image with lot of backdround noise at threshold level.

I convert a image to a binary to extract a feature. The problem is that the there is lot of noise (high connectivity).

I use MATLAB auto threshold finding function for finding threshold hold.

The problem is that the noise is at the same level (image attached)

How can I solve this problem? How a potential noise removal, optimizing threshold should look like?

Note: a 10 % threshold variation already removes all noise

Thanks in advance

Answer 1

[Image Analyst]

If you know the approximate size range of the center bar, you can use bwareafilt(). If you know it's location, you can take the blob with its centroid closest to the x value of the expected location. Attach your original image and segmentation routine if you need more help.

TUTORIAL: How to ask a question (on Answers) and get a fast answer