1 | initial version |
I have read@miki answer in this post and disable CV_OCL_RUN and CV_IPP_RUN in my own copy of opencv canny. results is here :
#include<opencv2/opencv.hpp>
#include "opencv2/core/ocl.hpp"
#include<iostream>
using namespace cv;
using namespace cv;
/*
Using Canny's Criteria to Derive a Recursively Implemented Optimal Edge Detector International Journal of Computer Vision,167-187 (1987)
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.476.5736&rep=rep1&type=pdf
http://www.esiee.fr/~coupriem/Algo/algoima.html
*/
class ParallelGradientDericheYCols: public ParallelLoopBody
{
private:
Mat &img;
Mat &im1;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheYCols(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
im1(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
long i,j,nb;
long tailleSequence=(img.rows>img.cols)?img.rows:img.cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a1=1,a2=0;
double a3=0,a4=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
int rows=img.rows,cols=img.cols;
kp=pow(1-exp(-p),2.0)/exp(-p);
a1=0,a2=kp*exp(-p),a3=-kp*exp(-p),a4=0;
b1=2*exp(-p);
b2=-exp(-2*p);
switch(img.depth()){
case CV_8U :
{
unsigned char *c1;
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
// Filtre causal vertical
c1 = (unsigned char*)img.ptr(0)+nb;
f2 = (float*)im1.ptr(0)+nb;
f2 += j;
c1+=j;
i=0;
g1[i] = (a1 + a2 +b1+b2)* *c1 ;
g1[i] = (a1 + a2 )* *c1 ;
i++;
c1+=cols;
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1+b2) * g1[i-1];
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1) * g1[i-1];
i++;
c1+=cols;
for (i=2;i<rows;i++,c1+=cols)
g1[i] = a1 * *c1 + a2 * c1[-cols] +b1*g1[i-1]+
b2 *g1[i-2];
// Filtre anticausal vertical
c1 = (unsigned char*)img.ptr(0)+nb;
c1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a3+a4+b1+b2)* *c1;
g2[i] =(a3+a4)* *c1;
i--;
c1-=cols;
g2[i] = a3* c1[cols] + a4 * c1[cols]+(b1+b2)*g2[i+1];
g2[i] = a3* c1[cols] + a4 * c1[cols]+(b1)*g2[i+1];
i--;
c1-=cols;
for (i=rows-3;i>=0;i--,c1-=cols)
g2[i] = a3*c1[cols] +a4* c1[2*cols]+
b1*g2[i+1]+b2*g2[i+2];
for (i=0;i<rows;i++,f2+=cols)
*f2 = (float)(g1[i]+g2[i]);
}
}
}
break;
case CV_16S :
case CV_16U :
{
unsigned short *c1;
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
c1 = ((unsigned short*)img.ptr(0))+nb;
f2 = ((float*)im1.ptr(0))+nb;
f2 += j;
c1+=j;
i=0;
g1[i] = (a1 + a2 +b1+b2)* *c1 ;
g1[i] = (a1 + a2 )* *c1 ;
i++;
c1+=cols;
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1+b2) * g1[i-1];
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1) * g1[i-1];
i++;
c1+=cols;
for (i=2;i<rows;i++,c1+=cols)
g1[i] = a1 * *c1 + a2 * c1[-cols] +b1*g1[i-1]+
b2 *g1[i-2];
// Filtre anticausal vertical
c1 = ((unsigned short*)img.ptr(0))+nb;
c1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a3+a4+b1+b2)* *c1;
g2[i] =(a3+a4)* *c1;
i--;
c1-=cols;
g2[i] = (a3+a4)* c1[cols] +(b1+b2)*g2[i+1];
g2[i] = (a3+a4)* c1[cols] +(b1)*g2[i+1];
i--;
c1-=cols;
for (i=rows-3;i>=0;i--,c1-=cols)
g2[i] = a3*c1[cols] +a4* c1[2*cols]+
b1*g2[i+1]+b2*g2[i+2];
c1 = ((unsigned short*)img.ptr(0))+nb+j;
for (i=0;i<rows;i++,f2+=cols,c1+=cols)
*f2 = 0**c1+(float)(g1[i]+g2[i]);
}
}
}
break;
case CV_32S :
break;
case CV_32F :
break;
case CV_64F :
break;
default :
return ;
}
delete g1;
delete g2;
};
ParallelGradientDericheYCols& operator=(const ParallelGradientDericheYCols &) {
return *this;
};
};
class ParallelGradientDericheYRows: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheYRows(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
long i,j,nb;
long tailleSequence=(img.rows>img.cols)?img.rows:img.cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
int rows=img.rows,cols=img.cols;
// for(int y = range.start; y < range.end; y++)
p=alphaMoyenne;
k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
a5=k,a6=k*exp(-p)*(p-1);
a7=k*exp(-p)*(p+1),a8=-k*exp(-2*p);
b3=2*exp(-p);
b4=-exp(-2*p);
for (nb=0;nb<img.channels();nb++)
{
for (i=range.start;i<range.end;i++)
{
f2 = ((float*)dst.ptr(i))+nb;;
f1 = ((float*)img.ptr(i))+nb;
j=0;
g1[j] = (a5 +a6+b3+b4)* *f1 ;
g1[j] = (a5 +a6)* *f1 ;
j++;
f1++;
g1[j] = a5 * f1[0]+a6*f1[j-1]+(b3+b4) * g1[j-1];
g1[j] = a5 * f1[0]+a6*f1[j-1]+(b3) * g1[j-1];
j++;
f1++;
for (j=2;j<cols;j++,f1++)
g1[j] = a5 * f1[0] + a6 * f1[-1]+b3*g1[j-1]+b4*g1[j-2];
f1 = ((float*)img.ptr(0))+nb;;
f1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a7+a8+b3+b4)* *f1;
g2[j] = (a7+a8)* *f1;
j--;
f1--;
g2[j] = (a7+a8) * f1[1] +(b3+b4) * g2[j+1];
g2[j] = (a7+a8) * f1[1] +(b3) * g2[j+1];
j--;
f1--;
for (j=cols-3;j>=0;j--,f1--)
g2[j] = a7*f1[1]+a8*f1[2]+b3*g2[j+1]+b4*g2[j+2];
for (j=0;j<cols;j++,f2++)
*f2 = (float)(g1[j]+g2[j]);
}
}
delete g1;
delete g2;
};
ParallelGradientDericheYRows& operator=(const ParallelGradientDericheYRows &) {
return *this;
};
};
class ParallelGradientDericheXCols: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheXCols(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
int rows=img.rows,cols=img.cols;
long i,j,nb;
long tailleSequence=(rows>cols)?rows:cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
// double a1=1,a2=0;
// double a3=0,a4=0;
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
p=alphaMoyenne;
k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
a5=k,a6=k*exp(-p)*(p-1);
a7=k*exp(-p)*(p+1),a8=-k*exp(-2*p);
b3=2*exp(-p);
b4=-exp(-2*p);
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
f1 = (float*)img.ptr(0)+(nb);
f1+=j;
i=0;
g1[i] = (a5 + a6 +b3+b4)* *f1 ;
g1[i] = (a5 + a6 )* *f1 ;
i++;
f1+=cols;
g1[i] = a5 * *f1 + a6 * f1[-cols] + (b3+b4) * g1[i-1];
g1[i] = a5 * *f1 + a6 * f1[-cols] + (b3) * g1[i-1];
i++;
f1+=cols;
for (i=2;i<rows;i++,f1+=cols)
g1[i] = a5 * *f1 + a6 * f1[-cols] +b3*g1[i-1]+
b4 *g1[i-2];
// Filtre anticausal vertical
f1 = (float*)img.ptr(0)+(nb);
f1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a7+a8+b3+b4)* *f1;
