There is rapidly increasing interest in the application of satellite/airborne-based remote sensing in the globe, and a changeover from manual techniques of data collection to digital automated techniques is under way. Therefore, integration between image s captured from old and new earth observation systems is crucial issue because this integration will provide the information society by a very useful tool for many applications. This proposed research project is focused on the exploitation of linear features, instead of point features which may not be exist or difficult to be achieved, to develop a generic technique for image registration and matching. This technique will contribute to multi-sensor image matching, LIDAR image matching, and registration of high-resolution satellite images. To meet the need for a generic model, this proposed project is an extension to the research project in which a new model named "Line-Based Transformation Model (LBTM)" was developed.
The research will investigate the performance of the LBTM for image-to-image and image-to-map registration, and further steps will be taken to improve the LBTM for image matching. The approach of this research will involve three main phases: First, verification of the use of the LBTM for registration of different types of images; second, modification of the LBTM for multi-sensor and LIDAR image matching; third, evaluation and optimisation for application cases. Generally, the work proposed in this research will provide a step forward to maximize the benefits of using data from different sources and it will help to develop the process of map production and map updating. Several applications will get direct benefit from the project, some of which are: applications of fast-response to disaster map ping, monitoring of environment, urban growth and change detection.
Call for proposal
See other projects for this call