Cal Poly
Object Detection and Classification for Man-made Structures To Aid in Search and Rescue (SAR)
Abstract
dc:description.abstractThis project explores the use of various object detection and classification models to identify permanent man-made structures, specifically cell towers, water towers, watch towers as well as its corresponding shadows (attributes), from aerial imagery. The goal is to evaluate those models and test the feasibility of current models to detect man-made structures in aerial images, in an effort to support search and rescue operations as well as update network infrastructure databases. Multiple object detection models, as well as classification models, including YOLOv8 and VGG19 are implemented and trained on annotated aerial images to determine the presence or absence of man-made structures in the imagery. This work contributes to automating geospatial analysis for search-and-rescue operations and infrastructure identification.
Degree
thesis:*- Name thesis:degree_name
- MS in Computer Science
- Discipline thesis:degree_discipline
- Computer Science
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Nakagawa, Mio A
- Contributors dc:contributor
-
- Franz Kurfess
- Computer Science
- College of Engineering
Subjects
dc:subject × 4Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.calpoly.edu/theses/3333