Analyzing Urbanization of Northern Menomonie through Supervised and Unsupervised Classification

Abstract

Urbanization refers to concentrations of human populations. Urban expansion is monitored through the alteration of landscapes into residential, commercial, and industrial uses. Continually tracking the movement of urbanization allows for the assessment of climate vulnerability within an urban system. Image classification is an image scanning technique that analyzes the individual pixels of the image to classify objects within a picture. Pixels themselves do not reflect patterns but the grouping of pixels relays spatial trends. There are two main types of Image Classification: Supervised classification: Requires user input to key in multiple examples of the object of interest within the imagery. Unsupervised classification: Automatic clustering of similar pixels by a pre-trained model. Purpose: The purpose of this study is to analyze the urbanization of Northern Menomonie over the previous two decades. Study this change through supervised and unsupervised classification models. Analyze the change in land usage over time. We’re interested in determining if unsupervised classification—the automated and more time-efficient method—can produce a classified map as accurate and detailed as the user-intensive supervised method.