#include <iostream>
#include "opencv2/opencv.hpp"
using namespace std;
using namespace cv;
生成高斯卷积核 kernel
*/
void Gaussian_kernel(int kernel_size, int sigma, Mat &kernel)
{
const double PI = 3.1415926;
int m = kernel_size / 2;
kernel = Mat(kernel_size, kernel_size, CV_32FC1);
float s = 2 * sigma * sigma;
for (int i = 0; i < kernel_size; i++)
{
for (int j = 0; j < kernel_size; j++)
{
int x = i - m;
int y = j - m;
kernel.at<float>(i, j) = exp(-(x * x + y * y) / s) / (PI * s);
}
}
}
计算梯度值和方向
imageSource 原始灰度图
imageX X方向梯度图像
imageY Y方向梯度图像
gradXY 该点的梯度幅值
theta 梯度方向角度 theta=arctan(imageY/imageX)
*/
void GradDirection(const Mat imageSource, Mat &imageX, Mat &imageY, Mat &gradXY, Mat &theta)
{
imageX = Mat::zeros(imageSource.size(), CV_32SC1);
imageY = Mat::zeros(imageSource.size(), CV_32SC1);
gradXY = Mat::zeros(imageSource.size(), CV_32SC1);
theta = Mat::zeros(imageSource.size(), CV_32SC1);
int rows = imageSource.rows;
int cols = imageSource.cols;
Mat.step参数指图像的一行实际占用的内存长度,以字节为基本单位,
因为opencv中的图像会对每行的长度自动补齐(8的倍数),
编程时尽量使用指针,指针读写像素是速度最快的,使用at函数最慢。
*/
int stepXY = imageX.step;
int step = imageSource.step;
Mat::data的默认类型为uchar*,但很多时候需要处理其它类型,如float、int,
此时需要将data强制类型转换,如:
Mat src(1000,1000,CV_32F);
float* myptr = (float*)src.data;
无论Mat的type是何种类型,Mat::data均为uchar*
*/
uchar *PX = imageX.data;
uchar *PY = imageY.data;
uchar *P = imageSource.data;
uchar *XY = gradXY.data;
for (int i = 1; i < rows - 1; i++)
{
for (int j = 1; j < cols - 1; j++)
{
int a00 = P[(i - 1)*step + j - 1];
int a01 = P[(i - 1)*step + j];
int a02 = P[(i - 1)*step + j + 1];
int a10 = P[i*step + j - 1];
int a11 = P[i*step + j];
int a12 = P[i*step + j + 1];
int a20 = P[(i + 1)*step + j - 1];
int a21 = P[(i + 1)*step + j];
int a22 = P[(i + 1)*step + j + 1];
double gradX = double(a02 + 2 * a12 + a22 - a00 - 2 * a10 - a20);
double gradY = double(a00 + 2 * a01 + a02 - a20 - 2 * a21 - a22);
imageX.at<int>(i, j) = abs(gradX);
imageY.at<int>(i, j) = abs(gradY);
if (gradX == 0)
{
gradX = 0.000000000001;
}
theta.at<int>(i, j) = atan(gradY / gradX) * 57.3;
theta.at<int>(i, j) = (theta.at<int>(i, j) + 360) % 360;
gradXY.at<int>(i, j) = sqrt(gradX*gradX + gradY * gradY);
}
}
在经过处理后,需要用convertScaleAbs()函数将其转回原来的uint8形式,
否则将无法显示图像,而只是一副灰色的窗口。
函数原型为
void convertScaleAbs(InputArray src, OutputArray dst,
double alpha = 1, double beta = 0);
其中可选参数alpha是伸缩系数,beta是加到结果上的一个值,结果返回uint8类型的图片
功能:实现将原图片转换为uint8类型
*/
convertScaleAbs(imageX, imageX);
convertScaleAbs(imageY, imageY);
convertScaleAbs(gradXY, gradXY);
}
局部非极大值抑制
沿着该点梯度方向,比较前后两个点的幅值大小,若该点大于前后两点,则保留,
若该点小于前后两点任意一点,则置为0;
imageInput 输入得到梯度图像
imageOutput 输出的非极大值抑制图像
theta 每个像素点的梯度方向角度
imageX X方向梯度
imageY Y方向梯度
*/
void NonLocalMaxValue(const Mat imageInput, Mat &imageOutput, const Mat &theta, const Mat &imageX, const Mat &imageY)
{
imageOutput = imageInput.clone();
int cols = imageInput.cols;
int rows = imageInput.rows;
for (int i = 1; i < rows - 1; i++)
{
for (int j = 1; j < cols - 1; j++)
{
if (0 == imageInput.at<uchar>(i, j))continue;
int g00 = imageInput.at<uchar>(i - 1, j - 1);
int g01 = imageInput.at<uchar>(i - 1, j);
int g02 = imageInput.at<uchar>(i - 1, j + 1);
int g10 = imageInput.at<uchar>(i, j - 1);
int g11 = imageInput.at<uchar>(i, j);
int g12 = imageInput.at<uchar>(i, j + 1);
int g20 = imageInput.at<uchar>(i + 1, j - 1);
int g21 = imageInput.at<uchar>(i + 1, j);
int g22 = imageInput.at<uchar>(i + 1, j + 1);
int direction = theta.at<int>(i, j);
int g1 = 0;
int g2 = 0;
int g3 = 0;
int g4 = 0;
double tmp1 = 0.0;
double tmp2 = 0.0;
double weight = fabs((double)imageY.at<uchar>(i, j) / (double)imageX.at<uchar>(i, j));
if (weight == 0)
weight = 0.0000001;
关于这些公式的含义
https://www.cnblogs.com/love6tao/p/5152020.html 有解释
形如g10 * (1 - weight) + g20 * (weight)都是用两边的像素计算得到的亚像素