g2[i] =(a7+a8)* *f1;
i--;
f1-=cols;
g2[i] = (a7+a8)* f1[cols] +(b3+b4)*g2[i+1];
g2[i] = (a7+a8)* f1[cols] +(b3)*g2[i+1];
i--;
f1-=cols;
for (i=rows-3;i>=0;i--,f1-=cols)
g2[i] = a7*f1[cols] +a8* f1[2*cols]+
b3*g2[i+1]+b4*g2[i+2];
for (i=0;i<rows;i++,f2+=cols)
{
f2 = ((float*)dst.ptr(i))+(j*img.channels());
*f2 = (float)(g1[i]+g2[i]);
}
}
}
delete g1;
delete g2;
};
ParallelGradientDericheXCols& operator=(const ParallelGradientDericheXCols &) {
return *this;
};
};
class ParallelGradientDericheXRows: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheXRows(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
int rows=img.rows,cols=img.cols;
long i,j,nb;
long tailleSequence=(rows>cols)?rows:cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a1=1,a2=0;
double a3=0,a4=0;
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
kp=pow(1-exp(-p),2.0)/exp(-p);
a1=0,a2=kp*exp(-p),a3=-kp*exp(-p),a4=0;
b1=2*exp(-p);
b2=-exp(-2*p);
switch(img.depth()){
case CV_8U :
case CV_8S :
{
unsigned char *c1;
for (nb=0;nb<img.channels();nb++)
{
for (i=range.start;i<range.end;i++)
{
f1 = (float*)dst.ptr(i)+nb;
c1 = (unsigned char*)img.ptr(i)+(nb);
j=0;
g1[j] = (a1 +a2+b1+b2)* *c1 ;
g1[j] = (a1 +a2)* *c1 ;
j++;
c1++;
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1+b2) * g1[j-1];
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1) * g1[j-1];
j++;
c1++;
for (j=2;j<cols;j++,c1++)
g1[j] = a1 * c1[0] + a2 * c1[-1]+b1*g1[j-1]+b2*g1[j-2];
c1 = (unsigned char*)img.ptr(0)+(nb);
c1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a3+a4+b1+b2)* *c1;
g2[j] = (a3+a4)* *c1;
j--;
g2[j] = (a3+a4) * c1[1] +(b1+b2) * g2[j+1];
g2[j] = (a3+a4) * c1[1] +(b1) * g2[j+1];
j--;
c1--;
for (j=cols-3;j>=0;j--,c1--)
g2[j] = a3*c1[1]+a4*c1[2]+b1*g2[j+1]+b2*g2[j+2];
for (j=0;j<cols;j++,f1++)
*f1 = (float)(g1[j]+g2[j]);
}
}
}
break;
case CV_16S :
case CV_16U :
{
unsigned short *c1;
for (nb=0;nb<img.channels();nb++)
{
f1 = ((float*)dst.ptr(0))+nb;
for (i=range.start;i<range.end;i++)
{
c1 = ((unsigned short*)img.ptr(0))+(nb);
c1 += i*cols;
j=0;
g1[j] = (a1 +a2+b1+b2)* *c1 ;
g1[j] = (a1 +a2)* *c1 ;
j++;
c1++;
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1+b2) * g1[j-1];
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1) * g1[j-1];
j++;
c1++;
for (j=2;j<cols;j++,c1++)
g1[j] = a1 * c1[0] + a2 * c1[-1]+b1*g1[j-1]+b2*g1[j-2];
c1 = ((unsigned short*)img.ptr(0))+(nb);
c1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a3+a4+b1+b2)* *c1;
g2[j] = (a3+a4)* *c1;
j--;
c1--;
g2[j] = (a3+a4) * c1[1] +(b1+b2) * g2[j+1];
g2[j] = (a3+a4) * c1[1] +(b1) * g2[j+1];
j--;
c1--;
for (j=cols-3;j>=0;j--,c1--)
g2[j] = a3*c1[1]+a4*c1[2]+b1*g2[j+1]+b2*g2[j+2];
for (j=0;j<cols;j++,f1++)
*f1 = (float)(g1[j]+g2[j]);
}
}
}
break;
default :
return ;
}
delete g1;
delete g2;
};
ParallelGradientDericheXRows& operator=(const ParallelGradientDericheXRows &) {
return *this;
};
};
UMat GradientDericheY(UMat op, double alphaDerive,double alphaMoyenne)
{
Mat im1(op.size(),CV_32FC(op.channels()));
UMat imRes(op.rows,op.cols,CV_32FC(op.channels()));
cv::Mat mThis = op.getMat(cv::ACCESS_RW);
cv::Mat m2 = imRes.getMat(cv::ACCESS_RW);
ParallelGradientDericheYCols x(mThis,im1,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.cols), x,getNumThreads());
ParallelGradientDericheYRows xr(im1,m2,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.rows), xr,getNumThreads());
return imRes;
}
UMat GradientDericheX(UMat op, double alphaDerive,double alphaMoyenne)
{
Mat im1(op.size(),CV_32FC(op.channels()));
UMat imRes(op.rows,op.cols,CV_32FC(op.channels()));
cv::Mat mThis = op.getMat(cv::ACCESS_RW);
cv::Mat m2 = imRes.getMat(cv::ACCESS_RW);
ParallelGradientDericheXRows x(mThis,im1,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.rows), x,getNumThreads());
ParallelGradientDericheXCols xr(im1,m2,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.cols), xr,getNumThreads());
return imRes;
}
void CannyBis( InputArray _src, OutputArray _dst,
double low_thresh, double high_thresh,
int aperture_size, bool L2gradient ,InputOutputArray _dx,InputOutputArray _dy)
{
const int type = _src.type(), depth = CV_MAT_DEPTH(type), cn = CV_MAT_CN(type);
const Size size = _src.size();
CV_Assert( depth == CV_8U );
_dst.create(size, CV_8U);
if (!L2gradient && (aperture_size & CV_CANNY_L2_GRADIENT) == CV_CANNY_L2_GRADIENT)
{
// backward compatibility
aperture_size &= ~CV_CANNY_L2_GRADIENT;
L2gradient = true;
}
if ((aperture_size & 1) == 0 || (aperture_size != -1 && (aperture_size < 3 || aperture_size > 7)))
CV_Error(CV_StsBadFlag, "Aperture size should be odd");
if (low_thresh > high_thresh)
std::swap(low_thresh, high_thresh);
/* CV_OCL_RUN(_dst.isUMat() && (cn == 1 || cn == 3),
ocl_Canny(_src, _dst, (float)low_thresh, (float)high_thresh, aperture_size, L2gradient, cn, size))*/
Mat src = _src.getMat(), dst = _dst.getMat();
#ifdef HAVE_TEGRA_OPTIMIZATION
if (tegra::useTegra() && tegra::canny(src, dst, low_thresh, high_thresh, aperture_size, L2gradient))
return;
#endif
// CV_IPP_RUN(USE_IPP_CANNY && (aperture_size == 3 && !L2gradient && 1 == cn), ippCanny(src, dst, (float)low_thresh, (float)high_thresh))
#ifdef HAVE_TBB
if (L2gradient)
{
low_thresh = std::min(32767.0, low_thresh);
high_thresh = std::min(32767.0, high_thresh);
if (low_thresh > 0) low_thresh *= low_thresh;
if (high_thresh > 0) high_thresh *= high_thresh;
}
int low = cvFloor(low_thresh);
int high = cvFloor(high_thresh);
ptrdiff_t mapstep = src.cols + 2;
AutoBuffer<uchar> buffer((src.cols+2)*(src.rows+2));
uchar* map = (uchar*)buffer;
memset(map, 1, mapstep);
memset(map + mapstep*(src.rows + 1), 1, mapstep);
int threadsNumber = tbb::task_scheduler_init::default_num_threads();
int grainSize = src.rows / threadsNumber;
// Make a fallback for pictures with too few rows.
uchar ksize2 = aperture_size / 2;
int minGrainSize = 1 + ksize2;
int maxGrainSize = src.rows - 2 - 2*ksize2;
if ( !( minGrainSize <= grainSize && grainSize <= maxGrainSize ) )
{
threadsNumber = 1;
grainSize = src.rows;
}
tbb::task_group g;
for (int i = 0; i < threadsNumber; ++i)
{
if (i < threadsNumber - 1)
g.run(tbbCanny(Range(i * grainSize, (i + 1) * grainSize), src, map, low, high, aperture_size, L2gradient));
else
g.run(tbbCanny(Range(i * grainSize, src.rows), src, map, low, high, aperture_size, L2gradient));
}
g.wait();
#define CANNY_PUSH_SERIAL(d) *(d) = uchar(2), borderPeaks.push(d)
// now track the edges (hysteresis thresholding)
uchar* m;
while (borderPeaks.try_pop(m))
{
if (!m[-1]) CANNY_PUSH_SERIAL(m - 1);
if (!m[1]) CANNY_PUSH_SERIAL(m + 1);
if (!m[-mapstep-1]) CANNY_PUSH_SERIAL(m - mapstep - 1);
if (!m[-mapstep]) CANNY_PUSH_SERIAL(m - mapstep);
if (!m[-mapstep+1]) CANNY_PUSH_SERIAL(m - mapstep + 1);
if (!m[mapstep-1]) CANNY_PUSH_SERIAL(m + mapstep - 1);
if (!m[mapstep]) CANNY_PUSH_SERIAL(m + mapstep);
if (!m[mapstep+1]) CANNY_PUSH_SERIAL(m + mapstep + 1);
}
#else
Mat dx,dy;
if (_dx.empty())
{
Sobel(src, dx, CV_16S, 1, 0, aperture_size, 1, 0, BORDER_REPLICATE);
Sobel(src, dy, CV_16S, 0, 1, aperture_size, 1, 0, BORDER_REPLICATE);
}
else
{
dx = _dx.getMat(), dy = _dy.getMat();
}
if (L2gradient)
{
low_thresh = std::min(32767.0, low_thresh);
high_thresh = std::min(32767.0, high_thresh);
if (low_thresh > 0) low_thresh *= low_thresh;
if (high_thresh > 0) high_thresh *= high_thresh;
}
int low = cvFloor(low_thresh);
int high = cvFloor(high_thresh);
ptrdiff_t mapstep = src.cols + 2;
AutoBuffer<uchar> buffer((src.cols+2)*(src.rows+2) + cn * mapstep * 3 * sizeof(int));
int* mag_buf[3];
mag_buf[0] = (int*)(uchar*)buffer;
mag_buf[1] = mag_buf[0] + mapstep*cn;
mag_buf[2] = mag_buf[1] + mapstep*cn;
memset(mag_buf[0], 0, /* cn* */mapstep*sizeof(int));
uchar* map = (uchar*)(mag_buf[2] + mapstep*cn);
memset(map, 1, mapstep);
memset(map + mapstep*(src.rows + 1), 1, mapstep);
int maxsize = std::max(1 << 10, src.cols * src.rows / 10);
std::vector<uchar*> stack(maxsize);
uchar **stack_top = &stack[0];
uchar **stack_bottom = &stack[0];
/* sector numbers
(Top-Left Origin)
1 2 3
* * *
* * *
0*******0
* * *
* * *
3 2 1
*/
#define CANNY_PUSH(d) *(d) = uchar(2), *stack_top++ = (d)
#define CANNY_POP(d) (d) = *--stack_top
#if CV_SSE2
bool haveSSE2 = checkHardwareSupport(CV_CPU_SSE2);
#endif
// calculate magnitude and angle of gradient, perform non-maxima suppression.
// fill the map with one of the following values:
// 0 - the pixel might belong to an edge
// 1 - the pixel can not belong to an edge
// 2 - the pixel does belong to an edge
for (int i = 0; i <= src.rows; i++)
{
int* _norm = mag_buf[(i > 0) + 1] + 1;
if (i < src.rows)
{
short* _dx = dx.ptr<short>(i);
short* _dy = dy.ptr<short>(i);
if (!L2gradient)
{
int j = 0, width = src.cols * cn;
#if CV_SSE2
if (haveSSE2)
{
__m128i v_zero = _mm_setzero_si128();
for ( ; j <= width - 8; j += 8)
{
__m128i v_dx = _mm_loadu_si128((const __m128i *)(_dx + j));
__m128i v_dy = _mm_loadu_si128((const __m128i *)(_dy + j));
v_dx = _mm_max_epi16(v_dx, _mm_sub_epi16(v_zero, v_dx));
v_dy = _mm_max_epi16(v_dy, _mm_sub_epi16(v_zero, v_dy));
__m128i v_norm = _mm_add_epi32(_mm_unpacklo_epi16(v_dx, v_zero), _mm_unpacklo_epi16(v_dy, v_zero));
_mm_storeu_si128((__m128i *)(_norm + j), v_norm);
v_norm = _mm_add_epi32(_mm_unpackhi_epi16(v_dx, v_zero), _mm_unpackhi_epi16(v_dy, v_zero));
_mm_storeu_si128((__m128i *)(_norm + j + 4), v_norm);
}
}
#elif CV_NEON
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
vst1q_s32(_norm + j, vaddq_s32(vabsq_s32(vmovl_s16(vget_low_s16(v_dx))),
vabsq_s32(vmovl_s16(vget_low_s16(v_dy)))));
vst1q_s32(_norm + j + 4, vaddq_s32(vabsq_s32(vmovl_s16(vget_high_s16(v_dx))),
vabsq_s32(vmovl_s16(vget_high_s16(v_dy)))));
}
#endif
for ( ; j < width; ++j)
_norm[j] = std::abs(int(_dx[j])) + std::abs(int(_dy[j]));
}
else
{
int j = 0, width = src.cols * cn;
#if CV_SSE2
if (haveSSE2)
{
for ( ; j <= width - 8; j += 8)
{
__m128i v_dx = _mm_loadu_si128((const __m128i *)(_dx + j));
__m128i v_dy = _mm_loadu_si128((const __m128i *)(_dy + j));
__m128i v_dx_ml = _mm_mullo_epi16(v_dx, v_dx), v_dx_mh = _mm_mulhi_epi16(v_dx, v_dx);
__m128i v_dy_ml = _mm_mullo_epi16(v_dy, v_dy), v_dy_mh = _mm_mulhi_epi16(v_dy, v_dy);
__m128i v_norm = _mm_add_epi32(_mm_unpacklo_epi16(v_dx_ml, v_dx_mh), _mm_unpacklo_epi16(v_dy_ml, v_dy_mh));
_mm_storeu_si128((__m128i *)(_norm + j), v_norm);
v_norm = _mm_add_epi32(_mm_unpackhi_epi16(v_dx_ml, v_dx_mh), _mm_unpackhi_epi16(v_dy_ml, v_dy_mh));
_mm_storeu_si128((__m128i *)(_norm + j + 4), v_norm);
}
}
#elif CV_NEON
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
int16x4_t v_dxp = vget_low_s16(v_dx), v_dyp = vget_low_s16(v_dy);
int32x4_t v_dst = vmlal_s16(vmull_s16(v_dxp, v_dxp), v_dyp, v_dyp);
vst1q_s32(_norm + j, v_dst);
v_dxp = vget_high_s16(v_dx), v_dyp = vget_high_s16(v_dy);
v_dst = vmlal_s16(vmull_s16(v_dxp, v_dxp), v_dyp, v_dyp);
vst1q_s32(_norm + j + 4, v_dst);
}
#endif
for ( ; j < width; ++j)
_norm[j] = int(_dx[j])*_dx[j] + int(_dy[j])*_dy[j];
}
if (cn > 1)
{
for(int j = 0, jn = 0; j < src.cols; ++j, jn += cn)
{
int maxIdx = jn;
for(int k = 1; k < cn; ++k)
if(_norm[jn + k] > _norm[maxIdx]) maxIdx = jn + k;
_norm[j] = _norm[maxIdx];
_dx[j] = _dx[maxIdx];
_dy[j] = _dy[maxIdx];
}
}
_norm[-1] = _norm[src.cols] = 0;
}
else
memset(_norm-1, 0, /* cn* */mapstep*sizeof(int));
// at the very beginning we do not have a complete ring
// buffer of 3 magnitude rows for non-maxima suppression
if (i == 0)
