accés aux pixels d''une image en couleur et les mettre dans un tableau

[Natalie11]

je programme en C et j'utilise OpenCv
j'essaye de stocker la valeur des pixels dans une matrice

uchar mat[307200][3];
data      = (uchar *)image_ref->imageData;
       for(i=0;i<height;i++) 
	   {
		   for(j=0;j<width;j++) 
		   {
			   for(k=0;k<channels;k++)
			   {
		           mat[i][j]=data[i*step+j*channels+k] ;
				   
			   }
			   printf("%d\t",mat[i][j] );
		   }
	   }

mais le pb c'est que j'obtient pas la meme valeur que lorsque je fais

CvScalar s=cvGet2D(image_ref,i,j);

svp ou est le probleme??
mercii

[Salem]

The written language on this board is English.
Can you edit your post to English please.

[CommonTater]

The written language on this board is English.
Can you edit your post to English please.

I don't see the answer to her problem... but she says she's trying to import an image matrix into an array and having problems converting it to a 2d array.

And yes she needs to post in English from here on... I'm no translator.

[GReaper]

Hey, french!! God, although i knew it very well in the past, know i've completely forgotten it!

[CommonTater]

Hey, french!! God, although i knew it very well in the past, know i've completely forgotten it!

I haven't read, spoken or heard French since high school (Actually, most Canadians do not speak french, it's only required for government jobs)...

I'm amazed I took any meaning from it whatsoever...

[Natalie11]

i'm sorry for my french post and i rectifie that i said in english, sorry

I programming in C and I use OpenCV
I try to store the value of the pixels in a matrix

uchar mat[480][640];
uchar *data  = (uchar *)image->imageData;
       for(i=0;i<height;i++) 
	   {
		   for(j=0;j<width;j++) 
		   {
			   for(k=0;k<channels;k++)
			   {
		           mat[i][j]=data[i*step+j*channels+k] ;
				   
			   }
			   
		   }
                  printf("n\");
	   }

I want to know if is it just or no and how can I transforme my color picture to a gray level ?
thinks

[Salem]

Presumably, your channels are say 0,1,2 to represent the RGB components?

Convert RGB image or colormap to grayscale - MATLAB

Algorithm

rgb2gray converts RGB values to grayscale values by forming a weighted sum of the R, G, and B components:

0.2989 * R + 0.5870 * G + 0.1140 * B

- Get your 3 values for RGB
- Normalise the values in the range [0..1), by dividing by 256
- Apply the above function to get a brightness in the range [0..1)
- Scale back to whatever resolution of bits you have for your grayscale.

[Natalie11]

yes Salem my channels are represented by 0,1 et 2
that's what i do

       // convert to gray

       for(i=0;i<height;i++) 
	   {
		   for(j=0;j<width;j++) 
		   {
			   
			           mat[i][j]= (data[i*step+j*channels]*0.114 
						          +data[i*step+j*channels+1]*0.587
								  +data[i*step+j*channels+2]*0.2989)/256 ;
     
		   }
		 
	   }

        // apply the matrix for my picture

	   for(i=0;i<height;i++) 
	   {
		   for(j=0;j<width;j++) 
		   {
				data[i*step+j]=mat[i][j] ;
		   }
	   }

please tell me that is just else give some clues , i need the help
thanks

[msh]

Maybe?

mat[i][j]= ((data[i * step + j * channels] / 256) * 0.114 +
            (data[i * step + j * channels + 1] / 256) * 0.587 +
            (data[i * step + j * channels + 2] /256) * 0.2989);

[Salem]

Use a debugger, or add some print statements, to observe what is going on.
Are the RGB values you're pulling out of your data array correct?
What about your value of step - are you stepping correctly through the array?

Or create a really small "image" consisting of say 10 RGB values, and feed that into your conversion function to see what values emerge.

[Natalie11]

thanks for all

I managed to convert my image to gray level without dividing by 256 because when I divide I obtain an image completely white !!!!!

[anduril462]

For future reference, the white image is probably because you were doing integer division: Question 3.15.

[Natalie11]

it's me again )

I managed to convert my image to gray level and I applied the Sobel filter for edge detection and I haven't got good result .I send you the original_picture , sobel_picture and the code of my fonction and try to help me please
thanks

	int ConvY[3][3]={1,2,1,0,0,0,-1,-2,-1};
	int ConvX[3][3]={1,0,-1,2,0,-2,1,0,-1};
	double mX,mY,x,y;
	int i,j,k,l;
	double moy,g;


	for(i=(pImage->height)-2;i>=1;i--)
		for(j=(pImage->width)-2;j>=1;j--)
		{
			mX=0;
			mY=0;
			for( k=-1;k<=1;k++)
			{
				for(l=-1;l<=1;l++)
					
				{
					mX+=(double)pImage->imageData[(i+k)*pImage->widthStep+(j+l)*pImage->nChannels]*ConvX[k+1][l+1];
					mY+=(double)pImage->imageData[(i+k)*pImage->widthStep+(j+l)*pImage->nChannels]*ConvY[k+1][l+1];	
				}
				moy=sqrt((double)mX*mX+mY*mY);
				if(moy>255) {moy=255;}

				pImage->imageData[i*pImage->widthStep+j*pImage->nChannels]=moy;
				pImage->imageData[i*pImage->widthStep+j*pImage->nChannels+2]=pImage->imageData[i*pImage->widthStep+j*pImage->nChannels];
				pImage->imageData[i*pImage->widthStep+j*pImage->nChannels+1]=pImage->imageData[i*pImage->widthStep+j*pImage->nChannels];
		
			}
		}
}

this is the original_picture Image : original_picture.jpg

and the sobel_picture Image : sobel_picture.jpg

thanks

[GReaper]

If you're trying to convert RGB values to a single GrayScale value, you've got this all wrong!
This's how i do it, and it works pretty fine:

Example:

GrayScale = (R + G + B)/3

[Salem]

I don't suppose you bothered to do any research on this subject?
It might be a "gray" answer, but it won't be the equivalent grayscale image.