Hi,I ran into the term "analysis of the vertical and horizontal histogram" explained as "showing the intensity differences between successive rows, pixel-wise". I didn't get the point here. I wonder if there is an OpenCV implementation for this operation or should I implement it by myself? Any clue will help. Thanks.
The histogram calculation function is exactly what you need. What they mean is that you have to perform the histogram calculation on rows and colums. As input you see the image, which is of the Mat type. This can be either a 2D image, or a 1D row or column.
how can that show the intensity differences between successive rows by performing histogram calculation on rows or colums?
I couldn't use the histogram functions descriped in openCV doc , because it try to find color histogram , and the question try to find projection in on X-axis and Y-axis for black and white image. i used the following code the find vertical and herozontal projection.
//2-find x ,y projections (horizontal projection)
Mat horizontal(ROIimg.cols,1,CV_32S);//horizontal histogram
horizontal = Scalar::all(0);
Mat vertical(ROIimg.rows,1,CV_32S);//vertical histogram
vertical = Scalar::all(0);
for(int i=0;i<ROIimg.cols;i++)
{
horizontal.at<int>(i,0)=countNonZero(ROIimg(Rect(i,0,1,ROIimg.rows)));
}
for(int i=0;i<ROIimg.rows;i++)
{
vertical.at<int>(i,0) = countNonZero(ROIimg(Rect(0,i,ROIimg.cols,1)));
}
where ROIimg is balck and white
Hi @berak i'm just was expecting that reduce will work as resize and the output will be the same image type one chanel 8 bit per pixel . anyway i tried the function and it crash when called , i use the following
Mat vertMat ; Mat horzMat ; reduce(ROIimg,vertMat,1,CV_REDUCE_SUM); reduce(ROIimg,horzMat,0,CV_REDUCE_SUM);
also i used the following code and also crash
Mat vertMat (ROIimg.rows,1,CV_32S);
Mat horzMat (1,ROIimg.cols,CV_32S);
vertMat = Scalar::all(0);
horzMat = Scalar::all(0);
reduce(ROIimg,vertMat,1,CV_REDUCE_SUM);
reduce(ROIimg,horzMat,0,CV_REDUCE_SUM);
how does this function work
Is your input image one channel? The output is the same number of channels as the input.
thanks @break , it seems this function working correctly i tested the following code and it's working
Mat imgHist = imread("C:\\Temp\\Testreduce.png",IMREAD_GRAYSCALE);
threshold(imgHist,imgHist,128,255,CV_THRESH_OTSU);
Mat vertMat, horzMat;
reduce(imgHist,vertMat,1,CV_REDUCE_SUM,CV_32S);
reduce(imgHist,horzMat,0,CV_REDUCE_SUM,CV_32S);
Also the ROI is working ,i just added the last parameter to CV_32S ,
Asked: 2013-07-30 02:58:32 -0600
Seen: 6,389 times
Last updated: Nov 10 '16