Processing and Recognition of the Earth Surface Snapshots
The Methods, Algorithms and Software of Halftone Image Processing and 2D, 3D Objects Recognition on the Earth Surface Snapshots
Tech Area / Field
- INF-IMA/High-Definition Imaging and Displays/Information and Communications
- INF-SOF/Software/Information and Communications
3 Approved without Funding
Belarussian State University of Informatics and Radioelectronics, Belarus, Minsk
- University of Cambridge / Engineering Department, UK, Cambridge\nUniversity of Paisley, UK, Paisley\nThe University of Hull / Department of Computer Science, UK, Hull\nUniversity of Milan / Department of Information Technologies, Italy, Milan\nYonsei University / Department of Computer Science, Korea, Seoul\nVal de Marne University / Senart Institute of Technology / Intelligence in Instrumentations and Systems Laboratory, France, Liesant\nUniversita di Reggio Calabria, Italy, Reggio Calabria
Project summaryThe purpose of the project is the development of methods, algorithms and software for high-performance processing the gray-scale earth surface images and 2D, 3D objects recognition.
The current state in the field of the project development and importance of the proposed project.
The modern geoinformation technologies are widely applied in cartography, agriculture and forestry, in the field of water resources management and land tenure, in evaluation of the environment and other areas of human activity. The tasks of maps processing, decoding of space snapshots, the analysis of earth snapshots are solved with the use of images preprocessing, the extraction of informative features, the creation of classifiers for the analysis and recognition of graphic images, metric characteristic measurement of the chosen objects, their transformation and comparison.
The interesting results of earth snapshots processing and recognition have been obtained and published in the following papers: Gruen and R. Nevatia, guest eds. Special Issue on Automatic Building Extraction from Aerial Images // Computer Vision and Image Understanding, v. 72 (2), Nov. 1998, pp. 99-214, A. Vailaya, A. Jain and H.I. Zhang. On Image Classification: City images vs. landscapes // Pattern Recognition, 1998, pp. 1921-1935.
The world leaders in the development and distribution of geoinformation systems (GIS) such as ERDAS Inc., INTERGRAPH, Geograph/Geodraw, ARC/INFO, ARCView, Panorama and others are mainly oriented on the class of cartographical tasks. Despite of wide technical possibilities the above-mentioned systems do not fully satisfy the customer requirements. Such important functions as Earth snapshot processing, decoding and recognition have not been represented completely enough.
Therefore the necessity of the development of new tools having got increased intellectual level of 2D objects decoding and identification, visualization and calculation of the metric characteristics of 3D objects compatible both with domestic and foreign systems is obvious. The project is intended to bring abovementioned possibilities into modern GIS.
The project influence on the progress in the development field.
The product to be developed will be based on original algorithms, which allow to increase the automation level of objects selection and identification on the earth surface snapshots. Most of used methods and algorithms will have the world scientific and applied novelty.
The project participants.
The collective of the participants of the project consists of scientists which are directly re-lated to development of weapon, systems of automatic navigation, systems of searching for targets, and other areas concerned with the weapon development, treatment and utilization.
Other experts, which have a great experience in a solution of the scientific and practical tasks in the field of image processing and 2D and 3D object recognition, could also be used in the weapon development and treatment process. For the last years scientific group members have published more than 50 scientific papers concerning the given area.
After working out the nine project tasks the following results will be obtained:
– The methods and algorithms for preprocessing the gray scale images of the Earth snapshots.
– The methods and algorithms for description of geometric shape of 2D objects on the Earth snapshots.
– The methods and algorithms for 2D object identification on the Earth snapshots.
– The methods and algorithms for 3D object reconstruction using its projections.
– The methods and algorithms for 3D object modelling and metric characteristics evaluation.
– The open architecture of the software for image processing and recognition of objects on the Earth surface snapshots.
– Software modules for data processing subsystems.
– Software application package for processing the snapshots of the earth surface images and 2D, 3D objects recognition.
– The laboratory installation for testing the algorithms of Earth image processing.
Software application package will be an independent product, compatible with the open geoinformation software technologies, and can be used in already existing technological lines of potential customers. Practical significance of the project is that modern technologies of image processing and pattern recognition can be implemented in GIS. Protection of the intelligent property is also supposed.
Project results application.