continue;
uchar* _map = map + mapstep*i + 1;
_map[-1] = _map[src.cols] = 1;
int* _mag = mag_buf[1] + 1; // take the central row
ptrdiff_t magstep1 = mag_buf[2] - mag_buf[1];
ptrdiff_t magstep2 = mag_buf[0] - mag_buf[1];
const short* _x = dx.ptr<short>(i-1);
const short* _y = dy.ptr<short>(i-1);
if ((stack_top - stack_bottom) + src.cols > maxsize)
{
int sz = (int)(stack_top - stack_bottom);
maxsize = std::max(maxsize * 3/2, sz + src.cols);
stack.resize(maxsize);
stack_bottom = &stack[0];
stack_top = stack_bottom + sz;
}
int prev_flag = 0;
for (int j = 0; j < src.cols; j++)
{
#define CANNY_SHIFT 15
const int TG22 = (int)(0.4142135623730950488016887242097*(1<<CANNY_SHIFT) + 0.5);
int m = _mag[j];
if (m > low)
{
int xs = _x[j];
int ys = _y[j];
int x = std::abs(xs);
int y = std::abs(ys) << CANNY_SHIFT;
int tg22x = x * TG22;
if (y < tg22x)
{
if (m > _mag[j-1] && m >= _mag[j+1]) goto __ocv_canny_push;
}
else
{
int tg67x = tg22x + (x << (CANNY_SHIFT+1));
if (y > tg67x)
{
if (m > _mag[j+magstep2] && m >= _mag[j+magstep1]) goto __ocv_canny_push;
}
else
{
int s = (xs ^ ys) < 0 ? -1 : 1;
if (m > _mag[j+magstep2-s] && m > _mag[j+magstep1+s]) goto __ocv_canny_push;
}
}
}
prev_flag = 0;
_map[j] = uchar(1);
continue;
__ocv_canny_push:
if (!prev_flag && m > high && _map[j-mapstep] != 2)
{
CANNY_PUSH(_map + j);
prev_flag = 1;
}
else
_map[j] = 0;
}
// scroll the ring buffer
_mag = mag_buf[0];
mag_buf[0] = mag_buf[1];
mag_buf[1] = mag_buf[2];
mag_buf[2] = _mag;
}
// now track the edges (hysteresis thresholding)
while (stack_top > stack_bottom)
{
uchar* m;
if ((stack_top - stack_bottom) + 8 > maxsize)
{
int sz = (int)(stack_top - stack_bottom);
maxsize = maxsize * 3/2;
stack.resize(maxsize);
stack_bottom = &stack[0];
stack_top = stack_bottom + sz;
}
CANNY_POP(m);
if (!m[-1]) CANNY_PUSH(m - 1);
if (!m[1]) CANNY_PUSH(m + 1);
if (!m[-mapstep-1]) CANNY_PUSH(m - mapstep - 1);
if (!m[-mapstep]) CANNY_PUSH(m - mapstep);
if (!m[-mapstep+1]) CANNY_PUSH(m - mapstep + 1);
if (!m[mapstep-1]) CANNY_PUSH(m + mapstep - 1);
if (!m[mapstep]) CANNY_PUSH(m + mapstep);
if (!m[mapstep+1]) CANNY_PUSH(m + mapstep + 1);
}
#endif
// the final pass, form the final image
const uchar* pmap = map + mapstep + 1;
uchar* pdst = dst.ptr();
for (int i = 0; i < src.rows; i++, pmap += mapstep, pdst += dst.step)
{
for (int j = 0; j < src.cols; j++)
pdst[j] = (uchar)-(pmap[j] >> 1);
}
}
void cvCannyBis( const CvArr* image, CvArr* edges, double threshold1,
double threshold2, int aperture_size )
{
cv::Mat src = cv::cvarrToMat(image), dst = cv::cvarrToMat(edges);
CV_Assert( src.size == dst.size && src.depth() == CV_8U && dst.type() == CV_8U );
Canny(src, dst, threshold1, threshold2, aperture_size & 255,
(aperture_size & CV_CANNY_L2_GRADIENT) != 0);
}
int lowThreshold=20;
int maxThreshold=20;
int const max_lowThreshold = 500;
int alDerive=100;
int alMean=100;
Mat sobel_x, sobel_y;
UMat img;
const char* window_name = "Edge Map";
/**
* @function CannyThreshold
* @brief Trackbar callback - Canny thresholds input with a ratio 1:3
*/
static void CannyThreshold(int, void*)
{
Mat dst;
double d=alDerive/100.0,m=alMean/100.0;
UMat rx= GradientDericheX(img,d,m);
UMat ry= GradientDericheY(img,d,m);
double minv,maxv;
minMaxLoc(rx,&minv,&maxv);
std::cout << minv << "\t" << maxv << "\\t";
minMaxLoc(ry,&minv,&maxv);
std::cout << minv << " " << maxv << "\n";
Mat mm;
mm=abs(rx.getMat(ACCESS_READ));
rx.getMat(ACCESS_READ).convertTo(sobel_x,CV_16S,1);
mm=abs(ry.getMat(ACCESS_READ));ry.getMat(ACCESS_READ).convertTo(sobel_y,CV_16S,1);
minMaxLoc(sobel_x,&minv,&maxv);
std::cout << minv << "\t" << maxv << "\\t";
minMaxLoc(sobel_y,&minv,&maxv);
std::cout << minv << " " << maxv << "\n";
CannyBis(img, dst,lowThreshold, maxThreshold, 3,false,sobel_x,sobel_y);
imshow("edges x", sobel_x.mul(50));
imshow("edges y", sobel_y.mul(50));
imshow( window_name, dst );
}
int main()
{
ocl::setUseOpenCL(false);
//imread("f:/lib/opencv/samples/data/lena.jpg",IMREAD_GRAYSCALE).copyTo(img);
imread("c:/Users/Laurent.PC-LAURENT-VISI/Downloads/F1A_glider_hook.jpg",IMREAD_GRAYSCALE).copyTo(img);
imshow("Original",img);
namedWindow( window_name, WINDOW_AUTOSIZE );
/// Create a Trackbar for user to enter threshold
createTrackbar( "Min Threshold:",window_name, &lowThreshold, max_lowThreshold, CannyThreshold );
createTrackbar( "Max Threshold:", window_name, &maxThreshold, max_lowThreshold, CannyThreshold );
createTrackbar( "Derive:",window_name, &alDerive, 400, CannyThreshold );
createTrackbar( "Mean:", window_name, &alMean, 400, CannyThreshold );
waitKey();
return 0;
}
2 | No.2 Revision |
I have readread @miki answer in this post and disable CV_OCL_RUN and CV_IPP_RUN in my own copy of opencv canny. results is here :
#include<opencv2/opencv.hpp>
#include "opencv2/core/ocl.hpp"
#include<iostream>
using namespace cv;
using namespace cv;
/*
Using Canny's Criteria to Derive a Recursively Implemented Optimal Edge Detector International Journal of Computer Vision,167-187 (1987)
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.476.5736&rep=rep1&type=pdf
http://www.esiee.fr/~coupriem/Algo/algoima.html
*/
class ParallelGradientDericheYCols: public ParallelLoopBody
{
private:
Mat &img;
Mat &im1;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheYCols(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
im1(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
long i,j,nb;
long tailleSequence=(img.rows>img.cols)?img.rows:img.cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a1=1,a2=0;
double a3=0,a4=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
