0

I'm doing my final year project on image processing in matlab. My project is identifying which lights are 'off' in a street in a static image.

enter image description here

These are the sequences of steps that i'm following :

enter image description here

I have done thresholding and image labelling i.e, labelling each light source from 1 and so on in thresholded image. The corresponding code is 

I = imread('st.jpg');
imshow(I);
G = rgb2gray(I);
imshow(G);
BW = im2bw(G,0.7);
imshow(BW);
f = fspecial('average',3);
I1 = filter2(f,BW,'same');
imshow(I1);
[Lw, nw] = bwlabel(I1);
imtool(Lw,[]);

s = strel('disk',5);
O = imopen(Lw,s);
figure, imshow(O,[]);

Now I need to create a data base of each light source and compare the test image(image with one/many light source off) with reference image to identify which one is off and display it.
Since houghpeaks for each light source is different. I want to extract houghpeaks for each light source using hough transform.
I have searched for it so much but i dint get anything. If someone helps me it would be a great help for my project.
Thanks already.

agua from mars
  • 16,428
  • 4
  • 61
  • 70
Sandeep Deepu
  • 1
  • To add an image inline, just click on the image button on the tool bar while you're writing your question : http://meta.stackexchange.com/questions/75491/how-to-upload-an-image-to-a-post – agua from mars Mar 25 '16 at 11:23
  • I'm confused, `hough` and `houghpeaks` functions in MATLAB are designed to detect lines ([here is an example](http://stackoverflow.com/a/2600292/97160)). How is that related to identifying if a light is on or off? – Amro Mar 25 '16 at 11:49
  • Yes. I din't quite understood that part either. But it might be based on intensity values of each light source obtained from the houghpeaks. And comparing with the test image where a light is off. Or can i use some distance algorithm there? Or if you know some other way to continue from there, please do suggest? – Sandeep Deepu Mar 25 '16 at 12:40

0 Answers0