Ask Your Question
0

Ideas to process challenging image

asked 2014-05-19 09:07:47 -0500

Hello everyone.

I'm working with Infra Red image that is an output of a 3D sensor. This sensors project a Infra Red pattern in order to draw a depth map, and, because of this, the IR image has a lot of white spots that reduce its quality. So, I want to process this image to make it smoother in order to make it possible to detect objects laying in the surface.

The original image looks like this: image description

My objective is to have something like this (which I obtained by blocking the IR projecter with my hand) : image description

I understand that this is a challenging problem. Any ideas?

Best regards

edit retag flag offensive close merge delete

Comments

Would be nice if you would post some feedback on your progress

Goosebumps gravatar imageGoosebumps ( 2014-06-02 08:50:05 -0500 )edit

3 answers

Sort by ยป oldest newest most voted
1

answered 2014-06-05 06:14:46 -0500

My main point on the noise removal was to have a cleaner image so it would be easier to detect objects. However, as I tried to find a solution for the problem, I realized that it was unrealistic to remove all noise from the image, since most of the image is actually noise.. So I had to find the objects despite the noise. Here is my aproach

1 - Initial image

1

2 - Background subtraction followed by opening operation to smooth noise

2

3 - Binary threshold

3

4 - Morphological operation close to make sure object has no edge discontinuities (necessary for thin objects)

4

5 - Fill holes + opening morphological operations to remove small noise blobs

5

6 - Detection

6

edit flag offensive delete link more
1

answered 2014-05-20 03:23:06 -0500

Goosebumps gravatar image

updated 2014-05-21 02:22:03 -0500

A morphological operation that could work is opening: image description

        int morphOpSize = 3;
        StructuringElementEx element =
            new StructuringElementEx(
                morphOpSize,
                morphOpSize,
                morphOpSize / 2,
                morphOpSize / 2,
                Emgu.CV.CvEnum.CV_ELEMENT_SHAPE.CV_SHAPE_ELLIPSE);

        filtered = gray.MorphologyEx(
            element,
            Emgu.CV.CvEnum.CV_MORPH_OP.CV_MOP_OPEN,
            2);

(sorry for my Emgu dialect)

Otherwise look for edge preserving low pass filtering.

edit flag offensive delete link more

Comments

The opening operation does remove some noise, but I think that before applying that it could be useful to apply a good noise removal operation that addresses those white spots.

Pedro Batista gravatar imagePedro Batista ( 2014-05-20 06:37:42 -0500 )edit
0

answered 2014-05-20 10:24:12 -0500

zerog80 gravatar image

Another filter with effect similar to morphological opening is the median... did you try it with a small radius? It should remove the majority of the noise while retaining the main shapes.

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2014-05-19 09:07:47 -0500

Seen: 460 times

Last updated: Jun 05 '14