A vortex is a region in which the flow revolves around an axis line, which may be straight or curved. High mass planets leave remarkable features in their parent protoplanetary disks (PPDs), namely a gap, spiral waves, vortices, and eccentricities. In the context of planet formation, vortices are good candidates to trap dust grains allowing them to grow to planetesimal or planets sizes.
Our aim is to figure out these vortices ( if present ) formed due to presence of rotation and helicity in a three-dimensional hydrodynamic helically rotating turbulence in a database, provided as an input. And then find out the clockwise or anticlockwise rotation of those vortices. Our approach is as follows,
- We are given a dataset in .h5 format. We are using a software tool named Visit Visualization that detects the vectors in the dataset and generate that vectored image in both 3D and 2D .
- From an image we are detecting the voritices using Hough Circle Detection Algorithm ( as in 2D vortex is displayed as circle ) and crop the shape as square to detect the rotation of vector in that shape.
- Now we are detecting the rotation of vectors using a Convolutional Neural Network Model. The training dataset for this model is generated by taking multiple screenshot of vector rotation from different position of original images. The test dataset is the cropped images of vortex.