VS2010实现opencv基于DCT的图像压缩

1.基于DCT的图像压缩方法是将一幅图像分割成矩形像素块，再分别对每一独立的像素块DCT变换、量化、编码和传输。

2.打开VS2010，新建Visual C++下Win32控制台应用程序demo，主程序如下：

// demo.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"
#include "highgui.h"
#include <math.h>
#include <cv.h>
#include "cxcore.h"
#define cvCvtPlaneToPix cvMerge
double PSNR_B = 0; 
double PSNR_G = 0; 
double PSNR_R = 0; 
double PSNR; 

int _tmain(int argc, _TCHAR* argv[])
{
 const char* imagename = "D:/demo/demo.jpg";
 IplImage *src; 
    CvScalar SrcPixel; 
    CvScalar DstPixel; 
    double sumB = 0; 
    double sumG = 0; 
    double sumR = 0; 
    double mseB; 
    double mseG; 
    double mseR; 
 
    src= cvLoadImage( imagename,1 ) ;   
    if( !src ) 
    { 
        printf("can't open the image...\n"); 
        return -1; 
    } 
    // YUV颜色空间  
    IplImage* YUVImage = cvCreateImage(cvSize(src->width,src->height), src->depth, 3); 
    IplImage* dst = cvCreateImage(cvSize(src->width,src->height), src->depth, 3); 
    // YUV颜色空间各通道  
    IplImage* Y = cvCreateImage(cvSize(src->width,src->height), IPL_DEPTH_8U, 1); 
    IplImage* U = cvCreateImage(cvSize(src->width,src->height), IPL_DEPTH_8U, 1); 
    IplImage* V = cvCreateImage(cvSize(src->width,src->height), IPL_DEPTH_8U, 1); 
     
     
    //cvNamedWindow( "Origin Image", CV_WINDOW_AUTOSIZE );  
    cvCvtColor(src, YUVImage, CV_BGR2YUV); //BGR→YUV  
    cvSplit( YUVImage, Y, U, V, NULL);//分割通道  
     
    CvMat* MatY = cvCreateMat(Y->height,Y->width,CV_64FC1); 
    CvMat* MatU = cvCreateMat(V->height,U->width,CV_64FC1); 
    CvMat* MatV = cvCreateMat(V->height,V->width,CV_64FC1); 
 
    CvMat* DCTY = cvCreateMat(Y->height, Y->width,CV_64FC1); 
    CvMat* DCTU = cvCreateMat(U->height, U->width,CV_64FC1); 
    CvMat* DCTV = cvCreateMat(V->height, V->width,CV_64FC1); 
     
    cvScale( Y, MatY ); 
    cvScale( U, MatU ); 
    cvScale( V, MatV ); 
 
    cvDCT(MatY, DCTY, CV_DXT_FORWARD); //余弦变换  
    cvDCT(MatU, DCTU, CV_DXT_FORWARD); //余弦变换  
    cvDCT(MatV, DCTV, CV_DXT_FORWARD); //余弦变换  
     
    //Y 通道压缩  
    for(int i = 0; i < Y->height; i++)   
    { 
        for(int j = 0; j < Y->width; j++) 
        { 
            double  element = CV_MAT_ELEM( *DCTY, double, i, j ); 
            if ( abs(element) < 10 ) 
                CV_MAT_ELEM( *DCTY, double, i, j ) = 0; 
        } 
    } 
 
    // U 通道压缩  
    for(int i = 0; i < U->height; i++)   
    { 
        for(int j = 0; j < U->width; j++) 
        { 
            double  element = CV_MAT_ELEM( *DCTU, double, i, j ); 
            if (     abs(element) < 20 ) 
                CV_MAT_ELEM( *DCTU, double, i, j ) = 0; 
        } 
    } 
 
    // V 通道压缩  
    for(int i = 0; i < V->height; i++)   
    { 
        for(int j = 0; j < V->width; j++) 
        { 
            double  element = CV_MAT_ELEM( *DCTV, double, i, j ); 
            if (     abs(element) < 20 ) 
                CV_MAT_ELEM( *DCTV, double, i, j ) = 0; 
        } 
    } 
    cvDCT(DCTY, MatY, CV_DXT_INVERSE); //余弦反变换  
    cvDCT(DCTU, MatU, CV_DXT_INVERSE);  
    cvDCT(DCTV, MatV, CV_DXT_INVERSE);  
 
    cvScale( MatY, Y ); 
    cvScale( MatU, U ); 
    cvScale( MatV, V ); 
     
    cvMerge( Y, U, V, NULL, YUVImage ); 
    cvCvtColor( YUVImage, dst, CV_YUV2BGR); //YUV→BGR  
 
    //  计算前后两幅图像的PSNR值  
    for(int i = 0; i < src->height; i++) 
    { 
        for(int j = 0; j < src->width; j++) 
        { 
            SrcPixel = cvGet2D( src, i, j ); 
            DstPixel = cvGet2D( dst, i, j ); 
            sumB += ( SrcPixel.val[0] - DstPixel.val[0] ) * ( SrcPixel.val[0] - DstPixel.val[0] ); 
            sumG += ( SrcPixel.val[1] - DstPixel.val[1] ) * ( SrcPixel.val[1] - DstPixel.val[1] ); 
            sumR += ( SrcPixel.val[2] - DstPixel.val[2] ) * ( SrcPixel.val[2] - DstPixel.val[2] ); 
             
        } 
    } 
    mseB = sumB / ((src->width) * (src->height)); //计算均方差  
    mseG = sumG / ((src->width) * (src->height));  
    mseR = sumR / ((src->width) * (src->height));  
 
    PSNR_B = 10.0 * ( log10( 255.0 * 255.0 / mseB ) ); 
    PSNR_G = 10.0 * ( log10( 255.0 * 255.0 / mseG ) ); 
    PSNR_R = 10.0 * ( log10( 255.0 * 255.0 / mseR ) ); 
    PSNR=(PSNR_B + PSNR_G + PSNR_R) / 3; 
     
    cvShowImage( "YImage", Y ); 
    cvShowImage( "UImage", U ); 
    cvShowImage( "VImage", V ); 
    cvShowImage( "DstImage", dst ); 
    cvSaveImage( "D:/demo/dstdemo.jpg", dst); 
 
    while( 1 ) 
    { 
        if( cvWaitKey(0) == 27 ) break; 
    } 
 
    cvDestroyWindow("YImage"); 
    cvDestroyWindow("UImage"); 
    cvDestroyWindow("VImage"); 
    cvDestroyWindow("DstImage"); 
 
     
    cvReleaseImage(&Y); 
    cvReleaseImage(&U); 
    cvReleaseImage(&V); 
    cvReleaseImage(&src); 
    cvReleaseImage(&dst); 
    cvReleaseImage(&YUVImage); 
 
    return 0; 

}

 

3.参考用，jpg图片越压缩越大，代码有待优化来达到压缩效果！