int rows=img.rows,cols=img.cols;
kp=pow(1-exp(-p),2.0)/exp(-p);
a1=0,a2=kp*exp(-p),a3=-kp*exp(-p),a4=0;
b1=2*exp(-p);
b2=-exp(-2*p);
switch(img.depth()){
case CV_8U :
{
unsigned char *c1;
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
// Filtre causal vertical
c1 = (unsigned char*)img.ptr(0)+nb;
f2 = (float*)im1.ptr(0)+nb;
f2 += j;
c1+=j;
i=0;
g1[i] = (a1 + a2 +b1+b2)* *c1 ;
g1[i] = (a1 + a2 )* *c1 ;
i++;
c1+=cols;
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1+b2) * g1[i-1];
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1) * g1[i-1];
i++;
c1+=cols;
for (i=2;i<rows;i++,c1+=cols)
g1[i] = a1 * *c1 + a2 * c1[-cols] +b1*g1[i-1]+
b2 *g1[i-2];
// Filtre anticausal vertical
c1 = (unsigned char*)img.ptr(0)+nb;
c1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a3+a4+b1+b2)* *c1;
g2[i] =(a3+a4)* *c1;
i--;
c1-=cols;
g2[i] = a3* c1[cols] + a4 * c1[cols]+(b1+b2)*g2[i+1];
g2[i] = a3* c1[cols] + a4 * c1[cols]+(b1)*g2[i+1];
i--;
c1-=cols;
for (i=rows-3;i>=0;i--,c1-=cols)
g2[i] = a3*c1[cols] +a4* c1[2*cols]+
b1*g2[i+1]+b2*g2[i+2];
for (i=0;i<rows;i++,f2+=cols)
*f2 = (float)(g1[i]+g2[i]);
}
}
}
break;
case CV_16S :
case CV_16U :
{
unsigned short *c1;
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
c1 = ((unsigned short*)img.ptr(0))+nb;
f2 = ((float*)im1.ptr(0))+nb;
f2 += j;
c1+=j;
i=0;
g1[i] = (a1 + a2 +b1+b2)* *c1 ;
g1[i] = (a1 + a2 )* *c1 ;
i++;
c1+=cols;
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1+b2) * g1[i-1];
g1[i] = a1 * *c1 + a2 * c1[-cols] + (b1) * g1[i-1];
i++;
c1+=cols;
for (i=2;i<rows;i++,c1+=cols)
g1[i] = a1 * *c1 + a2 * c1[-cols] +b1*g1[i-1]+
b2 *g1[i-2];
// Filtre anticausal vertical
c1 = ((unsigned short*)img.ptr(0))+nb;
c1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a3+a4+b1+b2)* *c1;
g2[i] =(a3+a4)* *c1;
i--;
c1-=cols;
g2[i] = (a3+a4)* c1[cols] +(b1+b2)*g2[i+1];
g2[i] = (a3+a4)* c1[cols] +(b1)*g2[i+1];
i--;
c1-=cols;
for (i=rows-3;i>=0;i--,c1-=cols)
g2[i] = a3*c1[cols] +a4* c1[2*cols]+
b1*g2[i+1]+b2*g2[i+2];
c1 = ((unsigned short*)img.ptr(0))+nb+j;
for (i=0;i<rows;i++,f2+=cols,c1+=cols)
*f2 = 0**c1+(float)(g1[i]+g2[i]);
}
}
}
break;
case CV_32S :
break;
case CV_32F :
break;
case CV_64F :
break;
default :
return ;
}
delete g1;
delete g2;
};
ParallelGradientDericheYCols& operator=(const ParallelGradientDericheYCols &) {
return *this;
};
};
class ParallelGradientDericheYRows: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheYRows(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
long i,j,nb;
long tailleSequence=(img.rows>img.cols)?img.rows:img.cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
int rows=img.rows,cols=img.cols;
// for(int y = range.start; y < range.end; y++)
p=alphaMoyenne;
k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
a5=k,a6=k*exp(-p)*(p-1);
a7=k*exp(-p)*(p+1),a8=-k*exp(-2*p);
b3=2*exp(-p);
b4=-exp(-2*p);
for (nb=0;nb<img.channels();nb++)
{
for (i=range.start;i<range.end;i++)
{
f2 = ((float*)dst.ptr(i))+nb;;
f1 = ((float*)img.ptr(i))+nb;
j=0;
g1[j] = (a5 +a6+b3+b4)* *f1 ;
g1[j] = (a5 +a6)* *f1 ;
j++;
f1++;
g1[j] = a5 * f1[0]+a6*f1[j-1]+(b3+b4) * g1[j-1];
g1[j] = a5 * f1[0]+a6*f1[j-1]+(b3) * g1[j-1];
j++;
f1++;
for (j=2;j<cols;j++,f1++)
g1[j] = a5 * f1[0] + a6 * f1[-1]+b3*g1[j-1]+b4*g1[j-2];
f1 = ((float*)img.ptr(0))+nb;;
f1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a7+a8+b3+b4)* *f1;
g2[j] = (a7+a8)* *f1;
j--;
f1--;
g2[j] = (a7+a8) * f1[1] +(b3+b4) * g2[j+1];
g2[j] = (a7+a8) * f1[1] +(b3) * g2[j+1];
j--;
f1--;
for (j=cols-3;j>=0;j--,f1--)
g2[j] = a7*f1[1]+a8*f1[2]+b3*g2[j+1]+b4*g2[j+2];
for (j=0;j<cols;j++,f2++)
*f2 = (float)(g1[j]+g2[j]);
}
}
delete g1;
delete g2;
};
ParallelGradientDericheYRows& operator=(const ParallelGradientDericheYRows &) {
return *this;
};
};
class ParallelGradientDericheXCols: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheXCols(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
int rows=img.rows,cols=img.cols;
long i,j,nb;
long tailleSequence=(rows>cols)?rows:cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
// double a1=1,a2=0;
// double a3=0,a4=0;
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
p=alphaMoyenne;
k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
a5=k,a6=k*exp(-p)*(p-1);
a7=k*exp(-p)*(p+1),a8=-k*exp(-2*p);
b3=2*exp(-p);
b4=-exp(-2*p);
for (nb=0;nb<img.channels();nb++)
{
for (j=range.start;j<range.end;j++)
{
f1 = (float*)img.ptr(0)+(nb);
f1+=j;
i=0;
g1[i] = (a5 + a6 +b3+b4)* *f1 ;
g1[i] = (a5 + a6 )* *f1 ;
i++;
f1+=cols;
g1[i] = a5 * *f1 + a6 * f1[-cols] + (b3+b4) * g1[i-1];
g1[i] = a5 * *f1 + a6 * f1[-cols] + (b3) * g1[i-1];
i++;
f1+=cols;
for (i=2;i<rows;i++,f1+=cols)
g1[i] = a5 * *f1 + a6 * f1[-cols] +b3*g1[i-1]+
b4 *g1[i-2];
// Filtre anticausal vertical
f1 = (float*)img.ptr(0)+(nb);
f1 += (rows-1)*cols+j;
i = rows-1;
g2[i] =(a7+a8+b3+b4)* *f1;
g2[i] =(a7+a8)* *f1;
i--;
f1-=cols;
g2[i] = (a7+a8)* f1[cols] +(b3+b4)*g2[i+1];
g2[i] = (a7+a8)* f1[cols] +(b3)*g2[i+1];
i--;
f1-=cols;
for (i=rows-3;i>=0;i--,f1-=cols)
g2[i] = a7*f1[cols] +a8* f1[2*cols]+
b3*g2[i+1]+b4*g2[i+2];
for (i=0;i<rows;i++,f2+=cols)
{
f2 = ((float*)dst.ptr(i))+(j*img.channels());
*f2 = (float)(g1[i]+g2[i]);
}
}
}
delete g1;
delete g2;
};
ParallelGradientDericheXCols& operator=(const ParallelGradientDericheXCols &) {
return *this;
};
};
class ParallelGradientDericheXRows: public ParallelLoopBody
{
private:
Mat &img;
Mat &dst;
double alphaMoyenne;
double alphaDerive;
bool verbose;
public:
ParallelGradientDericheXRows(Mat& imgSrc, Mat &d,double alm,double ald):
img(imgSrc),
dst(d),
alphaMoyenne(alm),
alphaDerive(ald),
