Qt、Opencv联合开发
Qt在可视化和界面开发方面拥有很强大的功能和便捷性，opencv是一个强大的开源图像处理库，大家常常需要两者结合起来用。
本文介绍个人开发使用方式，大家可以参考下。
qt安装
官网下载直接安装就可以，我直接用的vs2015座编译和调试器。安装后可以把bin加到系统path里。
opencv安装
我用的3.46，opencv官网有提供exe直接安装，或者可以下载源码自己编译，编译没有什需要注意的选中“WITH_OPENGL”、“WITH_QT”，将“WITH_IPP”取消。“ENABLE_PRECOMPILED_HEADERS”取消。安装后可以把bin加到系统path里。
路径按照自己安装位置
qt配置opencv环境
qt和opencv环境都搭建好后，我们联合开发。
为了便于以后利用，我们把opencv配置单独写进api里。
首先一路默认新建一个qt widgets application 程序
建好后，选中.pro文件，右键在资源管理器打开，可以看到生成文件。
在这个文件夹下我们新建文件夹Qt_OPENCV文件夹，方便以后移植。
在Qt_OPENCV文件夹里新建两个.txt文件，我们手写pri文件和头文件。（个人认为pri是qmake为了便于整理代码用的，和直接放入pro一模一样）
重命名为 qt_halcon.pri 和 QT_Halcon
用记事本打开这两个文件分别写入
//qt_opencv.pri
INCLUDEPATH+=E:/opencv346/build/include/opencv
INCLUDEPATH+=E:/opencv346/build/include/opencv2
INCLUDEPATH+=E:/opencv346/build/include
CONFIG(debug, debug|release):{
LIBS+=-LE:/opencv346/build/x64/vc14/lib\
-lopencv_world346d
}else:CONFIG(release, debug|release):{
LIBS+=-LE:/opencv346/build/x64/vc14/lib\
-lopencv_world346
}
HEADERS += \
$$PWD/qt_opencv.h
SOURCES += \
$$PWD/qt_opencv.cpp
需要注意 cv命名空间经常和其他冲突，所以不建议在qt环境下开头都加 /using namespace cv
//QT_Opencv
#ifndef QT_QTOPENCV_MODULE_H
#define QT_QTOPENCV_MODULE_H
#include qt_opencv.h
#endif
qt_opencv.pri就是引用halcon的头文件和lib文件
具体路径要根据自己安装来定
QT_Opencv就是为了方便我们其他文件调用Opencv的头文件
qt_opencv.h下文提供
建立好模板后我们需要调用，打开我们的pro文件，文件最后添加
// untitleda.pro文件添加
INCLUDEPATH += $$PWD/Qt_OPENCV
include ($$PWD/Qt_OPENCV/qt_opencv.pri)
#过程文件存放位置
MOC_DIR = temp/moc #指定moc命令将含Q_OBJECT的头文件转换成标准.h文件的存放目录
RCC_DIR = temp/rcc #指定rcc命令将.qrc文件转换成qrc_*.h文件的存放目录
UI_DIR = temp/ui #指定rcc命令将.qrc文件转换成qrc_*.h文件的存放目录
OBJECTS_DIR = temp/obj #指定目标文件(obj)的存放目录
添加后，选中.pro文件，右键qmake，可以看到添加完成。
此时，这个程序的halcon环境就配置完成，这个Qt_OPENCV文件夹就可以作为我们的模板，以后需要用到opencv，直接把这个文件夹复制进去就好。
mat qimage互转 Yx_opencv_ImgChange类
qt_opencv.h下Yx_opencv_ImgChange类
class Yx_opencv_ImgChange;
// 图像转换
class Yx_opencv_ImgChange
{
public:
Yx_opencv_ImgChange();
~Yx_opencv_ImgChange();
QImage cvMat2QImage(const Mat& mat); // Mat 改成 QImage
Mat QImage2cvMat(QImage image); // QImage 改成 Mat
QImage splitBGR(QImage src, int color); // 提取RGB分量
QImage splitColor(QImage src, String model, int color); // 提取分量
};
QImage Yx_opencv_ImgChange::cvMat2QImage(const Mat& mat) // Mat 改成 QImage
{
if (mat.type() == CV_8UC1) // 单通道
{
QImage image(mat.cols, mat.rows, QImage::Format_Indexed8);
image.setColorCount(256); // 灰度级数256
for (int i = 0; i < 256; i++)
{
image.setColor(i, qRgb(i, i, i));
}
uchar *pSrc = mat.data; // 复制mat数据
for (int row = 0; row < mat.rows; row++)
{
uchar *pDest = image.scanLine(row);
memcpy(pDest, pSrc, mat.cols);
pSrc += mat.step;
}
return image;
}
else if (mat.type() == CV_8UC3) // 3通道
{
const uchar *pSrc = (const uchar*)mat.data; // 复制像素
QImage image(pSrc, mat.cols, mat.rows, (int)mat.step, QImage::Format_RGB888); // R, G, B 对应 0,1,2
return image.rgbSwapped(); // rgbSwapped是为了显示效果色彩好一些。
}
else if (mat.type() == CV_8UC4)
{
const uchar *pSrc = (const uchar*)mat.data; // 复制像素
