c++ - Fast/Efficent Pixel Access in Magick++ -

As an educational exercise for me, I am writing an application that can average a bunch of images. It is often used in astrofotography to reduce the noise.

The library I am using is Magic ++ and I have actually succeeded in writing the application. But, unfortunately, its slow speed I am using this code: Lt; columns; column ++) {red.clear (); Blue.clear (); Green.clear (); For (i = 1; i <-10; i ++) {color RGB RGB (image [i]. Pixel color (column, row)); Red.push_back (rgb.red ()); Green.push_back (rgb.green ()); Blue.push_back (rgb.blue ()); } RedVal = Average (red); GreenVal = Average (green); BlueVal = Average (blue); Red Val = Red Will * Max RGB; Green val = green veal * maxrajba; Blue Val = Blueville * Max RGB; Color new RGB (redwall, greenwall, blueville); StackedImage.pixelColor (column, row, newRGB); }}

Average 10 images of the code by going through each pixel and adding pixel intensity of each channel to double vector average Then takes the vector as a parameter and the average of the result. This average is then used on the corresponding pixel in the stacked image - which is the resultant image it works just fine but as I mentioned, I am not happy with the speed on the core i5 machine It takes 2 minutes and 30 seconds. The picture is 8 megapixels and 16 bit TIFF. I think that it has a lot of data, but I have done it faster in other applications.

Is it slow to loop or pixel color (x, y) is the way to reach the pixels in an image? What is a fast way?

Why use vectors / arrays in all?

Why not

  double red = 0.0, blue = 0.0, green = 0.0; For (i = 1; i <-10; i ++) {color RGB RGB (image [i]. Pixel color (column, row)); Red + = rgb.red (); Blue = + rgb.blue (); Green + = rgb.green (); } Red / 10 =; Blue / = 10; Green / = 10;   
 This vector avoids 36 function calls per object per pixel.  
 and you can get even better performance by using the full  PixelCache  image instead of the original  image  objects  
 then " Low-level image pixel access "view section  
  pixelpacket * pix = cache [i] + row * column + column; Red + = pix- & gt; Red; Blue + = px-> Blue; Green + = picture-> Green;   
 Now you have removed 10 calls to pixel collar, 10 color RGB constructor and 30 accuser function in 10 pixels.  
  Note, this is all the principle; I have not tested any of these