Recognizing Non-Stationary Hydrodynamics in Ground-Based Image Sequences

Abstract

To an experienced waterman, non-stationary hydrodynamic processes have recognizable optical signatures that can aid in navigation of unknown waterways [1], however, the signature physics of light interactions with these turbulent water-surface processes at low grazing angles is not well understood. The objective of this research is to determine the fundamental signature physics of non-stationary hydrodynamic processes inground-based imagery of water flow fields from short (e.g., < 30s) image sequences. Specifically, we will focus our effort on the task of identifying the optical signature of flow field physics for surf-zone rip currents, which can impact the safety of critical equipment and personnel during littoral entry operations. We hypothesize that decreased wave breaking (global signature), increased surface roughness from turbulence and localized high-frequency wave chop/breaking(temporal, local feature), and offshore advecting foam (temporal, global feature) will create a recognizable pattern of light-reflection from low-angles that, when aggregated, identify a rip current (Figure 1). To test our hypothesis, we will employ a novel analysis framework that simultaneously exploits global, local, and temporal features using analytical and data-driven decomposition techniques which are then aggregated through highly efficient deep machine-learning architectures to understand the optical signatures of transient flow fields in ground-based imagery.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 26, 2022
Accession Number
AD1186960

Entities

People

  • Adam Collins

Organizations

  • Engineer Research and Development Center

Tags

DTIC Thesaurus Topics

  • Acquisition
  • Contractors
  • Copyrights
  • Data Acquisition
  • Data Management
  • Data Sets
  • Economic Security
  • Flow Fields
  • Grazing Angles
  • Ground Based
  • Intellectual Property
  • Low Angles
  • Measurement
  • Optical Signatures
  • Physics
  • Standards
  • Surface Roughness

Readers

  • Coastal Oceanography
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML