How to design for Fourier filters?

AI Thread Summary
To achieve edge enhancement in Fourier optics, high pass filtering involves converting the image to the Fourier domain, removing unwanted low-frequency components, and performing an inverse Fourier transform. The design of filters is complex, requiring careful consideration of filter dimensions and spatial frequency characterization. Specific methods for calculating filter dimensions are not commonly detailed in standard texts. For more specialized guidance, engaging with the comp.dsp USENET newsgroup is recommended. This approach can provide targeted assistance for designing filters tailored to specific optical image processing needs.
jaycee4u
Messages
2
Reaction score
0
I am working on an experiment on Fourier optics-optical image processing. In order to get edge enhancement, how do we actually go about high pass filtering. How to design filters? How to decide the filter dimensions? Can anyone please help me?
 
Physics news on Phys.org
Convert the image to the Fourier domain, lop off the low-frequency components you don't want, then take the inverse Fourier transform to get back an image. The design of filters is an immense topic, much too large to describe here.

- Warren
 
I have done that. But to get edge enhancement, hoe to calculate the dimensions of the high pass filter? How to characterize the spatial frequencies? I went through a few books but none gave a method to design filters for my purpose. Would you help me?
 
jaycee4u said:
I have done that. But to get edge enhancement, hoe to calculate the dimensions of the high pass filter? How to characterize the spatial frequencies? I went through a few books but none gave a method to design filters for my purpose. Would you help me?

i would take this question to the comp.dsp USENET newsgroup. Google Groups is one way to post there.
 
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. Someone who shows interest in science is initially a welcome development. So are fresh ideas from unexpected quarters. In contrast, there is a scientific community that is meticulously organized down to the last detail, allowing little to no external influence. With the invention of social media and other sites on the internet competing for content, unprecedented opportunities have opened up for...
I am going through this course on collision detection: https://siggraphcontact.github.io/ In this link is a PDF called course notes. Scrolling down to section 1.3, called constraints. In this section it is said that we can write bilateral constraints as ##\phi(\mathbf{x}) = 0## and unilateral constraints as ##\phi(\mathbf{x}) \ge 0##. I understand that, but then it says that these constraints call also be written as: $$\mathbf{J} \mathbf{u} = 0, \mathbf{J} \mathbf{u} \ge 0,$$ where...
Back
Top