Georeferencing multi-scale imagery in photogrammetry

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
Helsinki University of Technology | Diplomityö
Checking the digitized thesis and permission for publishing
Instructions for the author
Date
2007
Major/Subject
Fotogrammetria
Mcode
Maa-57
Degree programme
Language
en
Pages
viii + 64
Series
Abstract
Georeferencing is crucial to make imagery useful for extracting geospatial information. This thesis presents an economic and feasible approach to georeference multi-scale imagery to the common coordinate system, which is called local coordinate system, and consistency, completeness, and reliability of referenced spatial information can be simultaneously improved. Since georeferencing methods include direct georeferencing and indirect georeferencing, this presented approach is mainly performed by bundle block adjustment, which is based on indirect georeferencing method. In the method of georeferencing multi-scale imagery, the crucial part is based on recursive georeference. New points can be obtained from block adjustment of different levels and can be used as GCPs in next adjustment process. As iterative procedures are carried out, reliability will be greatly improved. The source data is from Finnish Jabal Haroun Project (FJHP). As there were no control points in Jabal Haroun area, control points are derived from provided satellite orthoimage, DEM and topographic map. These derived control points can be applied in 1:30,000 aerial image block adjustment. Tie points are selected between neighboring photographs or between the flight strips. As a result of block adjustment, the unified coordinates of adjusted control points and tie points are obtained. Then these coordinates can be used as control points or check points in 1:15000 aerial image block adjustment. These two blocks are referenced by using the same points in consistent coordinate system. Well-known collinearity equations are employed in the process of bundle block adjustment. Since redundant observations can improve the reliability of block, georeferencing multi-scale imagery considerably increases the number of redundancy. Least squares adjustment method is applied to achieve unified solution of parameters. Terrestrial images with different camera parameters can be referenced in local coordinate system by using control points from 1:15000 image block. This process is carried out by iWitness software. Spatial resection and intersection are executed to calculate interior orientation (I0) and exterior orientation (E0) of image by given initial values. Consequently, with desired accuracy different scale images are georeferenced in local coordinate system and EO of each image is obtained. A method of coordinate transformation based on EO is introduced in chapter 6. This method is implemented with Matlab. Breaklines in ground coordinate system are transformed to local coordinate system. Two sets of stereomodels are established based on same images with different E0. In these two coordinate systems, the breaklines keep same image coordinates. Collinearity equations are employed during the transformation. In the last chapter, Monoplotting software is implemented by Java programming language. The idea is that one image with known E0 is backprojected onto DEM. Object coordinates can be calculated by intersecting photogrammetric ray with DEM.
Description
Supervisor
Haggrén, Henrik
Thesis advisor
Nuikka, Milka
Keywords
georeference, IO, EO, GCP, bundle block adjustment, least squares adjustment, collinearity equations, resection, spatial intersection, monoplotting
Other note
Citation