Suchergebnisse

Implementation of environmental legislation can be supported by the production of maps with two information levels, one which identifies areas where a legal rule is enforced and another which identifies areas where there is an infringement of that rule. This paper investigates the efficiency of remote sensing methods and techniques for the production of such a map regarding a specific law case, the Greek Forest Law. To do this, we identified, with the use of KVR-1000 very high resolution imagery, potential illegal buildings in forests and forest areas, which were mapped at a scale of 1:5000. For this purpose, an image processing method was developed which integrates spectral and first-order texture signatures, detects disturbances in forest environment, and classifies them in order for the buildings to be identified among other disturbances. The texture signatures were selected after an investigation of the performance of first- and second-order statistical texture methods in detecting disturbances in forest environment. By the method developed, buildings are mapped by segmented parts of their edges. Based on a sample of 163 buildings, the accuracy of the method was calculated at approximately 89%, i.e. 11% of the buildings included in the sample were not detected. The map produced would probably not be acceptable in court, however it can adequately support administrative procedures in ministries, forest services, etc., regarding identification of illegal building, incorporation of forest areas in an existing city plan, organization of the forest and forest area register book, etc.