This page is a part of CVprimer.com, a wiki devoted to computer vision. It focuses on low level computer vision, digital image analysis, and applications. It is designed as an online textbook but the exposition is informal. It geared towards software developers, especially beginners, and CS students. The wiki contains mathematics, algorithms, code examples, source code, compiled software, and some discussion. If you have any questions or suggestions, please contact me directly.

Inspecting semiconductor wafers

From Computer Vision Primer

Jump to: navigation, search
Enlarge
Enlarge
Enlarge
Enlarge
Enlarge

An example of how Pixcavator may be useful in industrial applications. I found these pictures of semiconductor wafers in a white paper Geometric Search Techniques Break Through Traditional Grayscale Correlation Barriers by Steve Geraghty, Director of Operations, Coreco Imaging posted in Machine Vision Online [1]. These are the challenges faced by this industry. It took me about 15 minutes to run Pixcavator [2] a few times and assemble this Excel spreadsheet [3]. The results are quite encouraging.

I ran the program for each of the images with several area settings. Then I compared the results, the number of dark and light objects, with those of the original. For the scaled image, I had to scale the area settings too to have the outputs matched. The biggest challenge was contrast reversal, figure 2. Here you have to add the number of dark and the number of light objects to have reasonable results. It is clear that the quality of the matching can be improved dramatically.

Note that the complexity of the images makes the identification task easier not harder.

The example is related Image search.


Other case studies

Personal tools