verbose(false)
{}
void Verbose(bool b){verbose=b;}
virtual void operator()(const Range& range) const
{
if (verbose)
std::cout << getThreadNum()<<"# :Start from row " << range.start << " to " << range.end-1<<" ("<<range.end-range.start<<" loops)" << std::endl;
float *f1,*f2;
int rows=img.rows,cols=img.cols;
long i,j,nb;
long tailleSequence=(rows>cols)?rows:cols;
double *g1=new double[tailleSequence],*g2=new double[tailleSequence];
double p=alphaDerive;
double k=pow(1-exp(-p),2.0)/(1+2*p*exp(-p)-exp(-2*p));
double kp=pow(1-exp(-p),2.0)/exp(-p);
double a1=1,a2=0;
double a3=0,a4=0;
double a5=0,a6=kp*exp(-p),a7=-kp*exp(-p),a8=0;
double b1=0,b3=2*exp(-p);
double b2=0,b4=-2*exp(-p);
kp=pow(1-exp(-p),2.0)/exp(-p);
a1=0,a2=kp*exp(-p),a3=-kp*exp(-p),a4=0;
b1=2*exp(-p);
b2=-exp(-2*p);
switch(img.depth()){
case CV_8U :
case CV_8S :
{
unsigned char *c1;
for (nb=0;nb<img.channels();nb++)
{
for (i=range.start;i<range.end;i++)
{
f1 = (float*)dst.ptr(i)+nb;
c1 = (unsigned char*)img.ptr(i)+(nb);
j=0;
g1[j] = (a1 +a2+b1+b2)* *c1 ;
g1[j] = (a1 +a2)* *c1 ;
j++;
c1++;
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1+b2) * g1[j-1];
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1) * g1[j-1];
j++;
c1++;
for (j=2;j<cols;j++,c1++)
g1[j] = a1 * c1[0] + a2 * c1[-1]+b1*g1[j-1]+b2*g1[j-2];
c1 = (unsigned char*)img.ptr(0)+(nb);
c1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a3+a4+b1+b2)* *c1;
g2[j] = (a3+a4)* *c1;
j--;
g2[j] = (a3+a4) * c1[1] +(b1+b2) * g2[j+1];
g2[j] = (a3+a4) * c1[1] +(b1) * g2[j+1];
j--;
c1--;
for (j=cols-3;j>=0;j--,c1--)
g2[j] = a3*c1[1]+a4*c1[2]+b1*g2[j+1]+b2*g2[j+2];
for (j=0;j<cols;j++,f1++)
*f1 = (float)(g1[j]+g2[j]);
}
}
}
break;
case CV_16S :
case CV_16U :
{
unsigned short *c1;
for (nb=0;nb<img.channels();nb++)
{
f1 = ((float*)dst.ptr(0))+nb;
for (i=range.start;i<range.end;i++)
{
c1 = ((unsigned short*)img.ptr(0))+(nb);
c1 += i*cols;
j=0;
g1[j] = (a1 +a2+b1+b2)* *c1 ;
g1[j] = (a1 +a2)* *c1 ;
j++;
c1++;
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1+b2) * g1[j-1];
g1[j] = a1 * c1[0]+a2*c1[j-1]+(b1) * g1[j-1];
j++;
c1++;
for (j=2;j<cols;j++,c1++)
g1[j] = a1 * c1[0] + a2 * c1[-1]+b1*g1[j-1]+b2*g1[j-2];
c1 = ((unsigned short*)img.ptr(0))+(nb);
c1 += i*cols+cols-1;
j=cols-1;
g2[j] = (a3+a4+b1+b2)* *c1;
g2[j] = (a3+a4)* *c1;
j--;
c1--;
g2[j] = (a3+a4) * c1[1] +(b1+b2) * g2[j+1];
g2[j] = (a3+a4) * c1[1] +(b1) * g2[j+1];
j--;
c1--;
for (j=cols-3;j>=0;j--,c1--)
g2[j] = a3*c1[1]+a4*c1[2]+b1*g2[j+1]+b2*g2[j+2];
for (j=0;j<cols;j++,f1++)
*f1 = (float)(g1[j]+g2[j]);
}
}
}
break;
default :
return ;
}
delete g1;
delete g2;
};
ParallelGradientDericheXRows& operator=(const ParallelGradientDericheXRows &) {
return *this;
};
};
UMat GradientDericheY(UMat op, double alphaDerive,double alphaMoyenne)
{
Mat im1(op.size(),CV_32FC(op.channels()));
UMat imRes(op.rows,op.cols,CV_32FC(op.channels()));
cv::Mat mThis = op.getMat(cv::ACCESS_RW);
cv::Mat m2 = imRes.getMat(cv::ACCESS_RW);
ParallelGradientDericheYCols x(mThis,im1,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.cols), x,getNumThreads());
ParallelGradientDericheYRows xr(im1,m2,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.rows), xr,getNumThreads());
return imRes;
}
UMat GradientDericheX(UMat op, double alphaDerive,double alphaMoyenne)
{
Mat im1(op.size(),CV_32FC(op.channels()));
UMat imRes(op.rows,op.cols,CV_32FC(op.channels()));
cv::Mat mThis = op.getMat(cv::ACCESS_RW);
cv::Mat m2 = imRes.getMat(cv::ACCESS_RW);
ParallelGradientDericheXRows x(mThis,im1,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.rows), x,getNumThreads());
ParallelGradientDericheXCols xr(im1,m2,alphaMoyenne,alphaDerive);
parallel_for_(Range(0,mThis.cols), xr,getNumThreads());
return imRes;
}
void CannyBis( InputArray _src, OutputArray _dst,
double low_thresh, double high_thresh,
int aperture_size, bool L2gradient ,InputOutputArray _dx,InputOutputArray _dy)
{
const int type = _src.type(), depth = CV_MAT_DEPTH(type), cn = CV_MAT_CN(type);
const Size size = _src.size();
CV_Assert( depth == CV_8U );
_dst.create(size, CV_8U);
if (!L2gradient && (aperture_size & CV_CANNY_L2_GRADIENT) == CV_CANNY_L2_GRADIENT)
{
// backward compatibility
aperture_size &= ~CV_CANNY_L2_GRADIENT;
L2gradient = true;
}
if ((aperture_size & 1) == 0 || (aperture_size != -1 && (aperture_size < 3 || aperture_size > 7)))
CV_Error(CV_StsBadFlag, "Aperture size should be odd");
if (low_thresh > high_thresh)
std::swap(low_thresh, high_thresh);
/* CV_OCL_RUN(_dst.isUMat() && (cn == 1 || cn == 3),
ocl_Canny(_src, _dst, (float)low_thresh, (float)high_thresh, aperture_size, L2gradient, cn, size))*/
Mat src = _src.getMat(), dst = _dst.getMat();
#ifdef HAVE_TEGRA_OPTIMIZATION
if (tegra::useTegra() && tegra::canny(src, dst, low_thresh, high_thresh, aperture_size, L2gradient))
return;
#endif
// CV_IPP_RUN(USE_IPP_CANNY && (aperture_size == 3 && !L2gradient && 1 == cn), ippCanny(src, dst, (float)low_thresh, (float)high_thresh))
#ifdef HAVE_TBB
if (L2gradient)
{
low_thresh = std::min(32767.0, low_thresh);
high_thresh = std::min(32767.0, high_thresh);
if (low_thresh > 0) low_thresh *= low_thresh;
if (high_thresh > 0) high_thresh *= high_thresh;
}
int low = cvFloor(low_thresh);
int high = cvFloor(high_thresh);
ptrdiff_t mapstep = src.cols + 2;
AutoBuffer<uchar> buffer((src.cols+2)*(src.rows+2));
uchar* map = (uchar*)buffer;
memset(map, 1, mapstep);
memset(map + mapstep*(src.rows + 1), 1, mapstep);
int threadsNumber = tbb::task_scheduler_init::default_num_threads();
int grainSize = src.rows / threadsNumber;
// Make a fallback for pictures with too few rows.
