# How to calculate dimension of an interference pattern?

3 views (last 30 days)
Biniam on 29 Mar 2023
Answered: Jack on 29 Mar 2023
Hello, I'm novice to the field of optics and image processing. Recently, working on IR images for defect/void detection on wafer bonding. Due to the curse of resolution, it's difficult to get the image of small particles using IRT imaging (I got this idea from the papers I reviewed). So, is it possible to compute the dimension (area) of Newton's rings?

Jack on 29 Mar 2023
Yes, it is possible to compute the dimension (area) of Newton's rings in an IR image.
Newton's rings are a pattern of interference fringes that result from the reflection of light between two surfaces with a slight curvature, such as a convex lens placed on a flat surface. The radius of curvature of the lens and the distance between the lens and the surface determine the spacing of the fringes, which can be used to measure the thickness or flatness of the surface.
To compute the area of the Newton's rings in an IR image, you can use image processing techniques such as edge detection, segmentation, and morphological operations. Here are the general steps you can follow:
1. Convert the IR image to grayscale and adjust the contrast to enhance the visibility of the Newton's rings.
2. Apply an edge detection algorithm to extract the edges of the rings. This can be done using techniques such as Canny edge detection or Sobel edge detection.
3. Apply a morphological operation such as dilation or closing to connect broken edges and fill gaps in the edge map.
4. Apply a segmentation algorithm such as region growing or watershed segmentation to segment the Newton's rings from the background and other objects in the image.
5. Calculate the area of the segmented Newton's rings using a method such as the pixel count or the region properties of the segmented object.
Note that the exact details of the image processing steps will depend on the characteristics of your IR image and the specific properties of the Newton's rings you are trying to measure. You may need to experiment with different parameters and algorithms to achieve the best results.