不过他w和1-w应该写反了
*/
if (weight > 1)
{
weight = 1 / weight;
}
if ((0 <= direction && direction < 45) || 180 <= direction && direction < 225)
{
tmp1 = g10 * (1 - weight) + g20 * (weight);
tmp2 = g02 * (weight)+g12 * (1 - weight);
}
if ((45 <= direction && direction < 90) || 225 <= direction && direction < 270)
{
tmp1 = g01 * (1 - weight) + g02 * (weight);
tmp2 = g20 * (weight)+g21 * (1 - weight);
}
if ((90 <= direction && direction < 135) || 270 <= direction && direction < 315)
{
tmp1 = g00 * (weight)+g01 * (1 - weight);
tmp2 = g21 * (1 - weight) + g22 * (weight);
}
if ((135 <= direction && direction < 180) || 315 <= direction && direction < 360)
{
tmp1 = g00 * (weight)+g10 * (1 - weight);
tmp2 = g12 * (1 - weight) + g22 * (weight);
}
if (imageInput.at<uchar>(i, j) < tmp1 || imageInput.at<uchar>(i, j) < tmp2)
{
imageOutput.at<uchar>(i, j) = 0;
}
}
}
}
双阈值的机理是:
指定一个低阈值A,一个高阈值B,选取占直方图总数70%为B,且B为1.5到2倍大小的A;
灰度值小于A的,置为0,灰度值大于B的,置为255;
*/
void DoubleThreshold(Mat &imageInput, const double lowThreshold, const double highThreshold)
{
int cols = imageInput.cols;
int rows = imageInput.rows;
for (int i = 0; i < rows; i++)
{
for (int j = 0; j < cols; j++)
{
double temp = imageInput.at<uchar>(i, j);
temp = temp > highThreshold ? (255) : (temp);
temp = temp < lowThreshold ? (0) : (temp);
imageInput.at<uchar>(i, j) = temp;
}
}
}
连接处理:
灰度值介于A和B之间的,考察该像素点临近的8像素是否有灰度值为255的,
若没有255的,表示这是一个孤立的局部极大值点,予以排除,置为0;
若有255的,表示这是一个跟其他边缘有“接壤”的可造之材,置为255,
之后重复执行该步骤,直到考察完最后一个像素点。
其中的邻域跟踪算法,从值为255的像素点出发找到周围满足要求的点,把满足要求的点设置为255,
然后修改i,j的坐标值,i,j值进行回退,在改变后的i,j基础上继续寻找255周围满足要求的点。
当所有连接255的点修改完后,再把所有上面所说的局部极大值点置为0;(算法可以继续优化)。
参数1,imageInput:输入和输出的梯度图像
参数2,lowTh:低阈值
参数3,highTh:高阈值
*/
void DoubleThresholdLink(Mat &imageInput, double lowTh, double highTh)
{
int cols = imageInput.cols;
int rows = imageInput.rows;
for (int i = 1; i < rows - 1; i++)
{
for (int j = 1; j < cols - 1; j++)
{
double pix = imageInput.at<uchar>(i, j);
if (pix != 255)continue;
bool change = false;
for (int k = -1; k <= 1; k++)
{
for (int u = -1; u <= 1; u++)
{
if (k == 0 && u == 0)continue;
double temp = imageInput.at<uchar>(i + k, j + u);
if (temp >= lowTh && temp <= highTh)
{
imageInput.at<uchar>(i + k, j + u) = 255;
change = true;
}
}
}
if (change)
{
if (i > 1)i--;
if (j > 2)j -= 2;
}
}
}
for (int i = 0; i < rows; i++)
{
for (int j = 0; j < cols; j++)
{
if (imageInput.at<uchar>(i, j) != 255)
{
imageInput.at<uchar>(i, j) = 0;
}
}
}
}
int main()
{
Mat grayImage = imread("I.bmp", 0);
imshow("gray image", grayImage);
Mat gausKernel;
int kernel_size = 5;
double sigma = 1;
Gaussian_kernel(kernel_size, sigma, gausKernel);
Mat gausImage;
filter2D(grayImage, gausImage, grayImage.depth(), gausKernel);
imshow("gaus image", gausImage);
imwrite("高斯滤波后的图像.bmp", gausImage);
Mat imageX, imageY, imageXY;
Mat theta;
GradDirection(gausImage, imageX, imageY, imageXY, theta);
imshow("XY grad", imageXY);
imwrite("梯度模值.bmp", imageXY);
Mat localImage;
NonLocalMaxValue(imageXY, localImage, theta, imageX, imageY);
imshow("Non local maxinum image", localImage);
imwrite("非极大性值抑制.bmp", localImage);
DoubleThreshold(localImage, 60, 100);
DoubleThresholdLink(localImage, 60, 100);
imshow("canny image", localImage);
imwrite("双阈值和边缘连接.bmp", localImage);
Mat temMat;
void Canny( InputArray image, OutputArray edges,
double threshold1, double threshold2,
int apertureSize = 3, bool L2gradient = false );
输入图像 输出图像 低阈值 高阈值 算子大小 是否采用更精确的方式计算图像梯度
void Canny( InputArray dx, InputArray dy,
OutputArray edges,
double threshold1, double threshold2,
bool L2gradient = false );
输入图像x,y方向的导数 输出图像 低阈值 高阈值 算子大小 是否采用更精确的方式计算图像梯度
*/
Canny(grayImage, temMat, 60, 100);
imshow("opencv canny image", temMat);
imwrite("opencv自带的canny算法得到的效果图.bmp", temMat);
waitKey(0);
return 0;
}