True orthophoto generation

Kurzfassung

T hroughout this Bachelor thesis methods for generating true orthophoto imagery from aerial and satellite imagery and based on digital elevation models are investigated, compared and an application for generating true orthophotos developed. New developments in image and digital processing still increase the interest in orthophotos. The opportunity for imagery with greater detail and resolution results in the demand for greater quality and accuracy of the orthophoto. But the orthophoto reaches the end of the line due to its limitations. Rough terrain or significant difference in elevation leads to inconsistency in accuracy and scale. Therfore new methods and algorithms to overcome these limitations are required. Accurate true orthophotos fullfill all the requirements, but most of the existing approaches to generate true orthophotos imply a 3D model of the desired earth surface area, which is very time and cost-intensive in the generation process. To generate true orthophotos fast and cheap, the German Aerospace Center takes another route and wants to extend their existing image processing software by true orthophoto generation based on fully automated generated elevation models. The four general steps of the true orthophoto generation process are rectification of the source images and locating as well as marking occluded areas, due to differences in elevation and inconsistency in accuracy and scale, followed by seamline placement combined with mosaicking and feathering. To rectify the images is found to be the most crucial part of the process. The overall goals of this thesis are to devise a method to create true orthophoto, investigate problems and solutions for generating orthophotos, implement methods that are capable to create true orthophoto imagery, optimized regarding quality and computing time, fully automated and evaluate the solutions through developed test methods. The result shows that the implemented methods give reasonable results in a noticeable time.