之前写过一个java生成缩略图的算法，最近使用发现大图片缩小以后，效果很差。特意搞到一个生成高清缩略图的算法。相对清晰不少，不过比使用photoshop软件生成的缩略图还是会差一点，不仔细对比，是看不出来的。
高清，都是相对的。这个生成以后效果还不错。
import java.awt.image.BufferedImage;
public class ImgHigh {
private int width;
private int height;
private int scaleWidth;
double support = (double) 3.0;
double PI = (double) 3.14159265358978;
double[] contrib;
double[] normContrib;
double[] tmpContrib;
int startContrib, stopContrib;
int nDots;
int nHalfDots;
/**
* Start: Use Lanczos filter to replace the original algorithm for image
* scaling. Lanczos improves quality of the scaled image modify by :blade
* */
public BufferedImage imageZoomOut(BufferedImage srcBufferImage, int w, int h) {
width = srcBufferImage.getWidth();
height = srcBufferImage.getHeight();
scaleWidth = w;
if (DetermineResultSize(w, h) == 1) {
return srcBufferImage;
}
CalContrib();
BufferedImage pbOut = HorizontalFiltering(srcBufferImage, w);
BufferedImage pbFinalOut = VerticalFiltering(pbOut, h);
return pbFinalOut;
}
/**
* 决定图像尺寸
* */
private int DetermineResultSize(int w, int h) {
double scaleH, scaleV;
scaleH = (double) w / (double) width;
scaleV = (double) h / (double) height;
// 需要判断一下scaleH，scaleV，不做放大操作
if (scaleH >= 1.0 && scaleV >= 1.0) {
return 1;
}
return 0;
} // end of DetermineResultSize()
private double Lanczos(int i, int inWidth, int outWidth, double Support) {
double x;
x = (double) i * (double) outWidth / (double) inWidth;
return Math.sin(x * PI) / (x * PI) * Math.sin(x * PI / Support)
/ (x * PI / Support);
} // end of Lanczos()
//
// Assumption: same horizontal and vertical scaling factor
//
private void CalContrib() {
nHalfDots = (int) ((double) width * support / (double) scaleWidth);
nDots = nHalfDots * 2 + 1;
try {
contrib = new double[nDots];
normContrib = new double[nDots];
tmpContrib = new double[nDots];
} catch (Exception e) {
System.out.println("init contrib,normContrib,tmpContrib " + e);
}
int center = nHalfDots;
contrib[center] = 1.0;
double weight = 0.0;
int i = 0;
for (i = 1; i <= center; i++) {
contrib[center + i] = Lanczos(i, width, scaleWidth, support);
weight += contrib[center + i];
}
for (i = center - 1; i >= 0; i--) {
contrib[i] = contrib[center * 2 - i];
}
weight = weight * 2 + 1.0;
for (i = 0; i <= center; i++) {
normContrib[i] = contrib[i] / weight;
}
for (i = center + 1; i < nDots; i++) {
normContrib[i] = normContrib[center * 2 - i];
}
} // end of CalContrib()
// 处理边缘
private void CalTempContrib(int start, int stop) {
double weight = 0;
int i = 0;
for (i = start; i <= stop; i++) {
weight += contrib[i];
}
for (i = start; i <= stop; i++) {
tmpContrib[i] = contrib[i] / weight;
}
} // end of CalTempContrib()
private int GetRedValue(int rgbValue) {
int temp = rgbValue & 0x00ff0000;
return temp >> 16;
}
private int GetGreenValue(int rgbValue) {
int temp = rgbValue & 0x0000ff00;
return temp >> 8;
}
private int GetBlueValue(int rgbValue) {
return rgbValue & 0x000000ff;
}
private int ComRGB(int redValue, int greenValue, int blueValue) {
return (redValue << 16) + (greenValue << 8) + blueValue;
}
// 行水平滤波
private int HorizontalFilter(BufferedImage bufImg, int startX, int stopX,
int start, int stop, int y, double[] pContrib) {
double valueRed = 0.0;
double valueGreen = 0.0;
double valueBlue = 0.0;
int valueRGB = 0;
int i, j;
for (i = startX, j = start; i <= stopX; i++, j++) {
valueRGB = bufImg.getRGB(i, y);
valueRed += GetRedValue(valueRGB) * pContrib[j];
valueGreen += GetGreenValue(valueRGB) * pContrib[j];
valueBlue += GetBlueValue(valueRGB) * pContrib[j];
}
valueRGB = ComRGB(Clip((int) valueRed), Clip((int) valueGreen),
Clip((int) valueBlue));
return valueRGB;
} // end of HorizontalFilter()
// 图片水平滤波
private BufferedImage HorizontalFiltering(BufferedImage bufImage, int iOutW) {
int dwInW = bufImage.getWidth();
int dwInH = bufImage.getHeight();
int value = 0;
