Lecturer, University Department of Electronics,
B.R.A.Bihar University Muzaffarpur.
Email: [email protected]
Dr. Ashok Kumar Sinha
Professor 0f Electronics & Instrumentation,
Galgotia College of Engineering Greater Noida(U.P),India
Email: [email protected]
The modern era of science and technology s is the era of expert skill. Everywhere in every domain relevant expertise is needed. The major problems in accessing a human expert in a particular field are unavailability and scarcity of real experts and if the human expert is available then there may be problem for common people in making contact with him. Consultation may be very expensive and human expert may feel the repetitive job uninteresting. This in turn may affect expert’s efficiency. The other major problems that are being faced by the human expert are the limitation of his memory and processing inability of all the essential knowledge required in the process of decision-making. As a result of researches and developments, day by day, new knowledge in enormous amount is being added in every discipline and thus more relevant and accurate advice can be taken from a human expert if his own knowledge is updated which is not an easy task. Human experts are bounded by limitations and it is quite difficult for a human expert to consider all the essential factors while taking decision. Something is always escaped and remains unattended. Thus some tool or assistance is needed even for an expert to update his knowledge and get help in decision making process. In the researches and developments in science and technology, attempts have been always made to overcome the problems of people. The advancements made in the discipline of Artificial Intelligence and Computer Science and Eng. have tackled the problems related to mental and intellectual processes of the people. Gradual advancements in these disciplines have enhanced our cognitive capabilities. From very beginning scientists and researchers of AI have been trying to produce systems that can behave like an intelligent being. In course of such developments scientists, researchers and other related resource persons realized the clue of human expertise in a particular field and tried to encode and assimilate the knowledge and experience of human experts in computer that led to the notion of development of expert systems in different domains. An expert system is a computer-based program that uses knowledge, facts and different reasoning techniques to solve problems that normally require the abilities of human experts. The program asks series of questions about the concerned problem and gives appropriate advice based on its store of knowledge. The knowledge, the expert system uses is made up of either rules or experience information about the behavior of elements of a particular subject domain. Such systems can be designed for specific hardware and software configurations, or they can be software systems that are designed to run on a general-purpose computer. Expert systems of today support many problem solving activities such as decision making, knowledge fusing, designing, and planning, forecasting, regulating, controlling, monitoring, identifying, diagnosing, prescribing, interpreting, explaining, training etc. using different techniques and it is expected that future expert systems will support even more activities. In the beginning, expert systems were developed in the chemical and scientific domains and by the end of 1970s expert systems were operating in the medicine, chemical, education, natural resources and science domains. Expert system started to gain popularity in the early 1980s. The announcement of successful operational systems like PROSPECTOR [1,2] a natural resources system that evaluates geographic sites for potential mineral deposits of commercial interest, MYCIN , medical consulting system, etc. catalysed the expert system technology. The availability of powerful tools to develop expert system have made possible creation of large number of expert systems in different domains. With the increasing interest in expert systems and availability of more powerful tools many more expert systems are being developed for new domains. The knowledge engineers are trying to use better knowledge representation and control strategies to develop more powerful expert systems. Today the most important applications of expert systems are those in which expert systems access databases used for other purpose or linked to other types of software and systems. Many valuable applications in future will combine conventional data processing with expert system technology. The linkage of expert systems with database management systems, management information systems, decision support systems, process control, office automation, geographical information system will tend to take place due to combination of knowledge engineering and information engineering .
This paper presents essentials of design and development technology of expert system and their application in the natural resource database management to facilitate the remote sensing applications.
