QT+opencv实现带通处理图片

//参数分别为 mat，类型（高通是0，带通是1，低通是2），径向中心，带宽
Mat FourierTransform(Mat image, int type, float D0, float w)
{
    image.convertTo(image, CV_32F);
    int Width =image.cols;
    int height = image.rows;
    //扩充图像
    copyMakeBorder(image, image, 50, 50, 50, 50, BORDER_REFLECT);

    vector<Mat> channels;
    split(image, channels);  //分离RGB通道
    Mat image_B = channels[0];
    //选取最适合做fft的宽和高

    int m1 = getOptimalDFTSize(image_B.rows);
    int n1 = getOptimalDFTSize(image_B.cols);
    Mat padded;
    //填充
    copyMakeBorder(image_B, padded, 0, m1 - image_B.rows, 0, n1 - image_B.cols, BORDER_REFLECT, Scalar::all(0));
    Mat planes[] = { Mat_<float>(padded), Mat::zeros(padded.size(), CV_32F) };
    Mat complexI;
    merge(planes, 2, complexI);  //planes[0], planes[1]是实部和虚部

    dft(complexI, complexI ,DFT_SCALE|DFT_COMPLEX_OUTPUT);
//    dft(complexI, complexI ,DFT_REAL_OUTPUT+DFT_INVERSE+DFT_SCALE);
    split(complexI, planes);

    //定义幅度谱和相位谱
    Mat ph, mag, idft;
    phase(planes[0], planes[1], ph);
    magnitude(planes[0], planes[1], mag);  //由实部planes[0]和虚部planes[1]得到幅度谱mag和相位谱ph

    //重新排列傅里叶图像中的象限，使得原点位于图像中心
    int cx = mag.cols / 2;
    int cy = mag.rows / 2;
    Mat q0(mag, Rect(0, 0, cx, cy));       //左上角图像划定ROI区域
    Mat q1(mag, Rect(cx, 0, cx, cy));      //右上角图像
    Mat q2(mag, Rect(0, cy, cx, cy));      //左下角图像
    Mat q3(mag, Rect(cx, cy, cx, cy));     //右下角图像

    //变换左上角和右下角象限
    Mat tmp;
    q0.copyTo(tmp);
    q3.copyTo(q0);
    tmp.copyTo(q3);

    //变换右上角和左下角象限
    q1.copyTo(tmp);
    q2.copyTo(q1);
    tmp.copyTo(q2);

    //高通滤波
    int i=0;
    int j=0;
    switch (type) {
    case 0 :{
        for (i = 0; i < mag.cols;i++){
            for (j = 0; j < mag.rows; j++){
                float r0=sqrt(pow(abs(float(i-mag.cols/2)),2)+pow(abs(float(j-mag.rows/2)),2));
                mag.at<float>(j, i)=(mag.at<float>(j, i))*(1.0 - float(1.0/(1.0+pow(r0/D0,2*1)))+5);
            }
        }
        break;
    }
    //带通
    case 1 :{
        float r0;
        for (i = 0; i < mag.cols;i++){
            for (j = 0; j < mag.rows; j++){
                r0=sqrt(pow(i-mag.cols/2,2)+pow(j-mag.rows/2,2));
                if(w!=0 && r0 != 0){
                    mag.at<float>(j, i) =(mag.at<float>(j, i)) *(1-(1/(1+pow((r0*w)/(pow(r0,2)-pow(D0,2)),2*1))));
                }
                //高斯
//                if(w != 0 && r0 != 0){
//                    mag.at<float>(j, i) = saturate_cast<float>(mag.at<float>(j, i) * exp(-pow((pow(r0,2)-pow(D0,2))/(r0*w),2)));
//                }
                //理想
//                 if(w != 0 && r0 != 0){
//                      if(r0 >= D0-(w/2.0) && r0 <= D0+(w/2.0)){
//                          mag.at<float>(j, i) = 1;
//                      }else{
//                          mag.at<float>(j, i) = 0;
//                      }
//                 }

//                if(j > mag.rows/2-10 && j <mag.rows/2+10 ){
//                    if(i < mag.cols/2-50 || i > mag.cols/2+50){
//                        mag.at<float>(j, i) = 0;
//                    }
//                }
//                if(i > mag.cols/2-20 && i <mag.cols/2+20 ){
//                    if(j < mag.rows/2-80 || j > mag.rows/2+80){
//                        mag.at<float>(j, i) = 0;
//                    }
//                }

            }

        }

        break;
    }
        //低通
    case 2 :{
        for (i = 0; i < mag.cols;i++){
            for (j = 0; j < mag.rows; j++){
                if (abs(i - mag.cols / 2) < mag.cols / 5 && abs(j - mag.rows / 2) < mag.rows / 5)
                    mag.at<float>(j, i) = 0;
            }
        }
        break;
    }
    default:
        break;
    }
    q0.copyTo(tmp);
    q3.copyTo(q0);
    tmp.copyTo(q3);

    //变换右上角和左下角象限
    q1.copyTo(tmp);
    q2.copyTo(q1);
    tmp.copyTo(q2);

    //傅里叶逆变换
    polarToCart(mag, ph, planes[0], planes[1]);  //由幅度谱mag和相位谱ph恢复实部planes[0]和虚部planes[1]
    merge(planes, 2, idft);
    cv::idft(idft, idft,DFT_REAL_OUTPUT);
    image_B = idft(Rect(0, 0, image.cols & -2, image.rows & -2));
    image_B.copyTo(channels[0]);

    merge(channels, image);
    image=image(Rect(50,50,Width,height));
    image.convertTo(image, CV_16UC1);
    return image;
}

实现高通，低通和带通及带通的不同算法。