// Create QImage with same dimensions as input Mat
QImage image(pSrc,mat.cols, mat.rows, (int)mat.step, QImage::Format_ARGB32); // B,G,R,A 对应 0,1,2,3
return image.copy();
}
else
{
return QImage();
}
}
Mat Yx_opencv_ImgChange::QImage2cvMat(QImage image) // QImage改成Mat
{
Mat mat;
switch (image.format())
{
case QImage::Format_ARGB32:
case QImage::Format_RGB32:
case QImage::Format_ARGB32_Premultiplied:
mat = Mat(image.height(), image.width(), CV_8UC4, (void*)image.constBits(), image.bytesPerLine());
break;
case QImage::Format_RGB888:
mat = Mat(image.height(), image.width(), CV_8UC3, (void*)image.constBits(), image.bytesPerLine());
cv::cvtColor(mat, mat, CV_BGR2RGB);
break;
case QImage::Format_Indexed8:
mat = Mat(image.height(), image.width(), CV_8UC1, (void*)image.constBits(), image.bytesPerLine());
break;
}
return mat;
}
QImage Yx_opencv_ImgChange::splitBGR(QImage src, int color) // 提取RGB分量
{
Mat srcImg, dstImg;
srcImg = QImage2cvMat(src);
if (srcImg.channels() == 1)
{
QMessageBox message(QMessageBox::Information, QString::fromLocal8Bit("提示"), QString::fromLocal8Bit("该图像为灰度图像。"));
message.exec();
return src;
}
else
{
vector<Mat> m;
split(srcImg, m);
vector<Mat>Rchannels, Gchannels, Bchannels;
split(srcImg, Rchannels);
split(srcImg, Gchannels);
split(srcImg, Bchannels);
Rchannels[1] = 0; Rchannels[2] = 0;
Gchannels[0] = 0; Gchannels[2] = 0;
Bchannels[0] = 0; Bchannels[1] = 0;
merge(Rchannels, m[0]);
merge(Gchannels, m[1]);
merge(Bchannels, m[2]);
dstImg = m[color]; // 分别对应B、G、R
QImage dst = cvMat2QImage(dstImg);
return dst;
}
}
QImage Yx_opencv_ImgChange::splitColor(QImage src, String model, int color) // 提取分量
{
Mat img = QImage2cvMat(src);
Mat img_rgb, img_hsv, img_hls, img_yuv, img_dst;
if (img.channels() == 1)
{
QUIHelper::showMessageBoxError("该图像为灰度图像。");
return src;
}
else
{
vector <Mat> vecRGB, vecHsv, vecHls, vecYuv;
img_hsv.create(img.rows, img.cols, CV_8UC3);
img_hls.create(img.rows, img.cols, CV_8UC3);
cvtColor(img, img_rgb, CV_BGR2RGB);
cvtColor(img, img_hsv, CV_BGR2HSV);
cvtColor(img, img_hls, CV_BGR2HLS);
cvtColor(img, img_yuv, CV_BGR2YUV);
split(img_rgb, vecRGB);
split(img_hsv, vecHsv);
split(img_hls, vecHls);
split(img_yuv, vecYuv);
if(model == "RGB")
img_dst = vecRGB[color];
else if (model == "HSV")
img_dst = vecHsv[color];
else if (model == "HLS")
img_dst = vecHls[color];
else if (model == "YUV")
img_dst = vecYuv[color];
else
img_dst = img;
QImage dst = cvMat2QImage(img_dst);
return dst;
}
}
opencv图像改变 Yx_opencv_Enhance类
qt_opencv.h下Yx_opencv_Enhance类
// 图像改变
class Yx_opencv_Enhance
{
public:
Yx_opencv_Enhance();
~Yx_opencv_Enhance();
QImage Normalized(QImage src, int kernel_length); // 简单滤波
QImage Gaussian(QImage src, int kernel_length); // 高斯滤波
QImage Median(QImage src, int kernel_length); // 中值滤波
QImage Sobel(QImage src, int kernel_length); // sobel边缘检测
QImage Laplacian(QImage src, int kernel_length); // laplacian边缘检测
QImage Canny(QImage src, int kernel_length, int lowThreshold); // canny边缘检测