The developed methods and algorithms will be implemented as portable programs with open architecture and will be compatible with the most of existing GIS. The developed software products can also be used as a basis for the construction of the open independent systems of earth surface monitoring for various purposes such as agriculture and forestry, water resources management and land tenure, evaluation of the environment and other areas of human activity. In rare cases some finishing adaptation of the software under interests and restrictions of a concrete customer may be required.
The correspondence of project researches to the ISTC purposes.
The work on the project corresponds to the purposes and tasks ISTC, as the project realization will allow to:
– Redirect the efforts and skills of specialists having earlier been involved in missile technology and nuclear weapon development to solving tasks concerned to ecological land tenure, forestry and other problems, what is especially actual for the solution of the complex tasks of liquidation of emergency situation consequences.
– Create methods, algorithms and software image processing tools oriented to potential customer tasks, that may be a basis for commercial product.
– Integrate the scientists directly concerned with weapon development and treatment in to worldwide scientific community of image processing and pattern recognition of the earth surface snapshots.
The total amount of the project.
The duration of the project is three calendar years. The project group consists of 19 scientists and experts and auxiliary 7 workers from Belorussian State University of Informatics and Radioelectronics.
It is necessary to fulfil the following tasks:
1. Development of methods and algorithms for preprocessing the gray scale images.
2. Development of methods and algorithms for description of geometric shape of 2D objects.
3. Development of methods and algorithms for identification of 2D objects.
4. Development of methods and algorithms for 3D object reconstruction using its projections.
5. Development of methods and algorithms for modeling 3D representations of topographic objects.
6. Development of architecture of the software for Earth image processing and recognition.
7. Design of the data processing software.
8. Debugging and testing all the software as whole package.
9. Manufacturing the laboratory installation for experimental testing the algorithms for image processing and objects recognition.
The weapon scientists will develop the ideology and technique of the data processing and object recognition, and the architecture of the system. Other experts will formalize all the tasks to computer algorithms level and will develop the software. Auxiliary staff will create laboratory installation and solve the tasks of the technical programming.
At first the tasks 1-5, and task 9 concerned with the methods and algorithms for data processing and object recognition development will be worked out. Task 9 should be completely worked out before tasks 4 and 5 could be started. At second the tasks 6-8 concerned with algorithm and software implementation will be worked out.
The role of foreign collaborators.
The following possibilities are submitted to collaborator:
– the share publications of project outcomes;
– to get access to outcomes obtained during the project fulfilment, ability to get as the partial programs, libraries and routines, as the complete software product, and ability to have access to raw information and intermediate results of project;
– to take a part at a stage of project tasks detalization and software adaptation to the collaborator needs and interests;
– to evaluate the intermediate and final outcomes, obtained during project fulfilment;
– consultation on problems of protection of rights on the intellect property created during the project working out;
– to share the property rights on the "know-how" created during the project working out with developer, and joint actions on commercialization of the project outcomes.
Technical approach and methodology.
The authors of the project have developed the common methodology of the project fulfilment and carried out necessary basic investigations concerned with image processing, 2D object recognition, and 3D object reconstruction using the set of 2D projections. The authors have reached essential results in the field of gray scale image processing, using the neural networks and cellular logic for the object recognition. Large experience in that area allow to think the utilizing proposed approach and methods results in the succesful processing GIS objects too. At now project participants have evaluation versions of some tested algorithms.
The processing methods are based on discriminant functions calculation, cellular logic, mathematical morphology, orthogonal transformations in various bases, fast direct and inverse wavelet transforms. Using the developed methods will allow to reduce a computational expenses and increase the reliability of object recognition.
The methods of features extraction, description of the geometric shapes for objects, construction of classifiers will be utilized with using geometric transforms, cellular logic, orthogonal and moment functions, structural and statistical analysis, neural networks.
The methods and algorithms for surfaces and 3D object construction will be based on searching for congruent geometric objects and textures on the raster and vector representations of projections.
The developed programs and routines will be used to make up software package intended for processing the gray-scale aero-photographs, electronic photo plans and information from the CCD-camera of the laboratory installation.
The International Science and Technology Center (ISTC) is an intergovernmental organization connecting scientists from Kazakhstan, Armenia, Tajikistan, Kyrgyzstan, and Georgia with their peers and research organizations in the EU, Japan, Republic of Korea, Norway and the United States.
ISTC facilitates international science projects and assists the global scientific and business community to source and engage with CIS and Georgian institutes that develop or possess an excellence of scientific know-how.