1

Using CImg: I read here that you can change an individual pixel's RGB data like this:

CImg<float> img("filename.jpg");
// Change the (100,200) pixel to Red.
img(100,200,0,0) = 255; //R
img(100,200,0,1) = 0; //G
img(100,200,0,2) = 0; //B

but img(100, 200, 0, 0) returns the type unsigned char *, which is obviously not a variable as the above snippet implies. When I run the above code, I get "error C2106: '=' : left operand must be l-value" in my build output.

A potential solution would be to use a different version of the CImg constructor, which takes raw pixel data as its first parameter, but I can't find any information on how to format the data before running it through the constructor - the first parameter is describes as a template in line 9671 of CImg.h.

Any help would be greatly appreciated; I've been at this for a while.

Pere5
  • 35
  • 2
  • 7

2 Answers2

1

The expression img(100,200,0,0) is not a unsigned char* but a unsigned char& which makes a huge difference. Of course you can write img(100,200,0,0) = value without any problems (except if your img variable is const), I use it all the time. Note also that the third argument can be omitted in your case, as it is always 0, so you can write img(x,y,0)=255; img(x,y,1)=1; img(x,y,2)=1;.

Look at the examples provided in the CImg package, you will see that many code uses the CImg<T>::operator()like this. Hopefully !

bvalabas
  • 41
  • 1
0

You can use the function draw_point.

From the reference manual:

CImg& draw_point ( const int x0, const int y0, const int z0, const tc ∗const color, const float opacity = 1 )

Cpp Forever
  • 900
  • 8
  • 16