Investigation of Computational Problem Area Detection on Earthen Spillways Across Varying Image Resolutions
Abstract
This thesis work investigates the ability of segmentation algorithms to segment vegetal problem areas in aerial images of earthen spillways. Four different segmentation algorithms were applied to aerial images of spillways that were resized to 8 different resolutions. Segmentation results from each algorithm and spillway resolution were analyzed on the basis of the percentage of pixels segmented for each problem area using a manually segmented image as a ground truth for comparison.
Collections
- OSU Theses [15752]