Object-based Dune Analysis: Automated dune mapping and pattern characterization for Ganges Chasma and Gale crater, Mars

Abstract

A method that enables the automated mapping and characterization of dune fields on Mars is described. Using CTX image mosaics, the introduced Object-based Dune Analysis (OBDA) technique produces an objective and reproducible mapping of dune morphologies over extensive areas. The data set thus obtained integrates a large variety of data, allowing a simple cross-analysis of dune patterns, spectral and morphometric information, and mesoscale wind models. Two dune fields, located in Gale crater and Ganges Chasma, were used to test and validate the methodology. The segmentation of dune-related morphologies is highly efficient, reaching overall accuracies of 95%. In addition, we show that the automated segmentation of slipface traces is also possible with expected accuracies of 85-90%. A qualitative and quantitative comparison of the final outputs with photointerpretations is performed, and the precision of the directional characterization of the dune patterns is evaluated. We demonstrate a good agreement between the OBDA outputs and the photointerpreted dune morphologies, with local trend deviations below 45°f or 80-95% of the mapped areas. Because the developed algorithm is tuned for the recognition of linear features from the imagery, the slipfaces of small barchans can be preferentially overlooked owing to their small extent at the spatial resolution of the CTX mosaics. Dune types composed of longer linear morphologies are much better represented, including correct mapping of secondary structures. Having proved the effectiveness and accuracy of the mapping procedure, we discuss its future applications for the improvement of dune catalogs on Mars.

Dune‐like dynamic of Martian Aeolian large ripples

“…The COSI-Corr displacement maps are further improved by removing jitter artifacts ( Figure S2). We then used the Object-based Ripple Analysis technique [Vaz and Silvestro, 2014] to characterize the directional distribution of the LRs, evaluate their spatial distribution and correlate bed form pattern characteristics with migration rates and morphometric settings [Vaz and Silvestro, 2014;Vaz et al, 2015] ( Figure S3).…”

Migrating meter-scale bedforms on Martian dark dunes: Are terrestrial aeolian ripples good analogues?

“…These methods were previously introduced and validated Vaz et al, 2015a) and are in this study used to: 1) map the dune patterns at a regional scale using a CTX (Mars Reconnaissance Orbiter Context Camera) mosaic; 2) map and characterize the morphology and migration rates of slipfaces using a HiRISE digital terrain model (DTM) and stereo images; 3) map meter-scale ripples and integrate the mapped patterns with the migration displacement fields and a simplified topographic wind effect model. The final stage of the data analysis consists in the clustering of the mapped features at two different scales.…”

“…This is the same type of output described in Vaz et al (2015a), and is used in this work as input to a principal component analysis (PCA) and clustering algorithm (see Section 2.4.1). To apply these techniques we computed for each grid node: the length-weighted mean vector magnitude ( R L ); the circular standard deviation (CSTD); several parameters that characterize the trend of the mapped structures, namely the azimuth and frequency of the primary kernel modes (k mode1 and k freq1 ) and the modal ratio (k mrat Þ which measures the degree of bimodality of the trends ; the mean, standard deviation and median lengths of the mapped features ( L, s L and med L respectively); and the total number of aeolian features as well as the number of features classified as slipfaces (n all and n slipfaces ).…”

Semiautomated Identification and Characterization of Dunes at Hargraves Crater, Mars

“…Moreover, manual digitization of the individual dune parameters (e.g., shape, size, area, and elevation) from the lower spatial resolution THEMIS images is a very arduous, labor‐intensive, and time‐consuming task. Thus, a semiautomated method from high spatial resolution images, for example, the Context Camera (CTX; Malin et al, 2007) or the High Resolution Imaging Science Experiment (HiRISE; McEwen et al, 2007), is more efficient at maximizing the extraction of significant surface morphological information about Martian dune fields (Bandeira et al, 2010; Vaz et al, 2015).…”

“…Extracting dune parameters from orbital data sets is not only cumbersome but also involves some degree of user bias (Hugenholtz et al, 2012). That user bias yields difficulties in observational integration and standardization (Vaz et al, 2015). Recently, studies have applied automated detection of dune characteristics through the object‐based image identifications (e.g., Bandeira et al, 2010, 2011; Sholes et al, 2013; Silvestro et al, 2013; Vaz & Silvestro, 2014; Vaz et al, 2015).…”