BufferedImage pbOut = new BufferedImage(iOutW, dwInH,
BufferedImage.TYPE_INT_RGB);
for (int x = 0; x < iOutW; x++) {
int startX;
int start;
int X = (int) (((double) x) * ((double) dwInW) / ((double) iOutW) + 0.5);
int y = 0;
startX = X - nHalfDots;
if (startX < 0) {
startX = 0;
start = nHalfDots - X;
} else {
start = 0;
}
int stop;
int stopX = X + nHalfDots;
if (stopX > (dwInW - 1)) {
stopX = dwInW - 1;
stop = nHalfDots + (dwInW - 1 - X);
} else {
stop = nHalfDots * 2;
}
if (start > 0 || stop < nDots - 1) {
CalTempContrib(start, stop);
for (y = 0; y < dwInH; y++) {
value = HorizontalFilter(bufImage, startX, stopX, start,
stop, y, tmpContrib);
pbOut.setRGB(x, y, value);
}
} else {
for (y = 0; y < dwInH; y++) {
value = HorizontalFilter(bufImage, startX, stopX, start,
stop, y, normContrib);
pbOut.setRGB(x, y, value);
}
}
}
return pbOut;
} // end of HorizontalFiltering()
private int VerticalFilter(BufferedImage pbInImage, int startY, int stopY,
int start, int stop, int x, double[] pContrib) {
double valueRed = 0.0;
double valueGreen = 0.0;
double valueBlue = 0.0;
int valueRGB = 0;
int i, j;
for (i = startY, j = start; i <= stopY; i++, j++) {
valueRGB = pbInImage.getRGB(x, i);
valueRed += GetRedValue(valueRGB) * pContrib[j];
valueGreen += GetGreenValue(valueRGB) * pContrib[j];
valueBlue += GetBlueValue(valueRGB) * pContrib[j];
// System.out.println(valueRed+ "-> "+Clip((int)valueRed)+ " <- ");
//
// System.out.println(valueGreen+ "-> "+Clip((int)valueGreen)+
// " <- ");
// System.out.println(valueBlue+ "-> "+Clip((int)valueBlue)+ " <- "+
// "--> ");
}
valueRGB = ComRGB(Clip((int) valueRed), Clip((int) valueGreen),
Clip((int) valueBlue));
// System.out.println(valueRGB);
return valueRGB;
} // end of VerticalFilter()
private BufferedImage VerticalFiltering(BufferedImage pbImage, int iOutH) {
int iW = pbImage.getWidth();
int iH = pbImage.getHeight();
int value = 0;
BufferedImage pbOut = new BufferedImage(iW, iOutH,
BufferedImage.TYPE_INT_RGB);
for (int y = 0; y < iOutH; y++) {
int startY;
int start;
int Y = (int) (((double) y) * ((double) iH) / ((double) iOutH) + 0.5);
startY = Y - nHalfDots;
if (startY < 0) {
startY = 0;
start = nHalfDots - Y;
} else {
start = 0;
}
int stop;
int stopY = Y + nHalfDots;
if (stopY > (int) (iH - 1)) {
stopY = iH - 1;
stop = nHalfDots + (iH - 1 - Y);
} else {
stop = nHalfDots * 2;
}
if (start > 0 || stop < nDots - 1) {
CalTempContrib(start, stop);
for (int x = 0; x < iW; x++) {
value = VerticalFilter(pbImage, startY, stopY, start, stop,
x, tmpContrib);
pbOut.setRGB(x, y, value);
}
} else {
for (int x = 0; x < iW; x++) {
value = VerticalFilter(pbImage, startY, stopY, start, stop,
x, normContrib);
pbOut.setRGB(x, y, value);
}
}
}
return pbOut;
} // end of VerticalFiltering()
int Clip(int x) {
if (x < 0)
return 0;
if (x > 255)
return 255;
return x;
}
/**
* End: Use Lanczos filter to replace the original algorithm for image
* scaling. Lanczos improves quality of the scaled image modify by :blade
* */
}
外部的用法
BufferedImage outImage = new ImgHigh().imageZoomOut(src, scaledW, scaledH);
然后把outImage这个流保存为图片就好了。
this.saveAsFile(thumb, outImage, quality);
quality为图片质量，经过测试采用90%的质量，图片看起来不会有太大差别，但是文件会小到1/3。
附保存图片算法
public void saveAsFile(String outFileName, BufferedImage outImage,
int quality) {
BufferedOutputStream out;
try {
out = new BufferedOutputStream(
new FileOutputStream(outFileName));
JPEGImageEncoder encoder = JPEGCodec.createJPEGEncoder(out);
JPEGEncodeParam param = encoder
.getDefaultJPEGEncodeParam(outImage);
quality = Math.max(0, Math.min(quality, 100));
param.setQuality((float) quality / 100.0f, false);
encoder.setJPEGEncodeParam(param);
encoder.encode(outImage);
out.close();
} catch (Exception e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
具体缩略图缩小算法，请参考 http://java-er.com/blog/java-img-suo/