Authors: Joris van den Bossche, Alexandre Boucaud, Frédéric Schmidt & Anthony Lagain
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.
Open a terminal and
- install the
ramp-workflowlibrary (if not already done)
$ pip install git+https://github.com/paris-saclay-cds/ramp-workflow.git
- Follow the ramp-kits instructions from the wiki
Get started on this RAMP with the dedicated notebook.
We have built an AMI on the Oregon site of AWS. You can sign up and launch an instance following this blog post. When asked for the AMI, search for mars_craters_2_users. Both ramp-workflow and this kit are pre-installed, along with the most popular deep learning libraries. We will use p3.2xlarge instances to train your models. They cost about 3€/hour. Alternativaly you can also use p2.xlarge instances which cost 1€/hour and 3-4x slower than p3.2xlarge.
Go to the ramp-workflow wiki for more help on the RAMP ecosystem.