City Research Online

Abstract

Photogrammetry is a non-contact measurement technique of obtaining 3D information in the object space by processing 2D information on the image planes. This thesis mainly concerns the data reduction methods used in close range multiple camera photogrammetry.

The objective was to develop a fast method which could be used in real-time measurement systems. Traditional methods such as the bundle adjustment are investigated and the advantages and disadvantages are discussed. It was found that the bottle-neck which restricts the speed was the computation of matrices, especially the inverse of the coefficient matrix, in the simultaneous least squares estimation process.

In this thesis, an alternative method named the separate adjustment is developed and successfully used in close range photogrammetry. This method can give the same results as the traditional bundle adjustment, but with a significant saving of computation time and memory requirements. It was found that the computation time required by the separate adjustment is directly proportional to the numbers of the object points and cameras, and the memory required is independent of how many object points or cameras are involved. A thorough comparison between the two methods is given and some results from simulation and practical tests are presented.

The separate adjustment is a séparé least squares estimation process which estimates the unknown parameters separately in groups rather than simultaneously. In this way the sizes of the matrices in the least squares process are reduced. Therefore time and memory are reduced accordingly. Due to the special structures of the design matrix and the coefficient matrix in close range photogrammetry the separate adjustment method is very efficient and very easy to apply, especially when it is used in real-time applications.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) T Technology > TR Photography
Departments:	School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses School of Science & Technology > Engineering

Export

Downloads