2. Management of Natural Resources:
The economic growth of a region depends upon the proper exploitation of its natural resources. The resources like land; water, minerals, forests, fisheries and livestock are the natural gift and are transformable into tangible wealth on exploitation to produce agricultural, industrial and energy outputs. These are the most significant ingredients for stimulating the economic growth of the region.(fig.). The economic planning for a region or state or nation needs detail information of various items of natural resources. It is a fact that lot of data related to natural resources, already exists in scattered form at many places in government departments and in their files and are not easily available in a consolidated form when needed by planners. For the proper utilization, equitable distribution and optimal management of natural resources the most needed are an inventory of resources, present day utilization levels, or future utilization possibilities. Therefore, attempts are required to produce data from concerned sources in the standardized formats and put in an appropriate database. The set of activities related to data management on natural resources, such as data generation, data collection, compilation, storage, retrieval and processing are mutually interacting and inter-dependent which naturally open option for management as a system. In this connection the Department of Science and Technology, Govt. of India has launched the pilot project called as the Natural Resource Database Management System (NRDMS) , with following objectives:
- To evolve appropriate methodology for collection, collation, storage and processing of data on natural resources in a given region in totality.
- To evolve standardized format in which natural resources and socio-economic data could be presented in an integrated manner to establish linkage among various hierarchical units.
Figure 1. Components of Natural Resources
- To make assessment of natural resources of the areas under study.
- To make utilization of the information for the purpose of planning and development.
The significant developments in the field of electronics, computer science, artificial intelligence, communication, space-technology, availability of variety of sensors and platforms, remote sensing techniques and data processing techniques enable collection, analysis and interpretation of data with multi-objective approach at highly minimized cost. These data are being processed using modern computational techniques to provide the most up to date and useful information. The advancements in the information technology have made the area of natural resources management information rich. Information is coming in volumes and can be obtained by various means such as maps in conventional and digital form, aerial photographs surveys or experimental data, remotely sensed images and other forms of transmitted signal. Information is dynamic with intermittent or continuous changes in space and time and its enormous volume makes their handling very tough task for personnel associated with natural resources management without aid of modern data management system. In order to control flood of such information the application of knowledge and expertise related to natural resources is essential and will be helpful in streamlining, analyzing, managing, digesting, visualizing and integrating information in an efficient and affective manner. The inter-play  of knowledge and information shown in fig.1 makes the role of information technology more appropriate in intelligent decision-making. It is very straightforward to accept that intelligent decision can’t be taken without use of relevant knowledge. The role of knowledge and information is complementary in the intelligent decision-making.
3. Remote Sensing and knowledgebase:
In the recent years the technique of remote sensing has shown its superiority in data collection for natural resources management. It has been recognized that the value of data which is collected by known conventional means is considerably enhanced by the use of remote sensing and air-photo interpretation techniques which in turn calls for data of ground truths. The technique of remote sensing has been applied in almost every aspect such as atmosphere, geosphere ,biosphere, hydrosphere, and cryosphere together with environmental application and data collection systems etc. of natural resources management. Remotely sensed data/images are used to obtain necessary information on land under various crops, crop rotation and agricultural practices adopted, soil types, problems of land degradation, availability of water bodies (both surface and ground water) etc., which are very useful for agricultural development. The remotely sensed data/images can be taken even of inaccessible land and identification of unused land, waste land, degraded land etc. can be done by applying suitable technology and agricultural practices. The repetition coverage of space remote sensing is useful in detecting changes/degradation, unwanted happening, and correct measures can be taken in advance.
Remote sensing data/images have been used in water resource
management in citing various recharge structures through the preparation of thematic maps on land use/land cover, geomorphology, surface water bodies etc. and their combined analysis. Based on the land cover, slope, soil etc. it is possible to priorities areas in watersheds where there is need for immediate a forestation or other treatment to conserve soil. Satellite remote sensing data are useful in carrying out integrated sustainable development planning  at manageable units. The remote sensing data can be used for the preparation of a set of resource maps such as surface water bodies, ground water potential zones, ground water recharge site, type of soil, existing land use patterns etc. and the combination of these data with other information like meteorological data, socio-economic factors etc. can be used to identify the priority areas for various land use to meet the needs of the people without disturbing the ecology.
