A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
A high speed algorithm for circular object location
Pattern Recognition Letters
Numerical recipes in Pascal: the art of scientific computing
Numerical recipes in Pascal: the art of scientific computing
Algorithm for finding the center of circular fiducials
Computer Vision, Graphics, and Image Processing
Design of Fiducials for Accurate Registration Using Machine Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
Recognizing corners by fitting parametric models
International Journal of Computer Vision
Optimum circular fit to weighted data in multi-dimensional space
Pattern Recognition Letters
Localization of circular objects
Pattern Recognition Letters
Comments on "Design of Fiducials for Accurate Registration Using Machine Vision"
IEEE Transactions on Pattern Analysis and Machine Intelligence
Sub/spl mu/m Registration of Fiducial Marks Using Machine Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
Visible models for interactive pattern recognition
Pattern Recognition Letters
Journal of Mathematical Imaging and Vision
| Hi-index | 0.00 |
The photogrammetric exploitation of aerial images essentiallyrequires the accurate reconstruction of the imaging geometry. Thisespecially includes the determination of the orientation of thecamera. Usually, the orientation parameters are determined byspatial resection, knowing the exact coordinates of control points onthe ground and in the image. The reliability and accuracy of thisregistration task strongly depend on the selection of suitablelandmarks as well as on the precision obtained for landmarklocalization. In this contribution, we consider the problem ofautomatic landmark extraction for the purpose of aerial imageregistration. We suggest to use manhole covers as a specific type ofcircular landmarks which frequently occur in urban environments andwe introduce a model-based approach for localizing these featureswith high subpixel precision.Our approach is based on a parametric intensity model. Localizationof the landmarks is done by directly fitting this modelto the observed image intensities. Since we have an explicitdescription of the landmark it is possible to verify the result byexploiting the estimated parameters. We also address the problem oflandmark detection which can greatly be supported by templatematching. The template used is a prototype model which isgenerated from representative examples during a training phase. Thetraining scheme also provides initial values for the fittingprocedure as well as thresholds for the final verification step. Thefull approach has been tested on synthetic as well as on real imagedata.