QImage HoughLine(QImage src, int threshold, double minLineLength, double maxLineGap); // 线检测
QImage HoughCircle(QImage src, int minRadius, int maxRadius); // 圆检测
private:
Yx_opencv_ImgChange *imgchangeClass;
};
Yx_opencv_Enhance::Yx_opencv_Enhance()
{
imgchangeClass = new Yx_opencv_ImgChange;
}
Yx_opencv_Enhance::~Yx_opencv_Enhance()
{
}
QImage Yx_opencv_Enhance::Normalized(QImage src,int kernel_length) // 简单滤波
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
blur(srcImg, dstImg, Size(kernel_length, kernel_length), Point(-1, -1));
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::Gaussian(QImage src, int kernel_length) // 高斯滤波
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
GaussianBlur(srcImg, dstImg, Size(kernel_length, kernel_length), 0, 0);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::Median(QImage src, int kernel_length) // 中值滤波
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
medianBlur(srcImg, dstImg, kernel_length);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::HoughLine(QImage src, int threshold, double minLineLength, double maxLineGap) // 线检测
{
Mat srcImg, dstImg, cdstPImg;
srcImg = imgchangeClass->QImage2cvMat(src);
cv::Canny(srcImg, dstImg, 50, 200, 3); // Canny算子边缘检测
if (srcImg.channels() != 1)
cvtColor(dstImg, cdstPImg, COLOR_GRAY2BGR); // 转换灰度图像
else
cdstPImg = srcImg;
vector<Vec4i> linesP;
HoughLinesP(dstImg, linesP, 1, CV_PI / 180, threshold, minLineLength, maxLineGap);// 50,50,10
for (size_t i = 0; i < linesP.size(); i++)
{
Vec4i l = linesP[i];
line(cdstPImg, Point(l[0], l[1]), Point(l[2], l[3]), Scalar(0, 0, 255), 1, LINE_AA);
}
QImage dst = imgchangeClass->cvMat2QImage(cdstPImg);
return dst;
}
QImage Yx_opencv_Enhance::HoughCircle(QImage src, int minRadius, int maxRadius) // 圆检测
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat gray;
if (srcImg.channels() != 1)
cvtColor(srcImg, gray, COLOR_BGR2GRAY);
else
gray = srcImg;
medianBlur(gray, gray, 5); // 中值滤波，滤除噪声，避免错误检测
vector<Vec3f> circles;
HoughCircles(gray, circles, HOUGH_GRADIENT, 1, gray.rows / 16, 100, 30, minRadius, maxRadius); // Hough圆检测,100, 30, 1, 30
dstImg = srcImg.clone();
for (size_t i = 0; i < circles.size(); i++)
{
Vec3i c = circles[i];
Point center = Point(c[0], c[1]);
circle(dstImg, center, 1, Scalar(0, 100, 100), 3, LINE_AA); // 画圆
int radius = c[2];
circle(dstImg, center, radius, Scalar(255, 0, 255), 3, LINE_AA);
}
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::Sobel(QImage src, int kernel_length) // sobel
{
Mat srcImg, dstImg, src_gray;
srcImg = imgchangeClass->QImage2cvMat(src);
GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0, BORDER_DEFAULT); // 高斯模糊
if (srcImg.channels() != 1)
cvtColor(srcImg, src_gray, COLOR_BGR2GRAY); // 转换灰度图像
else
src_gray = srcImg;
Mat grad_x, grad_y, abs_grad_x, abs_grad_y;
cv::Sobel(src_gray, grad_x, CV_16S, 1, 0, kernel_length, 1, 0, BORDER_DEFAULT);
