Date of Award
6-27-2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Computer Science
First Advisor
Mariofanna Milanova
Abstract
This dissertation describes how to automate the reading of dials, gauges, and instruments using machine learning (ML) methods. This process can be described as analog to digital conversion through an air gap without any direct electronic connection. The goal is to convert the values of existing instruments into digital values for control and monitoring without requiring any intrusion, changes, or upgrades to the instruments being observed. Images of instrument faces can be distorted by various kinds of noise, but this can be overcome using a deep learning convolutional neural network (CNN) approach similar to handwriting recognition. One advantage of this approach is that it allows robots to adapt to the real world through specialized sensors without having to re-engineer existing infrastructure. In the process, an artificial interpolation method is developed that uses the entire decision vector to deduce in-between numerical values instead of a single categorial softmax value. This enables the inversion of affine transformations like rotation, scaling, translation, skewing, and perspective.
Recommended Citation
Warren, Lloyd Van, "Flying by ML or CNN Inversion of Affine Transforms" (2023). Theses and Dissertations. 1149.
https://research.ualr.edu/etd/1149
