RGB and Adjusting Brightness

[da4thrza]

Hi,

Im very new to objective c and the iphone sdk, so please excuse me if
this is a stupid question. I tried using your sample project and
noticed that you have code in place to convert it to grayscale. I went
into the image constructor and say where you are using

CGColorSpaceRef colorSpace=CGColorSpaceCreateDeviceGray();

I tried changing this to CGColorSpaceRef
colorSpace=CGColorSpaceCreateDeviceRGB(); but that didnt work and
simply changed the image to all black.

Could you please tell me how to keep the image in full color?

Also, i would like to adjust the brightness of the image. Is that done
by using the normalise function? If so, is there a way to tell it how
much to brighten an image?

Thanks in advance for your help!

[Chris Greening]

try this:

typedef struct _rgbPixel {

uint8_t a;

uint8_t b;

uint8_t g;

uint8_t r;

} RGBPixel;

RGBPixel *getRawValuesFromImage(UIImage *srcImage, int width, int height) {

RGBPixel *rgbImage=(RGBPixel *) malloc(width*height*sizeof(RGBPixel));

CGColorSpaceRef colorSpace=CGColorSpaceCreateDeviceRGB();

CGContextRef context=CGBitmapContextCreate(rgbImage,  width, height, 8, width*4, colorSpace, kCGBitmapByteOrder32Little|kCGImageAlphaNoneSkipLast);

CGContextSetInterpolationQuality(context, kCGInterpolationHigh);

CGContextSetShouldAntialias(context, NO);

CGContextDrawImage(context, CGRectMake(0, 0, width, height), [srcImage CGImage]);

CGContextRelease(context);

CGColorSpaceRelease(colorSpace);

return rgbImage;

}

The image normalisation just stretches the range of values so that they are in the rand 0-255. If you want to brighten an image then you need to increase the value of each pixel (but don't forget that the maximum pixel value is 255).

[Marcelo Roque]

Hey Chris, may I ask you something too? Don't know if you got my message, but I was wondering if you could post on that "extractConnectedRegion" and "findLargestStructure" sample code. I can't find a way to make that work properly with an "Cannyed"  image.

Thanks

2009/8/27 Chris Greening <cmgre...@gmail.com>

[Chris Greening]

Hi Marcelo,

Sorry, yes I got your message, I'm just catching up on the emails now.

The problem you might be having is that the canny algorithm is marking an EDGE as 0 and NOEDGE as 255.

The extract connection region and find largest structure are looking for pixels of 255 - so you are currently extracting the NOEDGE pixels instead of the EDGE pixels.

If you invert the canny image (set anything that is 0 to 255 and anything that is 255 to 0) you should get it to work.

Cheers

Chris.

[Marcelo Roque]

Hm, Chris, what would be the "std::vector<ImagePoint> *maxPoints"?

Marcelo

[Chris Greening]

I'm not sure I understand the question.

But I think if you invert the canny image and then feed it into the
findLargestStructure you should end up with a vector of points that
are the edge points.

[da4thrza]

Thanks a million Chris! Ill try this out on my Mac as soon as i get
home

> > Thanks in advance for your help!- Hide quoted text -
>
> - Show quoted text -

[Marcelo Roque]

I'm actually doing this...

" edges.image->findLargestStructure(edges.image.m_imageData);" but I am getting an error that goes:  "Confused by earlier error, bailing out."

Any idea?

2009/8/27 Chris Greening <cmgre...@googlemail.com>

[Chris Greening]

You should be doing something like this:

std::vector<ImagePoint> points;

edges.image->findLargestStructure(&points);

This will give a vector containing all the points in the image that make up the largest structure.

Cheers

Chris.

[Marcelo Roque]

Oh, think I understood. Thanks Chris!

Marcelo

2009/8/27 Chris Greening <cmgre...@gmail.com>

[da4thrza]

Hey Chris,

Sorry if i should already know the answer to this, but i have a
question about your response. The method you posted will return a
RGBPixel, but looking at the code you posted in your sample project,
you first make a call to createImage which in turn calls the
constructor to image (the one that was converting to greyscale).
Should i just create a new constructor for Image and take the code
within the getRawValuesFromImage method and paste it in there? If
thats the case, i will obviosuly remove the return rgbimage; line from
there, but do i also need to make teh call to initYptrs();?

Thanks again!

On 27 Aug, 13:36, Chris Greening <cmgreen...@gmail.com> wrote:

[Chris Greening]

Hi,

The code in the image class will only work with greyscale images so
you won't be able to plug in the the code I've given you and make it
work.

I'm afraid you'll have to write your own function to do full colour
image processing.

To access the pixels use:

RGBPixels *rgbData=getRawValuesFromImage....

Then to access a pixel at X and Y:

rgbData[x+y*width].r
rgbData[x+y*width].g
rgbData[x+y*width].b

Hope that helps. There's not much more I can tell you.

[da4thrza]

Got it... thanks again for all your help!

> >> - Show quoted text -- Hide quoted text -

[da4thrza]

Hey Chris,

Just wanted to thank you again... i got it all working last night.
Thanks for your help and willingness to share!

Reza

On 27 Aug, 15:47, Chris Greening <cmgreen...@gmail.com> wrote:

> >> - Show quoted text -- Hide quoted text -

[Franky Orson]

Hi Reza,

Any chance you would be willing to share your changes or submit them
to the project?

Thanks

--Frank