The application of remote sensing in agriculture has also produced praiseworthy result. Now remote sensing technology is capable of providing information about various agricultural resources which influences agricultural production directly viz. land, water and weather and also the related one such as forests and access to other agricultural information.
4. Expert System and Natural Resources management:
The knowledge related to natural resources may be structured or unstructured and can be organized in highly structured form to meet the requirements for making utilization of information received form the various sources. This calls for the utilization of Artificial Intelligence and Knowledge Engineering methods to represent and infer with such knowledge; software engineering techniques to manage system developments, information and control flows of models and data; intelligent system technology to process and display data. The domain specific knowledge used in decision-making can be represented in symbolic or asymbolic formalism in the most explicit and formal manner. The knowledge representation scheme of natural languages is the most sophisticated and this very act of knowledge representation have been captured by logic, and different formal methods such as propositional logic, predicate logic, fuzzy logic, semantic networks, frames etc. and related techniques have been developed to represent various types of knowledge that can be used by expert systems in decision making and reasoning.
In the recent years the asymbolic approach  for intelligence modeling has evolved in which neural networks, modelled after human brain, are connected to represent knowledge and to make inferences. In such systems the knowledge is encoded by connection strength and acquired through learning process . In geographical analysis, decision-making system like intelligent spatial decision support systems  have been developed to reason with structured or loosely structured knowledge. Here, again the expert system shell play core role in directing control flows and information flows. It provides facilities to represent and store domain specific knowledge acquired form experts or learning examples. It can also contain meta-knowledge for inference control and possess capacity of reasoning and inference of decision support system. It is the brain of decision support system. In another attempt by Leung , SDSS shell has been developed using fuzzy logic based expert system shell for the purpose of building SDSS to solve specific spatial problems in effective and efficient manner.
The object oriented programming approach has been found to be very effective approach in developing systems for the natural resources management. The object-oriented approach provides the way for user to perceive reality. It provides an effective user interface, enhances data reusability, maintainability, and extensibility through data encapsulation and inheritance. The object oriented database system overcomes many limitations such as limited query processing, lack of semantics, lack of extension mechanism, lack of handling recursions and data version etc. of relational database and thus provides somewhat better option for application in expert system.
Based on this approach Gahanna , Worbys  have developed system for geographical information manipulation and decision-making. Most of the current object oriented G I S designs emphasise processing of geometric data models and data structures. However, the concept based object oriented GIS by Leung  provides a spatial conceptual model which comprises spatial semantics, fundamental to spatial analysis, and an object oriented data model which provides an appropriate and effective representation of the spatial conceptual model for better database management. The remotely sensed data require their analysis, interpretation and preparation of databases at various levels of use and application in different decision-making systems. The need of interaction with these databases in the monitoring and identification of earth resources creates the opportunity for application of Artificial Intelligence in general and that of Expert Systems in particular. Different expert systems have been developed and are being developed in different institutions for every stage of application of remote sensing techniques in the natural resources management. National Remote Sensing Agency (NARSA), Bangalore has developed an expert system  in PROLOG, to access the databases of acquired/processed remote sensing data. This expert system provides access to databases by making queries relevant to type of data needed for a particular application. NARSA has also developed expert systems for interpretation of remote sensing data / images emulating the experiences and logical reasoning process used by human experts to derive information from remote sensing data/ images. Such expert systems  have been developed for interpretation in the area of soil studies, land use, land capability, geology etc. GIS use remotely sensed data of natural resources along with various ancillary information such as map data, statistical data, meteorological data etc. In such applications AI is again needed to interpret ancillary data for GIS and to provide natural language interface to GIS. Very few such successful systems have been developed. A rule based expert system to identify land use/ land cover using satellite data has been developed by K. Ramani and N. C. Guatam  . The other expert system  for image interpretation taking into account the elements of image interpretation for identification of land use / land cover classification, has been developed at APOLLO work station. This expert system has been developed in PROLOG. An expert system for soil classification was developed by Rao, et al  in M-PROLOG that can make identification and classification of soil up to great group levels. Several other expert systems have been developed and used successfully in various areas such as land form study, ground water, minimal exploration etc. For the urban planning  and urban scene analysis, different systems have been described by different scientists using different AI techniques. VISIONS system by Hanson and Riseman  , SPAM system by Meken  , are some examples of expert systems used in urban scene analysis. Junsen  has described a prolog-based approach for urban analysis. This incorporates modules for roads, rivers, settlements, texture analysis etc. It uses both the qualitative and quantitative knowledge about urban structures. Nagao, Matsuya, and Mori  have developed system for aerial urban scene analysis using black board approach. In this system there have been designed two object detection sub-systems that handle different subgroups of objects like roads, house etc. These sub-systems are picture driven and model driven. Picture driven system searches the existence of local areas by combining the characteristics regions while the model driven system picks up the candidate region by using the spectral and spatial relationship with already recognized objects.
