c++ - OpenCV image conversion goes wrong -

i have algorithm stuff. among them, there conversion works fine if i'm working on cv_8uc3 image goes wrong if file type c_16uc3. code:

             //new image created              mat3w img(100,100,vec3w(1000,0,0));              //image conversion - error!              cv::mat inputsource;              //saving image here work              img.convertto(inputsource, cv_64fc3);              //saving image here not work -> black image 

the problem cv_16uc3 image's processing result image of right dimensions black. problem in conversion because saving image right before give legit 1 while saving right after give white one.

edit:

i made changes: cut off useless code , added inputsource declaration. now, while trying stuff, arrived @ conclusion either haven't understood cv types, or strange happening. thought number in type indicating number of bits per channel. so, in head, cv_16uc3 3 channel 16bits per channel. idea strengthened fact image save during tests (before img.convertto) had matching bits per channel number. strange thing, saved inputsource (type cv_64fc3) 8bpc image.

what's missing?

you confused way imwrite , imread work in opencv. opencv documentation

imwrite

the function imwrite saves image specified file. image format chosen based on filename extension (see imread() list of extensions). 8-bit (or 16-bit unsigned (cv_16u) in case of png, jpeg 2000, , tiff) single-channel or 3-channel (with ‘bgr’ channel order) images can saved using function. if format, depth or channel order different, use mat::convertto() , , cvtcolor() convert before saving. or, use universal filestorage i/o functions save image xml or yaml format.

imread

the function imread loads image specified file , returns it. possible flags are:

• imread_unchanged : if set, return loaded image (with alpha channel, otherwise gets cropped).
• imread_grayscale : if set, convert image single channel grayscale image.
• imread_color : if set, convert image 3 channel bgr color image.
• imread_anydepth : if set, return 16-bit/32-bit image when input has corresponding depth, otherwise convert 8-bit.
• imread_anycolor : if set, image read in possible color format.

so case, cv_16u saved without conversion, while cv_64f converted , saved cv_8u. if want store double data, should use filestorage. should take care use imread image appropriate flag.

this example should clarify:

#include <opencv2\opencv.hpp> using namespace cv;  int main() {     // create 16-bit 3 channel image     mat3w img16uc3(100, 200, vec3w(1000, 0, 0));     img16uc3(rect(0, 0, 20, 50)) = vec3w(0, 2000, 0);      // convert 64-bit (double) 3 channel image     mat3d img64fc3;     img16uc3.convertto(img64fc3, cv_64fc3);      // save disk     imwrite("16uc3.png", img16uc3); // no conversion     imwrite("64fc3.png", img64fc3); // converted cv_8uc3      filestorage fout("64fc3.yml", filestorage::write);     fout << "img" << img64fc3; // no conversion     fout.release();       mat img_maybe16uc3_a = imread("16uc3.png" /*, imread_color*/);  // cv_8uc3     mat img_maybe16uc3_b = imread("16uc3.png", imread_anydepth);    // cv_16uc1     mat img_maybe16uc3_c = imread("16uc3.png", imread_unchanged);   // cv_16uc3       mat img_maybe64fc3_a = imread("64fc3.png" /*, imread_color*/);  // cv_8uc3     mat img_maybe64fc3_b = imread("64fc3.png", imread_anydepth);    // cv_8uc1     mat img_maybe64fc3_c = imread("64fc3.png", imread_unchanged);   // cv_8uc3      mat img_mustbe64fc3;     filestorage fin("64fc3.yml", filestorage::read);     fin["img"] >> img_mustbe64fc3; // cv_64fc3     fin.release();      return 0; }