Investigation of the Integrity of Tailings Dam Surfaces Using Aerial Imaging and Computer Vision
Loading...
Authors
Gomez Llerena, Jose Alberto
Issue Date
2023
Type
Thesis
Language
Keywords
Erosion , Machine learning , Monitoring , Seepage , Tailings
Alternative Title
Abstract
This thesis explores the critical analysis of seepage and rill erosion. This study tackles the difficulties in tracking and evaluating the condition of tailings dam surfaces by utilizing cutting-edge methods in computer vision and aerial imagery. The study uses multispectral imaging, a robust technique for detecting subtle changes in surface conditions that offer important insights into possible weak points in dam structures, to measure seepage. The thesis investigates the use of U-Net for semantic segmentation and Convolutional Neural Networks for picture classification in the setting of rill erosion. By enhancing the precision and effectiveness of rill erosion detection, this dual-methodology approach aims to enable the identification and delineation of erosion structures on dam surfaces. A thorough examination is made easier by the combination of CNN and U-Net, which helps to provide a more complex knowledge of the location and magnitude of rill erosion. The field of environmental monitoring and dam safety will benefit greatly from the research findings. In the end, the use of multispectral imaging and sophisticated computer vision algorithms helps identify and mitigate possible problems related to seepage and rill erosion by promoting a more thorough and proactive approach to the evaluation of tailings dam surfaces.