Michael Dille

Dr. Michael Dille


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 December 2024 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Carl LegleiterCarl Legleiter
Carl Legleiter
PKPaul Kinzel
Paul Kinzel
MDMichael Dille
Michael Dille
MVMassimo Vespignani
Massimo Vespignani
UWUland Wong
Uland Wong
JCJennifer M Cramer
Jennifer M Cramer
Image velocimetry has become an effective method of mapping flow conditions in rivers, but this analysis is typically performed in a post-processing mode after data collection is complete. In this study, we evaluated the potential to infer flow velocities in approximately real time as thermal images are being acquired from an uncrewed aircraft system (UAS). The sensitivity of thermal image velocimetry to environmental conditions was quantified by conducting 20 flights over four days and assessing the accuracy of image-derived velocity estimates via comparison to direct field measurements made with an acoustic Doppler current profiler (ADCP). This analysis indicated that velocity mapping was most reliable when the air was cooler than the water. We also introduced a workflow for River Velocity Measurement in Approximately Real Time (RiVMART) that involved transferring brief image sequences from the UAS to a ground station as distinct data packets. The resulting velocity fields were as accurate as those generated via post-processing. A new particle image velocimetry (PIV) algorithm based on staggered image sequences increased the number of image pairs available for a given image sequence duration and slightly improved accuracy relative to a standard PIV implementation. Direct, automated geo-referencing of image-derived velocity vectors based on information on the position and orientation of the UAS acquired during flight led to poor alignment with vectors that were geo-referenced manually by selecting ground control points from an orthophoto. This initial proof-of-concept investigation suggests that our workflow could enable highly efficient characterization of flow fields in rivers and might help support applications that require rapid response to changing conditions.

18 March 2025 in Frontiers in Remote Sensing
Verified
Journal Article
Open access
PDF file
MDMichael Dille
Michael Dille
MVMassimo Vespignani
Massimo Vespignani
JBJonathan Bruce
Jonathan Bruce
UWUland Wong
Uland Wong
Current streamgaging processes for river flow rate estimation are typically slow and often hazardous, leading to inadequate coverage across national waterways. This paper presents a semi-autonomous aerial monitoring system that is designed for rapid river flow gaging, building upon a recently developed sensor package that is mounted beneath a small uncrewed aerial vehicle. This package consists of, among other instruments, a mid-wave infrared camera that can be used to detect minute thermal variations in the water surface, from which a particle image velocimetry algorithm is used to extract flow estimation. The design and testing of this sensor package and velocimetry algorithm for field evaluation are discussed, and a simulation environment facilitating the development of algorithms for automatic a priori and live-adaptive vehicle trajectory planning is presented. The simulation environment captures a physically based approximation of vehicle flight characteristics, contains digital terrain models of field test sites, and incorporates water surface flow maps generated from numerical flow simulation data and real-world measurements. Field and simulation results are presented validating the design of the sensor package and the use of simulation as a digital twin for aerial streamgaging development. This framework and the lessons learned to date lay the foundation for accelerated improvements in waterway measurement for both routine and disaster response purposes requiring rapid deployment in novel locations.