Some of the routinely encountered algorithms.
12 Apr 2022 - fubar - Sreekar Guddeti
Some of the routinely encountered algorithms.
Algorithms are the hotspots of any program. They are the abstraction of logic that is coded into a program. Any program almost always an algorithm under the hood.
In the thesis, we have worked with hardware interfacing and image processing. For each set of measurement or analysis, there is an algorithm that is doing the heavy-lifting. Following are these algorithms.
In KerrPy we have used the Canny Edge detection algorithm to detect domain wall from a Kerr micrograph.
The algorithm was given by John F. Canny in 1986.
It is a multistage algorithm involving
PID loops are used in various engineering systems with closed feedback. In our Kerr microscopy system, we use PID loop to set in-plane field. The in-plane field component are configured in a closed feedback mechanism, where the field is set by voltage controlled current source and the resultant field is sensed by a Hall sensor.
X-ray reflectivity profiles for ultra-thin multilayer films exhibit variety of features like
These features can be modelled within the Paratt formalism using the following parameters for each of the layers
For a multilayer with 4 layers, this can lead to atleast 16 parameters, in addition to other parameters like the intensity scaling factor, initial $q_z$, substrate roughness, substrate refractive index.
Unless, we know the nominal values of the parameters, it is difficult to converge to the global minimum using conventional gradient based algorithms like simplex or gradient.
Even if we know the nominal values, it is highly likely for the trajectory of minimization to settle in a deep local minimum.
To avoid such inaccuracies, we have used genetic-algorithm based minimization.