4.1. Proposed Methodology:
The expert system comprises of three major modules viz., (1) Knowledge base (2) Inference engine and (3) User interface. In addition the expert system contains the following additional components
- Knowledge acquisition
- Work place
- Explanation (justifier)
- Knowledge refining
The proposed expert system development for natural resources management includes the following modules.
- Object identification: From the satellite images the various land-based natural resources are identified by image analysis. The image clustering and classification are done using fuzzy neural network model. The neural network parameters are estimated by supervised learning process for the known sites. The parameters thus estimated are used for identifying the different types of natural resources like forest, water resources, agricultural land and minerals for new sites.
- Modeling: The various land based image object identified above are further used to calibrate a mathematical model, the output of which yields the quantitative database for the natural resources
- Optimization: For optimal utilization of natural resources an intelligent decision support system needs to be developed by formulating a cost minimizing objective function while maximizing the economic benefits under physical constraints. The optimization can be implemented by formulating a neural network model. The parameters of the model could be estimated with the historical cases. These parameters could be used for future planning.
In this millennium the large production of food and agricultural commodities under condition of diminishing per capita arable land, irrigation water resources and expanding biotic and abiotic stresses will be required. So the efficient management of agriculture is now focal objective of any developing or developed country. Efficient agricultural management involves appropriate application of production and conservation practices for development of land and water resources. Substantial increase in crop production is possible by bringing additional land under cultivation and improved crop management technologies. The concept of precision farming requires scientific land and water use planning and extensive study of natural capital stocks and natures services. In the emerging knowledge intensive agricultural era, international cooperation is vital for taking benefits of new technologies to those who have so far been by passed by new knowledge and techniques.
The emerging biological and knowledge millennium will need appropriate national, regional and global system of knowledge creation and sharing. The agriculture in this 21st century  will be based on appropriate use of biotechnology, information technology and ecotechnology. The technique of remote sensing has been applied to bring more land under cultivation, flood monitoring, estimation of crop acreage and production, crop condition assessment etc. and data on socio-economic needs, integrated land water resource maps, highlighting priorities areas for action on agricultural development, soil conservation and afforestation are being prepared. Decision-making in agricultural management depends on the availability of such information. This again calls for the creation of information system for agricultural applications. The satellite information and conventional systems are required to be fused to create farmer’s based information system. Such integration will have tremendous scope for optimum crop management strategies at micro level. Remote Sensing, Geographical Information System (GIS), Global Positioning System (GPS) and the information technology together provide the requisite support for policy makers to take appropriate, precise and faster in situ and ex situ decisions. Thrust is being given to create database, full electronic access to available knowledge in the libraries of various information systems using AI techniques, and other agricultural information systems  in different institutions.
Obviously, Expert systems have large impact and scope for the application of remote sensing techniques in the monitoring and management of natural resources. Thus seeing the potentials of field large effort are required to be made to develop expert systems for natural resources management.
- Duda, R. O., et. al, “Development of the PROSPECTOR Consultation System for Mineral Exploration.”, The National Science Foundation, Artificial Intelligence Centre, SRI International, Menlo Park, October 1978.
