-

With the ever increasing the number of remote sensing satellites, advances in data fusion and the functionality of modern geographic information systems, the use of multi-image spatial information products is swiftly becoming commonplace. However, in order to meet the requirements of the user, each individual image making up the multi-image product needs to be expressed in the same geometric reference frame. This means the images have to be accurately registered to geodetic co-ordinate system (e.g. Digital maps).
Automatic Geo-Referencing involves a search to find the transform that yields the highest similarity between the input and the Geo-referenced data. It has been known that there is no single best registration technique for all types of data and applications. Due to the tremendous complications and complexities associated with the natural scenes appearing in satellite imageries and digital vector maps, fully automatic Geo-referencing process have faced serious obstacles and thus, only in a relatively simple imaging environment a reliable result is normally expected.
Proposed registration procedure of this paper matches satellite images with GIS data of the same scene that have been taken at different times, geometric structure and data types. The present approach is designed to be completely independent from the sensor type and any a prior information on the exterior orientation. The overall strategy for our proposed registration method may be expressed by the following interrelated three phases:
1- Multi-resolution Representation
2- Feature Extraction
3- Registration based on Genetic algorithm.
In this method, key features (salient points, intersections, and corners) in the reference and data based have been detected based on a multi-resolution representation of information. By construction of image and feature pyramids, in this stage, for each feature in the left feature pyramid, based on corresponding mathematical model, a search area is constructed on the corresponding right feature pyramid. Now to identify the conjugate features the Genetic algorithm is employed. The Genetic algorithm starts with the selection of population of features followed by the determination of a so called criterion function which can comprise different similarity measures (e.g. geometric discrepancy). Using this criterion function a new population is constructed by decomposition of the old population using a so called Cross-Over operator. The procedures are repeated until a small subset of the population with a specific pattern best satisfies the criterion function. The main advantage of Genetic algorithm is its fast rate of convergency compared to the other searching methods.
The potential of the proposed method is evaluated using IKONOS imagery and corresponding digital map in the urban and suburban area of Tehran City. Registration process is performed hierarchically using five–layer pyramids.
The proposed automatic registration method discussed in this paper, has proved to be very efficient and reliable for automatic registration of different satellite imageries with GIS data. The implemented methodology has the following characteristics:
• Utilization of a multi-resolution representation of information and mathematical models.
• Employing a Genetic algorithm for conjugate feature identification and modelling.
In spite of the success which is gained in the implementation of the presented method, the topic by no means is exhausted and still a great deal of research works are needed. These research works should be focused mainly on the development of a more sophisticated Genetic algorithm, interest operator and matching strategy. All of these are currently under investigation in our institute.