Opencv study NOTE 2: Use opencv for Image Comparison

Comparing the two images is similar, this is widely used. At the beginning, I thought this technology was very advanced (of course it was also very advanced) and I had to learn a lot of related knowledge to get started, at least you must have a certain understanding of all aspects before you can do this research.

However, opencv provides a set of APIS for this comparison, so that we can easily compare two images. This is the comparison of histograms. histogram is used in English for histograms, the principle is to convert the image into a histogram, and then compare the histogram. To some extent, the similarity of the image is actually reflected. The code below is as follows:

# Include "opencv2/highgui. HPP "# include" opencv/cv. HPP "// use int histogrambins = 256 to draw a histogram; float histogramrange1 [2] = {0,255}; float * histogramrange [1] = {& histogramrange1 [0]}; /** imagefile1: * imagefile2: * method: cocould be cv_comp_chisqr, cosine, cv_comp_correl, cv_comp_intersect */INT comparehist (const char * imagefile1, const char * imagefile2) {iplimage * image1 = cvloadimage (imagefile1, 0); iplimage * image2 = cvloadimage (imagefile2, 0); cvhistogram * histogram1 = cvcreatehist (1, & values, dimensions, histogramrange ); cvhistogram * histogram2 = cvcreatehist (1, & amp; #, metrics, histogramrange); cvcalchist (& image1, histogram1); cvcalchist (& image2, histogram2); histogram (histogram1, 1 ); cvnormalizehist (histogram2, 1); // cv_comp_chisqr, cv_comp_bhattacharyya can be used for histogram comparison. The smaller the value, the more similar the image is to printf ("cv_comp_chisqr: %. 4f \ n ", cvcomparehist (histogram1, histogram2, cv_comp_chisqr); printf (" cv_comp_bhattacharyya: %. 4f \ n ", cvcomparehist (histogram1, histogram2, histogram); // compare the cv_comp_correl and cv_comp_intersect histograms. A larger value indicates that the image is more similar to printf (" cv_comp_correl: %. 4f \ n ", cvcomparehist (histogram1, histogram2, cv_comp_correl); printf (" cv_comp_intersect: %. 4f \ n ", cvcomparehist (histogram1, histogram2, latency); cvreleaseimage (& image1); cvreleaseimage (& image2); histogram (& histogram1); cvreleasehist (& histogram2 ); return 0;} int main (INT argc, char * argv []) {comparehist (argv [1], argv [2]); // comparehist ("D: \ camera.jpg "," d: \ camera1.jpg "); Return 0 ;}

Now I use two photos taken by camera to compare them. camera.jpgand camera1.jpg

The running result of these two graphs is:

Cv_comp_chisqr: 0.2902

Cv_comp_bhattacharyya: 0.1707

Cv_comp_correl: 0.8017

Cv_comp_intersect: 0.8070

If you use the same image to compare the running results, it should be

Cv_comp_chisqr: 0.0000

Cv_comp_bhattacharyya: 0.0000

Cv_comp_correl: 1.0000

Cv_comp_intersect: 1.0000