Characterization of the land-cover and land-use by shape descritors in two areas in Ponta Grossa, PR, BR

This study aims to identify the spatial characteristics of the land use and occupation regarding shape (geometry) that these uses are presented in two (2) Landscape units of Hydrographic Basin Pitangui by satellite Geographic Object-Based Image Analysis (GEOBIA). The two (2) croppings of landscape units come from the IRS satellite - P6 (Indian Remote-Sensing Satellite) ResourceSat LISS III sensor. The preprocessings were performed, geometry and atmosphere correction, and the image became segmented in order to extract from it the spatial characteristics and identify the relationship between descriptors of the segments, which were generated from GEOBIA and the use, and occupation of the land. An relevant aspect of this study is that one (1) of landscape units is located in the First Paraná Plateau and the other in the Second Paraná Plateau where there are significant differences concerning geology. The GEOBIA allows segmenting the image by means of regions growth technique and it extracts a relational database considering the spectral, spatial (shape) and texture attributes of the image digital levels. Through this database, we used the descriptors of the spatial attribute (shape) to analyze through graphics and specialization of them with color, to check the relationship between these shape descriptors. It is possible to identify the landscape morphology with this
analysis, where small areas do not have defined shapes and the larger areas with geometrically defined shapes. It highlights the importance of analyzing the landscape and its relations with the use and pattern recognition and also the highlight of GEOBIA on this issue which enables to corroborate by the descriptors the visual analysis that the interpreter performs allowing more security and less subjectivity.