Drones4Discovery: High resolution field data using drones ...€¦ · using drones & digital...

Drones4Discovery: High resolution field data

using drones & digital pipelines

Steven Micklethwaite

Acknowledgements: Sam Thiele, Rob Glasgow, Monash University

“To sense our world in greater detail, with greater qualityover greater areas, and to develop the technology and dataanalytics that can do so for the benefit of Australianresearch, industry and education.”

DRONE DISCOVERY PLATFORM

Capability Developers

Capability Users

.

• Advanced remote sensing & sampling

technologies + HPC/Cloud + analytics

• For Research-Education-Industry Impact

• Driving interdisciplinary developments

(Science-Engineering-IT-Arts)

• Regulatory oversight with CASA, including

pilots (protecting research)

• Leveraging MTRPs LiDAR

Hyperspec

dolerite

tonalite

gabbro

Data Analytics

Visualisation

Flagship Workhorse

LiDAR Hyperspec (VNIR-SWIR)

Visualisation

6.1 X 104 nT

5.8 X 104 nT400 m0 m

UAV-mag

Thermal MEMS gravimetry

Original image (4000 x 6000 pixels)

Ground truth labelled image

Manual digitisation: 51 mins

HSV input, random sampling (500), scalevariation

Time taken: 82.34 seconds

Pixel accuracy: 75.05 %

Vegetation

Dolerite

Tonalite

Background

Water

Aplite

Dacite Gabbro

DATA ANALYTICS – CNN EXAMPLE

DATA FLUENCY: Photogrammetry example

e.g.3D point cloudOrthophotoDEM

Bemis, Micklethwaite et al., JSG 2014

FIELD OBSERVATIONS

DATACOLLECTION

DATAPROCESSING

INTERROGATIONEXPLORATION

e.g.FiltersScalar productsSome ML/DL

FIELD BASE MAP

RAW DATA PHASE … … INFORMATION PHASE

ANALYSISINTERPRETATION

VIZ. &REPORTING

… TIME (opportunity/financial cost)

EXTENDING PHASE

EXAMPLE LOCATION

Century Mine, Queensland

DATA PROCESSING: Photogrammetry example

Feature SelectionSIFT Algorithm

Bundle AdjustmentInterior and exterior

Geometry Infilldense point cloud

Texture Mappingphoto-realistic model

Reprojectionorthophoto

Mesh Generation3D model

InterpolationDEM

Courtesy M. Roach (UTAS)

Bemis, Micklethwaite et al., JSG 2014

Photogrammetry (SfM-MVS);

Metashape/Photoscan, Pix4D etc

INTERROGATION & EXPLORATION

METHOD DEVELOPMENT

Thiele et al., Solid Earth, 2017

Rapid fault, fracture, lithology

contact mapping

• Least-cost path solvers.

• Geologist remains involved in

interpretation.

• Future – CNN or other automated

ANALYSIS & INTERPRETATION

15 kilometres of dyke margin mapped @ 5-10 cm spatial resolution (fidelity)50,000 orientation and thickness measurements

e.g. Figshare

CLOUD COMPUTE: drones to information rapidly

e.g. PreViz, CAVE2

CONCLUSIONS

& QUESTIONS