OpenCV C++: Convert RGBA to HSL and then split channels

dom picture dom · Nov 29, 2012 · Viewed 15.3k times · Source

For some image segmentation work I'd like to use the lightness channel of an image in HSL color space.

To accomplish this I convert a RGBA image to RGB and then so HSL. After the color conversion I split the image into it's color planes using cv::mixChannels, which gives me black output for the saturation / lightness plane.

Code:

cv::Mat src;
cv::Mat hsl;

cv::cvtColor(srcRgba , src , CV_RGBA2RGB);
cv::cvtColor(src, hsl, CV_RGB2HLS);

cv::Mat hue = cv::Mat::Mat(hsl.size(), hsl.depth());
cv::Mat saturation = cv::Mat::Mat(hsl.size(), hsl.depth());
cv::Mat lightness = cv::Mat::Mat(hsl.size(), hsl.depth());

cv::Mat matsOut[] = { hue, saturation, lightness };

// hsv[0] => hue[0], hsv[1] => saturation[0], hsv[2] => lightness[0]
int ch[] = { 0,0, 1,0, 2,0 };

// number of elements in hsl -> 1
// number of elements in matsOut -> 3
// number of pairs in ch -> 3
cv::mixChannels(&hsl, 1, matsOut, 3, ch, 3);

Maybe I messed something up with cv::mixChannels?

EDIT

This is the cv::split code I used and the error Xcode gives me:

Code:

cv::Mat src;
cv::Mat hsl;

cv::cvtColor(srcRgba , src , CV_RGBA2RGB);
cv::cvtColor(src, hsl, CV_RGB2HLS);

std::vector<cv::Mat> hslChannels;
cv::split(hsl, hslChannels);

Error:

Undefined symbols for architecture x86_64:
  "cv::split(cv::Mat const&, std::__1::vector<cv::Mat, std::__1::allocator<cv::Mat> >&)", referenced from:
      hsvTest(cv::Mat) in test.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

EDIT 2

Got it, this works:

cv::Mat src;
cv::Mat hsl;

cv::cvtColor(srcRgba , src , CV_RGBA2RGB);
cv::cvtColor(src, hsl, CV_RGB2HLS);

cv::Mat hslChannels[3];
cv::split(hsl, hslChannels);

Now the hue plane is completely black, but saturation and lightness plane are looking okay …

Answer

Pablosproject picture Pablosproject · Nov 29, 2012

The error is that you're passing to cvSplit a vector of CVMat. Try to look here : http://docs.opencv.org/modules/core/doc/operations_on_arrays.html?highlight=cvsplit