cv::Sobel(src_gray, grad_y, CV_16S, 0, 1, kernel_length, 1, 0, BORDER_DEFAULT);
convertScaleAbs(grad_x, abs_grad_x); // 缩放，计算绝对值，并将结果转换为8位
convertScaleAbs(grad_y, abs_grad_y);
addWeighted(abs_grad_x, 0.5, abs_grad_y, 0.5, 0, dstImg);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::Laplacian(QImage src, int kernel_length) // laplacian
{
Mat srcImg, dstImg, src_gray;
srcImg = imgchangeClass->QImage2cvMat(src);
GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0, BORDER_DEFAULT); // 高斯模糊
if (srcImg.channels() != 1)
cvtColor(srcImg, src_gray, COLOR_BGR2GRAY); // 转换灰度图像
else
src_gray = srcImg;
Mat abs_dst; // 拉普拉斯二阶导数
cv::Laplacian(src_gray, dstImg, CV_16S, kernel_length, 1, 0, BORDER_DEFAULT);
convertScaleAbs(dstImg, dstImg); // 绝对值8位
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Enhance::Canny(QImage src, int kernel_length ,int lowThreshold) // canny
{
Mat srcImg, dstImg, src_gray, detected_edges;
srcImg = imgchangeClass->QImage2cvMat(src);
dstImg.create(srcImg.size(), srcImg.type());
if (srcImg.channels() != 1)
cvtColor(srcImg, src_gray, COLOR_BGR2GRAY); // 转换灰度图像
else
src_gray = srcImg;
blur(src_gray, detected_edges, Size(3, 3)); // 平均滤波平滑
cv::Canny(detected_edges, detected_edges, lowThreshold, lowThreshold * 3, kernel_length);
dstImg = Scalar::all(0);
srcImg.copyTo(dstImg, detected_edges);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
opencv图像几何变换 Yx_opencv_Enhance类
qt_opencv.h下Yx_opencv_Enhance类
// 图像几何变换
class Yx_opencv_Geom
{
public:
Yx_opencv_Geom();
~Yx_opencv_Geom();
QImage Resize(QImage src, int length, int width);
QImage Enlarge_Reduce(QImage src, int times);
QImage Rotate(QImage src, int angle);
QImage Rotate_fixed(QImage src, int angle);
QImage Flip(QImage src, int flipcode);
QImage Lean(QImage src, int x, int y);
private:
Yx_opencv_ImgChange *imgchangeClass; // 大小类
};
Yx_opencv_Geom::Yx_opencv_Geom()
{
imgchangeClass = new Yx_opencv_ImgChange;
}
Yx_opencv_Geom::~Yx_opencv_Geom()
{
}
QImage Yx_opencv_Geom::Resize(QImage src, int length, int width) // 改变大小
{
Mat matSrc, matDst;
matSrc = imgchangeClass->QImage2cvMat(src);
resize(matSrc, matDst, Size(length, width), 0, 0, INTER_LINEAR);// 线性插值
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
QImage Yx_opencv_Geom::Enlarge_Reduce(QImage src, int times) // 缩放
{
Mat matSrc, matDst;
matSrc = imgchangeClass->QImage2cvMat(src);
if (times > 0)
{
resize(matSrc, matDst, Size(matSrc.cols * abs(times+1), matSrc.rows * abs(times+1)), 0, 0, INTER_LINEAR);
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
else if (times < 0)
{
resize(matSrc, matDst, Size(matSrc.cols / abs(times-1), matSrc.rows / abs(times-1)), 0, 0, INTER_AREA);
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
else
{
return src;
}
}
QImage Yx_opencv_Geom::Rotate(QImage src, int angle) // 旋转
{
Mat matSrc, matDst,M;
matSrc = imgchangeClass->QImage2cvMat(src);
cv::Point2f center(matSrc.cols / 2, matSrc.rows / 2);
cv::Mat rot = cv::getRotationMatrix2D(center, angle, 1);