uchar ksize2 = aperture_size / 2;
int minGrainSize = 1 + ksize2;
int maxGrainSize = src.rows - 2 - 2*ksize2;
if ( !( minGrainSize <= grainSize && grainSize <= maxGrainSize ) )
{
threadsNumber = 1;
grainSize = src.rows;
}
tbb::task_group g;
for (int i = 0; i < threadsNumber; ++i)
{
if (i < threadsNumber - 1)
g.run(tbbCanny(Range(i * grainSize, (i + 1) * grainSize), src, map, low, high, aperture_size, L2gradient));
else
g.run(tbbCanny(Range(i * grainSize, src.rows), src, map, low, high, aperture_size, L2gradient));
}
g.wait();
#define CANNY_PUSH_SERIAL(d) *(d) = uchar(2), borderPeaks.push(d)
// now track the edges (hysteresis thresholding)
uchar* m;
while (borderPeaks.try_pop(m))
{
if (!m[-1]) CANNY_PUSH_SERIAL(m - 1);
if (!m[1]) CANNY_PUSH_SERIAL(m + 1);
if (!m[-mapstep-1]) CANNY_PUSH_SERIAL(m - mapstep - 1);
if (!m[-mapstep]) CANNY_PUSH_SERIAL(m - mapstep);
if (!m[-mapstep+1]) CANNY_PUSH_SERIAL(m - mapstep + 1);
if (!m[mapstep-1]) CANNY_PUSH_SERIAL(m + mapstep - 1);
if (!m[mapstep]) CANNY_PUSH_SERIAL(m + mapstep);
if (!m[mapstep+1]) CANNY_PUSH_SERIAL(m + mapstep + 1);
}
#else
Mat dx,dy;
if (_dx.empty())
{
Sobel(src, dx, CV_16S, 1, 0, aperture_size, 1, 0, BORDER_REPLICATE);
Sobel(src, dy, CV_16S, 0, 1, aperture_size, 1, 0, BORDER_REPLICATE);
}
else
{
dx = _dx.getMat(), dy = _dy.getMat();
}
if (L2gradient)
{
low_thresh = std::min(32767.0, low_thresh);
high_thresh = std::min(32767.0, high_thresh);
if (low_thresh > 0) low_thresh *= low_thresh;
if (high_thresh > 0) high_thresh *= high_thresh;
}
int low = cvFloor(low_thresh);
int high = cvFloor(high_thresh);
ptrdiff_t mapstep = src.cols + 2;
AutoBuffer<uchar> buffer((src.cols+2)*(src.rows+2) + cn * mapstep * 3 * sizeof(int));
int* mag_buf[3];
mag_buf[0] = (int*)(uchar*)buffer;
mag_buf[1] = mag_buf[0] + mapstep*cn;
mag_buf[2] = mag_buf[1] + mapstep*cn;
memset(mag_buf[0], 0, /* cn* */mapstep*sizeof(int));
uchar* map = (uchar*)(mag_buf[2] + mapstep*cn);
memset(map, 1, mapstep);
memset(map + mapstep*(src.rows + 1), 1, mapstep);
int maxsize = std::max(1 << 10, src.cols * src.rows / 10);
std::vector<uchar*> stack(maxsize);
uchar **stack_top = &stack[0];
uchar **stack_bottom = &stack[0];
/* sector numbers
(Top-Left Origin)
1 2 3
* * *
* * *
0*******0
* * *
* * *
3 2 1
*/
#define CANNY_PUSH(d) *(d) = uchar(2), *stack_top++ = (d)
#define CANNY_POP(d) (d) = *--stack_top
#if CV_SSE2
bool haveSSE2 = checkHardwareSupport(CV_CPU_SSE2);
#endif
// calculate magnitude and angle of gradient, perform non-maxima suppression.
// fill the map with one of the following values:
// 0 - the pixel might belong to an edge
// 1 - the pixel can not belong to an edge
// 2 - the pixel does belong to an edge
for (int i = 0; i <= src.rows; i++)
{
int* _norm = mag_buf[(i > 0) + 1] + 1;
if (i < src.rows)
{
short* _dx = dx.ptr<short>(i);
short* _dy = dy.ptr<short>(i);
if (!L2gradient)
{
int j = 0, width = src.cols * cn;
#if CV_SSE2
if (haveSSE2)
{
__m128i v_zero = _mm_setzero_si128();
for ( ; j <= width - 8; j += 8)
{
__m128i v_dx = _mm_loadu_si128((const __m128i *)(_dx + j));
__m128i v_dy = _mm_loadu_si128((const __m128i *)(_dy + j));
v_dx = _mm_max_epi16(v_dx, _mm_sub_epi16(v_zero, v_dx));
v_dy = _mm_max_epi16(v_dy, _mm_sub_epi16(v_zero, v_dy));
__m128i v_norm = _mm_add_epi32(_mm_unpacklo_epi16(v_dx, v_zero), _mm_unpacklo_epi16(v_dy, v_zero));
_mm_storeu_si128((__m128i *)(_norm + j), v_norm);
v_norm = _mm_add_epi32(_mm_unpackhi_epi16(v_dx, v_zero), _mm_unpackhi_epi16(v_dy, v_zero));
_mm_storeu_si128((__m128i *)(_norm + j + 4), v_norm);
}
}
#elif CV_NEON
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
vst1q_s32(_norm + j, vaddq_s32(vabsq_s32(vmovl_s16(vget_low_s16(v_dx))),
vabsq_s32(vmovl_s16(vget_low_s16(v_dy)))));
vst1q_s32(_norm + j + 4, vaddq_s32(vabsq_s32(vmovl_s16(vget_high_s16(v_dx))),
vabsq_s32(vmovl_s16(vget_high_s16(v_dy)))));
}
#endif
for ( ; j < width; ++j)
_norm[j] = std::abs(int(_dx[j])) + std::abs(int(_dy[j]));
}
else
{
int j = 0, width = src.cols * cn;
#if CV_SSE2
if (haveSSE2)
{
for ( ; j <= width - 8; j += 8)
{
__m128i v_dx = _mm_loadu_si128((const __m128i *)(_dx + j));
__m128i v_dy = _mm_loadu_si128((const __m128i *)(_dy + j));
__m128i v_dx_ml = _mm_mullo_epi16(v_dx, v_dx), v_dx_mh = _mm_mulhi_epi16(v_dx, v_dx);
__m128i v_dy_ml = _mm_mullo_epi16(v_dy, v_dy), v_dy_mh = _mm_mulhi_epi16(v_dy, v_dy);
__m128i v_norm = _mm_add_epi32(_mm_unpacklo_epi16(v_dx_ml, v_dx_mh), _mm_unpacklo_epi16(v_dy_ml, v_dy_mh));
_mm_storeu_si128((__m128i *)(_norm + j), v_norm);
v_norm = _mm_add_epi32(_mm_unpackhi_epi16(v_dx_ml, v_dx_mh), _mm_unpackhi_epi16(v_dy_ml, v_dy_mh));
_mm_storeu_si128((__m128i *)(_norm + j + 4), v_norm);
}
}
#elif CV_NEON
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
int16x4_t v_dxp = vget_low_s16(v_dx), v_dyp = vget_low_s16(v_dy);
int32x4_t v_dst = vmlal_s16(vmull_s16(v_dxp, v_dxp), v_dyp, v_dyp);
vst1q_s32(_norm + j, v_dst);
v_dxp = vget_high_s16(v_dx), v_dyp = vget_high_s16(v_dy);
v_dst = vmlal_s16(vmull_s16(v_dxp, v_dxp), v_dyp, v_dyp);
vst1q_s32(_norm + j + 4, v_dst);
}
#endif
for ( ; j < width; ++j)
