High-Level Project Summary
Impact crater cataloging is an important tool in the study of the geological history of planetary bodies in the SolarSystem, including dating of surface features and geologic mapping of surface processes. Catalogs of impact craters havebeen created by a diverse set of methods over many decades, including using visible or near infra-red imagery and digitalterrain models.We present an automated system for crater detection and cataloging using a digital terrain model (DTM) of Mars In the algorithm craters are first identified as rings or disks on samples of the DTM image using a convolutional neuralnetwork with a UNET architecture, and the location and size of the features are determined .
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Information on crater populations and spatial distributions provide important constraints on the geological history of planetary surfaces. Regional differences in crater distributions and population characteristics can be used to constrain geologic processes and stratigraphy.Automated CDAs have tunable parameters that can be optimized for the imagery or elevation data set being processed. In designing the algorithms, a curated list of crater locations and images are used in a “training” step to adjust these parameters. Once trained, the CDA can be applied to larger datasets from the same body or even applied to different planetary bodies through transfer learning.In this work, we used an automated CDA based on a Convolutional Neural Network to identify circular crater–like features in a martian Digital Terrain Model (DTM). I perform three experiments with the CDA to characterize its performance on the DTM under various assumptions.One of the first large global databases for Mars was created by Barlow (1988) using printed maps from Viking or-biters and included 25,826 craters with a diameter greater than 8km. This dataset has been updated since then(Barlow, 2003) with 42,283 craters, and other datasets are available (Rodionova et al. (2000) with 19,308 craters,The most comprehensive dataset for Mars craters is thatderived from the Thermal Emission Imaging System.Instrument by Robbins and Hynek (2012). The Robbins and Hynek (2012) dataset includes 383,343 craters with diameters greater than 1km, including 30,473 craters above 8km diameter. These craters were identified in 256 pixel/-degree resolution THEMIS IR imagery using a customized manual image processing pipeline.
Space Agency Data
Impact craters in planetary science are used to date and characterize planetary surfaces and study the geological history of planets. It is therefore an important task which traditionally has been achieved by means of visual inspection of images. The enormous number of craters, however, makes visual counting impractical. The challenge in this RAMP is to design an algorithm to automatically detect crater position and size based on satellite images.We have built an AMI . You can sign up and launch an instance . When asked for the AMI, search for mars_craters_2_users. Both ramp-workflow and this kit are preinstalled, along with the most popular deep learning libraries. We will use p3.2xlarge instances to train your models. They cost about 3€/hour. Alternatively you can also use p2.xlarge instances which cost 1€/hour and 3-4x slower than p3.2xlarge
Hackathon Journey
This hackathon gave us some motivation for our talent.This hackathon will be a moment for us.we gained some teamwork,hardwork and time management by this hackathon.This hackathon journy was a superb journy for us.Thanks for giving us this oppurtunity.There are lots of talented people in this world.By this hackathon they can prove wat they are.And finally we are thanking to this hackathon team and nasa.
References
R. E. Arvidson. Morphologic classification of Martian craters and some implications. Icarus, 22(3):264–271, 1974. ISSN 10902643.
doi: 10.1016/0019-1035(74)90176-6.
N. G. Barlow. Crater size-frequency distributions and a revised Martian relative chronology. Icarus, 75(2):285–305, 1988. ISSN 10902643. doi: 10.1016/0019-1035(88)90006-1.
N. G. Barlow. Revision of the ”Catalog of large martian impact
craters”. In Sixth International Conference on Mars, 2003.
N. G. Barlow. A review of Martian impact crater ejecta structures
and their implications for target properties. In T. Kenkmann,
F. H ̈orz, and A. Deutsch, editors, Large Meteorite Impacts III.
Geological Society of America, mar 2005. ISBN 9780813723846.
URL https://doi.org/10.1130/0-8137-2384-1.433.
N. G. Barlow and C. B. Perez. Martian impact crater ejecta morphologies as indicators of the distribution of subsurface volatiles.
Journal of Geophysical Research, 108(E8):5085, 2003. ISSN 01480227. doi: 10.1029/2002JE002036. URL http://doi.wiley.com/10.1029/2002JE002036.
F. Chollet. Keras, 2015
Tags
#Mars,#Space,#Nasa,#Images,#ArtificialIntelligence