cv::Rect bbox = cv::RotatedRect(center, matSrc.size(), angle).boundingRect();
rot.at<double>(0, 2) += bbox.width / 2.0 - center.x;
rot.at<double>(1, 2) += bbox.height / 2.0 - center.y;
cv::warpAffine(matSrc, matDst, rot, bbox.size());
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
QImage Yx_opencv_Geom::Rotate_fixed(QImage src, int angle) // 旋转90，180，270
{
Mat matSrc, matDst, M;
matSrc = imgchangeClass->QImage2cvMat(src);
M = getRotationMatrix2D(Point2i(matSrc.cols / 2, matSrc.rows / 2), angle, 1);
warpAffine(matSrc, matDst, M, Size(matSrc.cols, matSrc.rows));
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
QImage Yx_opencv_Geom::Flip(QImage src, int flipcode) // 镜像
{
Mat matSrc, matDst;
matSrc = imgchangeClass->QImage2cvMat(src);
flip(matSrc, matDst, flipcode); // flipCode==0 垂直翻转（沿X轴翻转）,flipCode>0 水平翻转（沿Y轴翻转）
// flipCode<0 水平垂直翻转（先沿X轴翻转，再沿Y轴翻转，等价于旋转180°）
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
QImage Yx_opencv_Geom::Lean(QImage src, int x, int y) // 倾斜
{
Mat matSrc, matTmp, matDst;
matSrc = imgchangeClass->QImage2cvMat(src);
matTmp = Mat::zeros(matSrc.rows, matSrc.cols, matSrc.type());
Mat map_x, map_y;
Point2f src_point[3], tmp_point[3], x_point[3], y_point[3];
double angleX = x / 180.0 * CV_PI ;
double angleY = y / 180.0 * CV_PI;
src_point[0] = Point2f(0, 0);
src_point[1] = Point2f(matSrc.cols, 0);
src_point[2] = Point2f(0, matSrc.rows);
x_point[0] = Point2f(matSrc.rows * tan(angleX), 0);
x_point[1] = Point2f(matSrc.cols + matSrc.rows * tan(angleX), 0);
x_point[2] = Point2f(0, matSrc.rows);
map_x = getAffineTransform(src_point, x_point);
warpAffine(matSrc, matTmp, map_x, Size(matSrc.cols + matSrc.rows * tan(angleX), matSrc.rows));
tmp_point[0] = Point2f(0, 0);
tmp_point[1] = Point2f(matTmp.cols, 0);
tmp_point[2] = Point2f(0, matTmp.rows);
y_point[0] = Point2f(0, 0);
y_point[1] = Point2f(matTmp.cols, matTmp.cols * tan(angleY));
y_point[2] = Point2f(0, matTmp.rows);
map_y = getAffineTransform(tmp_point, y_point);
warpAffine(matTmp, matDst, map_y, Size(matTmp.cols, matTmp.rows + matTmp.cols * tan(angleY)));
QImage dst = imgchangeClass->cvMat2QImage(matDst);
return dst;
}
opencv图像增强 Yx_opencv_Geom类
qt_opencv.h下Yx_opencv_Geom类
// 图像增强
class Yx_opencv_Gray
{
public:
Yx_opencv_Gray();
~Yx_opencv_Gray();
QImage Bin(QImage src, int threshold);
QImage Graylevel(QImage src);
QImage Reverse(QImage src); // 图像反转
QImage Linear(QImage src, int alpha, int beta); // 线性变换
QImage Gamma(QImage src, int gamma); // 伽马变换(指数变换)
QImage Log(QImage src, int c); // 对数变换
QImage Histeq(QImage src); // 直方图均衡化
private:
Yx_opencv_ImgChange *imgchangeClass;
};
Yx_opencv_Gray::Yx_opencv_Gray()
{
imgchangeClass = new Yx_opencv_ImgChange;
}
Yx_opencv_Gray::~Yx_opencv_Gray()
{
}
QImage Yx_opencv_Gray::Bin(QImage src, int threshold) // 二值化
{
Mat srcImg, dstImg,grayImg;
srcImg = imgchangeClass->QImage2cvMat(src);
if(srcImg.channels()!=1)
cvtColor(srcImg, grayImg, CV_BGR2GRAY);