_norm[j] = int(_dx[j])*_dx[j] + int(_dy[j])*_dy[j];
}
if (cn > 1)
{
for(int j = 0, jn = 0; j < src.cols; ++j, jn += cn)
{
int maxIdx = jn;
for(int k = 1; k < cn; ++k)
if(_norm[jn + k] > _norm[maxIdx]) maxIdx = jn + k;
_norm[j] = _norm[maxIdx];
_dx[j] = _dx[maxIdx];
_dy[j] = _dy[maxIdx];
}
}
_norm[-1] = _norm[src.cols] = 0;
}
else
memset(_norm-1, 0, /* cn* */mapstep*sizeof(int));
// at the very beginning we do not have a complete ring
// buffer of 3 magnitude rows for non-maxima suppression
if (i == 0)
continue;
uchar* _map = map + mapstep*i + 1;
_map[-1] = _map[src.cols] = 1;
int* _mag = mag_buf[1] + 1; // take the central row
ptrdiff_t magstep1 = mag_buf[2] - mag_buf[1];
ptrdiff_t magstep2 = mag_buf[0] - mag_buf[1];
const short* _x = dx.ptr<short>(i-1);
const short* _y = dy.ptr<short>(i-1);
if ((stack_top - stack_bottom) + src.cols > maxsize)
{
int sz = (int)(stack_top - stack_bottom);
maxsize = std::max(maxsize * 3/2, sz + src.cols);
stack.resize(maxsize);
stack_bottom = &stack[0];
stack_top = stack_bottom + sz;
}
int prev_flag = 0;
for (int j = 0; j < src.cols; j++)
{
#define CANNY_SHIFT 15
const int TG22 = (int)(0.4142135623730950488016887242097*(1<<CANNY_SHIFT) + 0.5);
int m = _mag[j];
if (m > low)
{
int xs = _x[j];
int ys = _y[j];
int x = std::abs(xs);
int y = std::abs(ys) << CANNY_SHIFT;
int tg22x = x * TG22;
if (y < tg22x)
{
if (m > _mag[j-1] && m >= _mag[j+1]) goto __ocv_canny_push;
}
else
{
int tg67x = tg22x + (x << (CANNY_SHIFT+1));
if (y > tg67x)
{
if (m > _mag[j+magstep2] && m >= _mag[j+magstep1]) goto __ocv_canny_push;
}
else
{
int s = (xs ^ ys) < 0 ? -1 : 1;
if (m > _mag[j+magstep2-s] && m > _mag[j+magstep1+s]) goto __ocv_canny_push;
}
}
}
prev_flag = 0;
_map[j] = uchar(1);
continue;
__ocv_canny_push:
if (!prev_flag && m > high && _map[j-mapstep] != 2)
{
CANNY_PUSH(_map + j);
prev_flag = 1;
}
else
_map[j] = 0;
}
// scroll the ring buffer
_mag = mag_buf[0];
mag_buf[0] = mag_buf[1];
mag_buf[1] = mag_buf[2];
mag_buf[2] = _mag;
}
// now track the edges (hysteresis thresholding)
while (stack_top > stack_bottom)
{
uchar* m;
if ((stack_top - stack_bottom) + 8 > maxsize)
{
int sz = (int)(stack_top - stack_bottom);
maxsize = maxsize * 3/2;
stack.resize(maxsize);
stack_bottom = &stack[0];
stack_top = stack_bottom + sz;
}
CANNY_POP(m);
if (!m[-1]) CANNY_PUSH(m - 1);
if (!m[1]) CANNY_PUSH(m + 1);
if (!m[-mapstep-1]) CANNY_PUSH(m - mapstep - 1);
if (!m[-mapstep]) CANNY_PUSH(m - mapstep);
if (!m[-mapstep+1]) CANNY_PUSH(m - mapstep + 1);
if (!m[mapstep-1]) CANNY_PUSH(m + mapstep - 1);
if (!m[mapstep]) CANNY_PUSH(m + mapstep);
if (!m[mapstep+1]) CANNY_PUSH(m + mapstep + 1);
}
#endif
// the final pass, form the final image
const uchar* pmap = map + mapstep + 1;
uchar* pdst = dst.ptr();
for (int i = 0; i < src.rows; i++, pmap += mapstep, pdst += dst.step)
{
for (int j = 0; j < src.cols; j++)
pdst[j] = (uchar)-(pmap[j] >> 1);
}
}
void cvCannyBis( const CvArr* image, CvArr* edges, double threshold1,
double threshold2, int aperture_size )
{
cv::Mat src = cv::cvarrToMat(image), dst = cv::cvarrToMat(edges);
CV_Assert( src.size == dst.size && src.depth() == CV_8U && dst.type() == CV_8U );
Canny(src, dst, threshold1, threshold2, aperture_size & 255,
(aperture_size & CV_CANNY_L2_GRADIENT) != 0);
}
int lowThreshold=20;
int maxThreshold=20;
int const max_lowThreshold = 500;
int alDerive=100;
int alMean=100;
Mat sobel_x, sobel_y;
UMat img;
const char* window_name = "Edge Map";
/**
* @function CannyThreshold
* @brief Trackbar callback - Canny thresholds input with a ratio 1:3
*/
static void CannyThreshold(int, void*)
{
Mat dst;
double d=alDerive/100.0,m=alMean/100.0;
UMat rx= GradientDericheX(img,d,m);
UMat ry= GradientDericheY(img,d,m);
double minv,maxv;
minMaxLoc(rx,&minv,&maxv);
std::cout << minv << "\t" << maxv << "\\t";
minMaxLoc(ry,&minv,&maxv);
std::cout << minv << " " << maxv << "\n";
Mat mm;
mm=abs(rx.getMat(ACCESS_READ));
rx.getMat(ACCESS_READ).convertTo(sobel_x,CV_16S,1);
mm=abs(ry.getMat(ACCESS_READ));ry.getMat(ACCESS_READ).convertTo(sobel_y,CV_16S,1);
minMaxLoc(sobel_x,&minv,&maxv);
std::cout << minv << "\t" << maxv << "\\t";
minMaxLoc(sobel_y,&minv,&maxv);
std::cout << minv << " " << maxv << "\n";
CannyBis(img, dst,lowThreshold, maxThreshold, 3,false,sobel_x,sobel_y);
imshow("edges x", sobel_x.mul(50));
imshow("edges y", sobel_y.mul(50));
imshow( window_name, dst );
}
int main()
{
ocl::setUseOpenCL(false);
//imread("f:/lib/opencv/samples/data/lena.jpg",IMREAD_GRAYSCALE).copyTo(img);
imread("c:/Users/Laurent.PC-LAURENT-VISI/Downloads/F1A_glider_hook.jpg",IMREAD_GRAYSCALE).copyTo(img);
imshow("Original",img);
namedWindow( window_name, WINDOW_AUTOSIZE );
/// Create a Trackbar for user to enter threshold
createTrackbar( "Min Threshold:",window_name, &lowThreshold, max_lowThreshold, CannyThreshold );
createTrackbar( "Max Threshold:", window_name, &maxThreshold, max_lowThreshold, CannyThreshold );
createTrackbar( "Derive:",window_name, &alDerive, 400, CannyThreshold );
createTrackbar( "Mean:", window_name, &alMean, 400, CannyThreshold );
waitKey();
return 0;
}