The identification of valley networks and channels is an essential tool for geomorphological interpretations of the fluvial, glacial and volcanic history of Mars. While the creation of valley networks by erosion is an accepted hypothesis, the flow of water as the sole cause has recently been put into question. To investigate the origin of the networks, their detailed properties have to be mapped at a global scale. In previous attempts of computer-generated global mapping, manual verification of the results was necessary. Herein, we present a novel algorithm to automatically extract valley networks in terms of extremal lines and compare the extraction results with an area that has already been manually mapped in the literature. Extremal lines are a subset of the topological skeleton from an elevation map. The skeleton encodes the essential information and can be iteratively simplified to obtain different levels of detail in the elevation map. The hierarchization process is thereby fully automatic and does not involve any algorithmic parameters. Using the measure separatrix persistence, we are able to assess the topological importance of each element of the skeleton. This enables a reduction to the most dominant extremal lines. This topological analysis allows for an unbiased extraction of all extremal lines in an elevation map. The data analyst is provided with a slider to choose an appropriate level of detail for further analysis.


List of all publications