else
dstImg = srcImg.clone();
cv::threshold(grayImg, dstImg, threshold, 255, THRESH_BINARY);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Graylevel(QImage src) // 灰度图像
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
dstImg.create(srcImg.size(), srcImg.type());
if (srcImg.channels() != 1)
cvtColor(srcImg, dstImg, CV_BGR2GRAY);
else
dstImg = srcImg.clone();
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Reverse(QImage src) // 图像反转
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
bitwise_xor(srcImg, Scalar(255), dstImg); // 异或
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Linear(QImage src, int alpha, int beta) // 线性变换
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
srcImg.convertTo(dstImg, -1, alpha/100.0, beta-100); // matDst = alpha * matTmp + beta
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Gamma(QImage src, int gamma) // 伽马变换(指数变换)
{
if (gamma < 0)
return src;
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat lookUpTable(1, 256, CV_8U); // 查找表
uchar* p = lookUpTable.ptr();
for (int i = 0; i < 256; ++i)
p[i] = saturate_cast<uchar>(pow(i / 255.0, gamma/100.0)*255.0); // pow()是幂次运算
LUT(srcImg, lookUpTable, dstImg); // LUT
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Log(QImage src, int c) // 对数变换
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat lookUpTable(1, 256, CV_8U); // 查找表
uchar* p = lookUpTable.ptr();
for (int i = 0; i < 256; ++i)
p[i] = saturate_cast<uchar>((c/100.0)*log(1 + i / 255.0)*255.0); // pow()是幂次运算
LUT(srcImg, lookUpTable, dstImg); // LUT
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Gray::Histeq(QImage src) // 直方图均衡化
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
if (srcImg.channels() != 1)
cvtColor(srcImg, srcImg, CV_BGR2GRAY);
else
dstImg = srcImg.clone();
equalizeHist(srcImg, dstImg); // 直方图均衡化
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
opencv图像腐蚀 Yx_opencv_Gray类
qt_opencv.h下Yx_opencv_Gray类
// 图像腐蚀
class Yx_opencv_Morp
{
public:
Yx_opencv_Morp();
~Yx_opencv_Morp();
QImage Erode(QImage src, int elem,int kernel,int times,int,int); // 腐蚀
QImage Dilate(QImage src, int elem, int kernel, int times,int,int); // 膨胀
QImage Open(QImage src, int elem, int kernel, int times,int,int); // 开运算
QImage Close(QImage src, int elem, int kernel, int times,int,int); // 闭运算
QImage Grad(QImage src, int elem, int kernel,int,int,int); // 形态学梯度
QImage Tophat(QImage src, int elem, int kernel,int,int,int); // 顶帽操作
QImage Blackhat(QImage src, int elem, int kernel,int,int,int); // 黑帽操作
private:
Yx_opencv_ImgChange *imgchangeClass;
};
Yx_opencv_Morp::Yx_opencv_Morp()
{
imgchangeClass = new Yx_opencv_ImgChange;
}
Yx_opencv_Morp::~Yx_opencv_Morp()
{
}
QImage Yx_opencv_Morp::Erode(QImage src, int elem, int kernel, int times, int, int) // 腐蚀
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
int erosion_type = 0;
if (elem == 0) { erosion_type = MORPH_RECT; }
else if (elem == 1) { erosion_type = MORPH_CROSS; }
