Using the Gradient Vector to Generate a Slope Class Map

Camillo Jorge Santos Oliveira1, José Luiz de Souza Pio2,3, Gilson Giraldi4, Arnaldo de Albuquerque Araújo1
1Núcleo de Processamento Digital de Imagens - NPDI/DCC/ICEx/UFMG
2Laboratório de Visão Computacional e Robótica - VerLab/DCC/ICEx/UFMG
3Departamento de Ciência da Computação - DCC/UFAM
4Laboratório Nacional de Computação Científica - LNCC/MCT

Abstract. This work aim to obtain the slope classes map from the topographic digitalized map information using a Geographic Information System (GIS). To do this, we use the directional derivative concept. It is related directly with the gradient vetor through the following property: "the maximum value of the directional derivative in a point is equal to the gradient vector modulus in that point and it occurs in the direction of this vector". Based on this property, we choose to subdivide the gradient field domain small regions (elements or cells) and we estimate the gradient modulus inside of each element.

[ project home page | full version | contact authors ]


Exemple of the geographic map, where we can see
a topographic map (contour lines).
Exemple of a slope classes map.
Having a topographic map in the scales 1:20.000 or 1:25.000, we digitalized the contour lines using a Geographic Information System (GIS) and we generate a file in the DXF format. We extract from DXF file, the Universal Transverse Mercator (UTM) coordinates from contour lines and we transform them to another small coordenate system (e.g. 500x500, 750x750, or ...).
Graphical sketch of the method applied to three contour lines. It was obtained after execution from line 3 until 14 of the algorithm. Graphical sketch of the slope classes map. It was obtained after
execution from line 14 until 16 of the algorithm.