Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 8

Bubble counting in the reactor using image

processing

Submitted By:
Rishikesh Meena(2013CH1010

Supervisor: Dr. Manojkumar Ramteke

Chemical Engineering Department


Indian Institute of Technology Delhi, New Delhi

Why count bubbles


Bubble count tell about rate of reaction

Lateral movement of bubble in bubble column react

This is done 3 steps


Using Fourier Counts
Using Support Vector Regression
Using contour validation using SVM

Fourier counts
It is applied to full image to
find count using gradients.
Image is divided into patches
Counts in each patches is
estimated using gradient in
images

Fourier transform of image

Support Vector regression


It is trained with image datasets with label counts
In the each patch of image counts is predicted by SVR
Total counts in the image will be summation of counts in
each patches

Contours detection and validation


Contours are found in the image
Circular contours are extracted and validated by SVM
Valid contours are send for counts

References
http://cns-alumni.bu.edu/~slehar/fourier/fourier.html
http://kernelsvm.tripod.com/
http://docs.opencv.org/2.4/doc/tutorials/imgproc/shaped
escriptors/find_contours/find_contours.html

Thank you

You might also like