const Source: TImageData;
ScaleX, ScaleY: Single; IpMode: TInterpolateMode = imDefault); overload;
// TGraphic对象缩放到Dest
procedure ImageScale(Dest: TImageData; const Source: TGraphic;
IpMode: TInterpolateMode = imDefault); overload;
procedure ImageScale(Dest: TImageData; x, y: Integer; const Source: TGraphic;
ScaleX, ScaleY: Single; IpMode: TInterpolateMode = imDefault); overload;
// TGpBitmap对象缩放到Dest
procedure ImageScale(Dest: TImageData;
const Source: TGpBitmap; IpMode: TInterpolateMode = imDefault); overload;
procedure ImageScale(Dest: TImageData; x, y: Integer;
const Source: TGpBitmap; ScaleX, ScaleY: Single;
IpMode: TInterpolateMode = imDefault); overload;
实现代码：
type
TInterpolateProc = procedure;
var
BicubicTable: Pointer;
BicubicSlope: Single;
BilinearTable: Pointer;
(*****************************************************************************
* typedef UINT ARGB *
* ARGB GetBilinearColor(int x(*256), int y(*256), void* Scan0, UINT Stride) *
* *
* int x0 = x / 256