| Natural resources management has become an important topic concerning the environment. Sound management encourages the protection of our environment and supports sustainable development. Policies and regulations regarding the environment require the development of monitoring systems. Remote sensing technologies are important tools in the elaboration of these systems. Regional monitoring permits resource managers to obtain accurate information on the status of these resources. Based on this information, it is possible to plan actions contributing to the preservation of the environment.; In this context, forest region monitoring requires the updating of forestry maps. The update can take place automatically with the assistance of change detection methods. The procedure takes place in several stages, where the possibility of introducing errors at each step exists. The final map may subsequently include numerous false alarms that must be corrected in order to obtain a reliable document. The traditional method is often preferred, because the interpreter identifies truly disturbed areas with great precision. According to the visual interpretation, the photo-interpreter uses various criteria, such as texture, appearance or context of the area to be identified. In order to reproduce this work automatically, it is necessary to develop a similar approach, such as an oriented-object method. In addition, the integration of information similar to that of the interpreter requires the use of intelligent methods. At the present time, these methods analyse the attributes of each object more accurately than conventional methods. In this case two groups of methods are distinguished; those based on the case-based reasoning (CBR), such as decision trees or neural networks; and those whose learning is supervised by the user, known as experts systems (ES).; For change identification in forest region, a method applying the second approach is preferred. A rule-based system has been elaborated for a study site in Labrador. Results of this approach reveal that the system provides more accurate results than the methods of the first approach. Indeed, the first group of methods depends on the samples. Otherwise, the rule-based system permits the conservation of the done work. |