- Duda, R, O, Et Al, ” A Computer Based Consultant for Mineral Exploration.” SRI International Menlo Park, CA, September 1979.
- Martime, J. & Oxman, S., “Building Expert Systems: A Tutorial.” Ph, New Jesery, 1988.
- Quereshy. M. N. , “Natural Resource Data Management”, Proc. of Workshop on Natural Resource Database Management System, (Patna) Dept. of Science and Tech. New Delhi, 1986.
- Leung, Y, “Intelligent Spatial Decision Support System,” Berlin, Springer-Verlag, 1997.
- Aribib, M.A. (Ed), “The Hand Book of Brain Theory and Neural Networks,” Cambridge; MIT, 1995.
- Leung, Y., ” Inference with Spatial Knowledge: An Artificial Neural Network Approach,” Geographical Systems, 1,1994.
- Leung , Y And Leung , K.S . ” An Intelligent Export System Shell for Knowledge Based Geographical Information System 2, Some Applications, International Journal of Geographical Information System 1993.
- Gahned, M.N. & Robot, S.A. “An Intelligent Object Oriented Geographical Information System”, International Journal of Geographical Information System 2, 1998
- Work M.F., ” Object Oriented Approaches to Geo- Referenced Information,” International Journal of Geographical Information System (1994)
- Leune, Y., Leung, K.S. & He J.Z., “A Concept Based Object Oriented GIS” International Journal of Geographical Information Science, 1993.
- Joseph F. ” Presidential Address at National Symposium Of ISRS On” Remote Sensing of Environment with Special Emphasis On Green Revolution” Journal of The Indian Society of Remote Sensing Vol. 23 December 1995.
- Deekshatulu, B.L. “Management of Natural Resources -Use of AI,” In AI and Expert System Technology in Indian Context”, Ed. By V.V. Sharma, Deekshatulu, Vishwandhan, TMH, New Delhi, 1991.
- Sharma, L.C. “Knowledge -Based Interpretation of Remotely Sensed Data” M.S. Thesis, IISC. , Bangalore, March 1991.
- Ramani, K. & Gautum, N.C., ” Expert System, for Identification of Land Use/Land Cover Categories from Satellite Data”, Proceedings of the Seminar on AI, Expert System & Knowledge Based System, April 1990, ISRO Hqts., Bangalore.
- Rao, D.P. ” Management of Natural Resources using Expert System Preliminary Result of Some Case Study ” Expert system Technology in Indian Context Ed. by V.V. Sharma, Deekshatulu, Vishwandhan, TMH, New Delhi, 1991.
- Rao, B.R.M., Sukanaya, H., Venkatraman; Parariy A., H., & Sharda” Development of an Expert System for Classification of Soil” Proceeding of Seminar on AI, ES and KB System, ISRO, Banglore, 1990.
- T. Kimal et. al (Ed), “Expert System Application to Urban Planning,” Springier Verlag 1990.
- Honson, A. R. & Riseman, E.M., “Visions: A Computer System for Interpreting Scenes”, Academic New York Press, 1978.
- Meksown, D.M., Wilson A.H., “Automating Knowledge Acquisition for Aerial Image Interpretation”, Computer Vision Graphics, 1990.
- Jensen, Dm, ” Knowledge Based Classification of an Urban Areas Using Textures and Context in Landsat Thematic Mapper Data” Photogrammetric Engg. and Remote Sensing, 56,6, pp 899-9041, June 1990.
- Nagioya, M., Matsuyama, T., & Mori, H.” Structural Analysis of Complex Aerial Photographic Blackboard System”, 1998, Addison Wesley.
- National research Council, USA, “Precision Agriculture in 21st Century”, National Academic Press, Washington D.C., USA, 1997.
- Rao, D.P., ” Valuable Information Aid”, The Hindu Survey of Indian Agriculture 1999, New Delhi