|About this Abstract
||Materials Science & Technology 2019
||Data Science for Material Property Interpretation
||Multi-modal Data Fusion and 3D Reconstruction of Serial Sectioning Data
||Megna Shah, Sean Donegan, Michael Uchic
|On-Site Speaker (Planned)
A more complete representation of microstructure often requires more than one sensor/modality. A challenge is to register all the modalities together and fuse the important signals from each into one representation. The workflows are further complicated when representing the microstructure in 3D. Here we describe data registration and fusion of 3D serial-sectioning data collected at AFRL. Optical images, BSE images and EBSD maps were collected on each section. Optical images were key for stack registration, BSE images provided the clearest representation of pores and precipitates, and EBSD maps offered crystallography. A non-linear distortion correction was applied to the BSE and EBSD data to register the three modalities. Finally, this data also was characterized by HEDM, and additional transformations were required to bring it into the correct reference frame for registration. The workflow to determine and apply all of the transformations, and reconstruct the data using DREAM.3D will be discussed.
||Definite: At-meeting proceedings