else if (elem == 2) { erosion_type = MORPH_ELLIPSE; }
Mat element = getStructuringElement(erosion_type,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel+1, kernel+1));
erode(srcImg, dstImg, element, Point(-1, -1), times);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Dilate(QImage src, int elem, int kernel, int times,int,int) // 膨胀
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
int dilation_type = 0;
if (elem == 0) { dilation_type = MORPH_RECT; }
else if (elem == 1) { dilation_type = MORPH_CROSS; }
else if (elem == 2) { dilation_type = MORPH_ELLIPSE; }
Mat element = getStructuringElement(dilation_type, Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
dilate(srcImg, dstImg, element, Point(-1, -1), times);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Open(QImage src, int elem, int kernel, int times, int, int) // 开运算
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat element = getStructuringElement(elem,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
morphologyEx(srcImg, dstImg, MORPH_OPEN, element, Point(-1, -1), times);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Close(QImage src, int elem, int kernel, int times, int, int) // 闭运算
{
Mat srcImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat element = getStructuringElement(elem,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
morphologyEx(srcImg, dstImg, MORPH_CLOSE, element, Point(-1, -1), times);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Grad(QImage src, int elem, int kernel, int, int, int) // 形态学梯度
{
Mat srcImg, grayImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat element = getStructuringElement(elem,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
if (srcImg.channels() != 1)
cvtColor(srcImg, grayImg, CV_BGR2GRAY);
else
grayImg = srcImg.clone();
morphologyEx(grayImg, dstImg, MORPH_GRADIENT, element);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Tophat(QImage src, int elem, int kernel,int,int,int) // 顶帽操作
{
Mat srcImg, grayImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat element = getStructuringElement(elem,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
if (srcImg.channels() != 1)
cvtColor(srcImg, grayImg, CV_BGR2GRAY);
else
grayImg = srcImg.clone();
morphologyEx(grayImg, dstImg, MORPH_TOPHAT, element);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
QImage Yx_opencv_Morp::Blackhat(QImage src, int elem, int kernel,int,int,int) // 黑帽操作
{
Mat srcImg, grayImg, dstImg;
srcImg = imgchangeClass->QImage2cvMat(src);
Mat element = getStructuringElement(elem,Size(2 * kernel + 3, 2 * kernel + 3), Point(kernel + 1, kernel + 1));
if (srcImg.channels() != 1)
cvtColor(srcImg, grayImg, CV_BGR2GRAY);
else
grayImg = srcImg.clone();
morphologyEx(grayImg, dstImg, MORPH_BLACKHAT, element);
QImage dst = imgchangeClass->cvMat2QImage(dstImg);
return dst;
}
来源：https://blog.csdn.net/a15005784320/article/details/98733765