detect scratches in image

this is the input image i want to detect scratches in this image but i don't know how to proceed please guide me. i want output image like this

[Walter Roberson]

You appear to have marked blobs, not scratches. Scratches are long thin marks such as appear near the bottom left of the bright area.

[mehroz irshad]

so how can i detect those blobs.

Answer 1

[Image Analyst]

Try a tophat or bottomhat filter, imtophat() or imbothat().

[mehroz irshad]

original = imread('fivTest.png');
figure, imshow(original);
se = strel('disk',12);
tophatFiltered = imtophat(original,se);
figure, imshow(tophatFiltered);
contrastAdjusted = imadjust(tophatFiltered);
figure, imshow(contrastAdjusted);

i have tried this code tophat filter but displays following error.

Error using images.internal.imageDisplayValidateParams>validateCData (line 119)
If input is logical (binary), it must be two-dimensional.

Error in images.internal.imageDisplayValidateParams (line 27)
common_args.CData = validateCData(common_args.CData,image_type);

Error in images.internal.imageDisplayParseInputs (line 78)
common_args = images.internal.imageDisplayValidateParams(common_args);

Error in imshow (line 222)
      images.internal.imageDisplayParseInputs({'Parent','Border','Reduce'},preparsed_varargin{:});

Error using imadjust>parseInputs (line 207)
Specified syntax IMADJUST(I) is only supported for 2-D grayscale images. Additional input arguments are required for use
with RGB images.

Error in imadjust (line 97)
    parseInputs(varargin{:});

[Walter Roberson]

original = rgb2gray( imread('fivTest.png') );

[Image Analyst]

You should also get a better photo. Your setup is horrible. Have you considered using more uniform illumination, board illumination (no line or point sources of light), crossed polarizers (to knock out that specular reflections right where you want to analyze), etc. It's always better to start with a good image than to try to fix us a crappy image afterwards in software.