lhospitallky / geoceles Goto Github PK

Geoceles (© 2018 Sampsa Pursiainen) is a simple tool for finite element based forward and inverse simulations in geoimaging of small planetary bodies. With Geoceles, one can segment a realistic multilayer geometry and generate a finite element mesh, if triangular surface grids (in ASCII DAT file format) are available. The current version also allows using a graphics card to speed up the mesh segmentation as well asforward (lead field) and inversion computations. Geoceles inherits from the Zeffiro Interface for Brain Imaging (© 2018 Sampsa Pursiainen). Instructions can be found at:

https://github.com/sampsapursiainen/zeffiro_interface/wiki

Article in which Geoceles has been used:

Sorsa, L. I., Takala, M., Bambach, P., Deller, J., Vilenius, E., Agarwal, J., Carroll, K., Karatekin, O., & Pursiainen, S. (2020). Tomographic inversion of gravity gradient field for a synthetic Itokawa model. Icarus, 336, 113425.

Articles in utilizing the IAS inversion method for asteroids:

Pursiainen, S., & Kaasalainen, M. (2013). Iterative alternating sequential (IAS) method for radio tomography of asteroids in 3D. Planetary and Space Science, 82, 84-98.

Pursiainen, S., and M. Kaasalainen. "Detection of anomalies in radio tomography of asteroids: Source count and forward errors." Planetary and Space Science 99 (2014): 36-47.

Pursiainen, S., & Kaasalainen, M. (2014). Sparse source travel-time tomography of a laboratory target: accuracy and robustness of anomaly detection. Inverse Problems, 30(11), 114016.

Pursiainen, Sampsa, and Mikko Kaasalainen. "Electromagnetic 3D subsurface imaging with source sparsity for a synthetic object." Inverse Problems 31.12 (2015): 125004.

Quick tips to start:

• Install Matlab Parallel Computing Toolbox

• If using Geoceles on a standard laptop or desktop computer without GPU, edit the row 4 of the geoceles.ini file into "UseGPU 0". The default is 1.

• Create a surface model in which layers are numerally in an ascending order from the innermost to the outermost, i.e., from inside Deep 1, Deep 2, Surface 1, Surface 2, etc.

• The surfaces and sensors can be scaled (via "scaling"). For example points distributed over a unit sphere can be scaled to 10 km orbit size using 10000 as the scaling factor. Similarly any surface model can be scaled to its actual size.

• The detail layers 1-4 can be placed deep inside and they render fine.

• One needs to choose the meshing resolution in the main window (it is the element size in meters), so that the meshing routine works appropriately. For example, 3 meters for a 200 m diameter object.

• After that the volume mesh can be created (volume mesh button).

• One needs to choose a suitable number of basis functions (lead field size in main window). For example, 500. The default in the box is 50000 which is kind of a large value.

• Then by pressing the lead field button the lead field will be created. One can choose between the vector gravimetry and scalar gravimetrly in the "imaging method" drop-down box. For me, the scalar version has been easier regarding choosing the inversion parameters.

• After generating the lead field, one needs to press" basis interpolation" if the LF interpolation option has not been checked. This will enable inverting the data and creating a 3D image out of the reconstruction.

• Then one needs to generate synthetic data (if there is no real data available) with the dialog box that opens in the "forward tools" menu. The real data can be imported in the same menu as well. In the data creation toolbox, one can create a small density anomaly to be localize. One can also change the layer densities in the main window before calculating the data. In that case, one needs to press the "update density" button after the update.

• Now the system can be inverted with one of the tools inthe inverse tools menu. For example IAS MAP Estimation. That will require shape and scale parameters as regularization parameters as the input. Also the number of iterations needs to be fixed. The reconstruction will be obtained after pressing start.

• The reconstruction can be visualized after the inverse iteration by choosing Recon. Volume or Recon. Surface as the visualization type in the main window and then clicking either visualize volume or visualize surfaces.