TY - THES

T1 - A Framework and Workflow for Hyper Resolution Image Visualization in Mineshaft Inspection

AU - Brandner, Michael

N1 - no embargo

PY - 2017

Y1 - 2017

N2 - This thesis presents a framework and workflow for the automatic preparation of data and images to enable the use of hyper resolution images for mineshaft inspections. The focus of the thesis is on the process from image registration to visualization of image data in the dimensions of 10^6 x 10^5 pixels. Concepts regarding image registration are presented, including phase correlation, non-rigid registration and homography. The thesis addresses the suitability of different web mapping libraries for building a visualization tool for deep mine shafts. Attention is focused on handling large volumes of data, resulting from the demand for a high image quality with a resolution of 1 pixel per millimetre for 2.5 x 10^4 m^2. A mapping tool is developed, using the Leaflet JavaScript library. The generation of a tile-layer as input for the mapping application is achieved, using GDAL for generic tiling of hyper resolution images. JSON is used as data exchange format and the JSON files provide meta-data associated with tag representations in the hyper resolution image. In order to guarantee consistency in connecting meta-data to pixel coordinates an example for the use of a common coordinate reference system is proposed. Finally, the functionality of the mapping application is presented and the framework is tested.

AB - This thesis presents a framework and workflow for the automatic preparation of data and images to enable the use of hyper resolution images for mineshaft inspections. The focus of the thesis is on the process from image registration to visualization of image data in the dimensions of 10^6 x 10^5 pixels. Concepts regarding image registration are presented, including phase correlation, non-rigid registration and homography. The thesis addresses the suitability of different web mapping libraries for building a visualization tool for deep mine shafts. Attention is focused on handling large volumes of data, resulting from the demand for a high image quality with a resolution of 1 pixel per millimetre for 2.5 x 10^4 m^2. A mapping tool is developed, using the Leaflet JavaScript library. The generation of a tile-layer as input for the mapping application is achieved, using GDAL for generic tiling of hyper resolution images. JSON is used as data exchange format and the JSON files provide meta-data associated with tag representations in the hyper resolution image. In order to guarantee consistency in connecting meta-data to pixel coordinates an example for the use of a common coordinate reference system is proposed. Finally, the functionality of the mapping application is presented and the framework is tested.

KW - automatic shaft monitoring

KW - hyper resolution images

KW - web mapping applications

KW - Leaflet

KW - automatische Schachtüberwachung

KW - hochauflösende Bilder

KW - Web-Mapping Anwendungen

KW - Leaflet

M3 - Master's Thesis

ER -