2D Convolution / Creating new filter

OpenCV function filter2D is used to create new linear filters.

void filter2D(InputArray src, OutputArray dst, int ddepth, InputArray kernel, Point anchor=Point(-1,-1), double delta=0, int borderType=BORDER_DEFAULT )
  • src – input image.
  • dst – output image of the same size and the same number of channels as src.
  • ddepth – desired depth of the destination image; if it is negative, it will be the same as src.depth(); the following combinations of src.depth() and ddepth are supported:
    • src.depth() = CV_8U, ddepth = -1/CV_16S/CV_32F/CV_64F
    • src.depth() = CV_16U/CV_16S, ddepth = -1/CV_32F/CV_64F
    • src.depth() = CV_32F, ddepth = -1/CV_32F/CV_64F
    • src.depth() = CV_64F, ddepth = -1/CV_64F
    when ddepth=-1, the output image will have the same depth as the source.
  • kernel – convolution kernel (or rather a correlation kernel), a single-channel floating point matrix; if you want to apply different kernels to different channels, process them individually.
  • anchor – anchor of the kernel that indicates the relative position of a filtered point within the kernel; the anchor should lie within the kernel; default value (-1,-1) means that the anchor is at the kernel center.
  • delta – optional value added to the filtered pixels before storing them in dst.
  • borderType – pixel extrapolation method (see borderInterpolate() for details).
A kernel is a fixed size array of numerical coefficients along with an anchor point in that array.

The code provided below is slight modification of code provided in OpenCV documentation.


  1. Load image
  2. Create a kernel to convolve with the input matrix ( here all elements of kernel is equal; so performs a low pass filter operation)
  3. Apply convolution (filter2D)
  4. Draw contours



#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/highgui/highgui.hpp"
#include <stdlib.h>
#include <stdio.h>

using namespace cv;

void conv2(Mat src, int kernel_size)
    Mat dst,kernel;
    kernel = Mat::ones( kernel_size, kernel_size, CV_32F )/ (float)(kernel_size*kernel_size);

    /// Apply filter
    filter2D(src, dst, -1 , kernel, Point( -1, -1 ), 0, BORDER_DEFAULT );
    namedWindow( "filter2D Demo", CV_WINDOW_AUTOSIZE );imshow( "filter2D Demo", dst );

int main ( int argc, char** argv )
    Mat src;

    /// Load an image
    src = imread( "1.jpg" );
    if( !src.data )  { return -1; }


    return 0;


  1. this is not even correct... filter2D does not do convolutions, but rather correlation.

    1. It does, but most of the filters are symmetric. Hence the convolutions just look like correlation.