| dc.description.abstract | This investigation proposes the use of geographic information system (GIS) for delineating landforms, starting from Digital Terrain Model (DTM) of great detail. The DTM is a product of LIDAR data (Light Detection and Ranging), and the methodology uses his derives: slope, curvature y texture. The proposed method can be replicated for extended areas with high detail level information. The study was done in a 5 square kilometer area located between the province of Cañar, canton Azogues, sector Jatumpamba, and the province of Azuay, canton Paute, sector San Cristóbal.
We analyzed the criteria and the combination of algorithms of the MDT to delineate the different landforms, and also the different software’s which process the information to generate MDT and it’s derives, with the unsupervised classification algorithm (ISODATA) proposed by "Iwahashi & Pike (2007)" which uses a geometric signature composed by slope, convexity and texture,
The results consisted of three types of classifications, in 8, 12 or 16 classes. The hierarchy is determined firstly by dividing the slope in four classes: gentle, moderate, steep and very steep. Secondly the convexity generates 2 types more: high and low, and lastly the texture divides in two classes; fine and course. Afterwards the landforms were submitted to a qualitative and quantitative verification in the field, which obtained an accuracy of 98% in texture, 99% in convexity, and 100% in slope, which made it possible to verify the feasibility of the landform classification method. | |
