Skip to content

Latest commit

 

History

History
480 lines (292 loc) · 12.8 KB

File metadata and controls

480 lines (292 loc) · 12.8 KB

System Documentation

System Overview

OMEGA is a mobile bathymetric survey system designed to collect underwater depth measurements along a traveled path and reconstruct a mapped representation of submerged terrain from filtered spatial samples. It combines field sensing, real-time validation, structured logging, and post-processing into a single acquisition and mapping pipeline.

Not every field measurement is equally useful for mapping. Depth values must be interpreted in context, evaluated against position and motion quality, and filtered before they are used to reconstruct a surface. OMEGA preserves both the complete raw log of the survey and a reduced mapping dataset containing only higher-quality accepted points.

The platform operates in four stages:

  1. Acquisition — sensor data is collected continuously during movement
  2. Validation — each observation is checked against quality rules in real time
  3. Logging — the system writes both raw and filtered datasets
  4. Reconstruction — mapping products are generated afterward from accepted points

This architecture allows the system to remain lightweight in the field while still supporting structured bathymetric reconstruction and geographic export in post-processing.


Purpose

The purpose of the system is to produce a usable spatial survey record from a moving field platform. It is intended to:

  • collect georeferenced underwater depth measurements
  • associate those measurements with environmental and motion context
  • reject observations that are unsuitable for mapping
  • preserve both raw and filtered data products
  • generate interpretable map outputs from accepted spatial samples

OMEGA is a full measurement-to-map workflow.


Architecture

The system is organized into four main functional layers.

1. Sensor Layer

The sensor layer gathers both mapping variables and measurement context.

Typical inputs include:

  • GPS

    • latitude
    • longitude
    • UTC time
    • speed
    • satellite count
    • HDOP
    • fix age
  • Depth sensor / sonar

    • depth below the sensor
  • DS18B20 temperature sensor

    • water temperature
  • BNO085 IMU

    • pitch
    • roll

These measurements are combined into structured observations that represent both the environment and the state of the survey platform.

2. Validation Layer

The validation layer determines whether a recorded observation is suitable for mapping. It evaluates:

  • GPS quality
  • motion stability
  • depth plausibility
  • spatial spacing

Its role is to prevent unreliable or redundant measurements from dominating the mapping dataset.

3. Logging Layer

The logging layer writes two parallel outputs:

  • a complete raw acquisition log
  • a filtered mapping-quality dataset

This separation preserves the full field record while creating a cleaner input for later reconstruction.

4. Reconstruction Layer

The reconstruction layer uses filtered spatial samples to generate:

  • scatter maps
  • interpolated depth surfaces
  • contour maps
  • KML exports
  • Google Earth overlays

Data Model

Each logged record is treated as a structured measurement event rather than a single sensor reading.

A typical observation may include:

  • UTC timestamp
  • latitude
  • longitude
  • depth
  • water temperature
  • pitch
  • roll
  • speed
  • satellite count
  • HDOP
  • fix age
  • acceptance state

This structure makes it possible to judge depth measurements in context rather than in isolation.

Core Mapping Representation

The mapping dataset is ultimately reduced to samples of the form:

(lat, lng, depth)

These accepted points serve as the basis for reconstruction.

Supporting Metadata

Additional fields are retained because they help determine whether a point should be trusted. These include GPS quality indicators, motion state, and environmental correction inputs.


Data Flow

The system follows a staged processing pipeline:

Sensor inputmeasurement assemblyreal-time validationraw log + filtered map logpost-processing and reconstructionmap outputs

Measurement Assembly

Sensor readings are combined into one observation that associates:

  • time
  • position
  • depth
  • motion state
  • environmental conditions

Validation

Each observation is checked against quality constraints. A measurement may be logged but rejected from the mapping dataset if its conditions are poor.

Dual Logging

The system writes:

  • all observations to the raw file
  • accepted observations to the mapping file

Reconstruction

Accepted mapping points are then used to produce visual and geographic outputs.


Filtering Logic

Real-time filtering is a central design feature of the system. It improves mapping quality by screening measurements as they are collected instead of waiting until the end of the survey.

GPS Constraints

A point is only eligible for mapping when position quality meets defined requirements, such as:

  • valid GPS fix
  • minimum satellite count
  • HDOP below threshold
  • acceptable fix age

This helps ensure that accepted depth samples are tied to credible locations.

Motion Constraints

The platform’s motion state affects both measurement geometry and spatial sampling quality. The system evaluates:

  • speed
  • pitch
  • roll

This reduces the likelihood that unstable measurements become control points in the reconstructed surface.

Depth Constraints

Depth values are tested for:

  • valid reading
  • plausible range
  • consistency relative to recent values
  • resistance to spikes or unstable jumps

These checks suppress noisy readings while preserving actual terrain variation.

Spatial Constraint

A minimum spacing rule is used to prevent accepted mapping points from clustering too densely. This improves point distribution and reduces oversampling of nearly identical locations.


Dual-File Logging Strategy

OMEGA writes two CSV datasets with different roles.

dataXX.csv

This is the complete acquisition log. It preserves the full field record and may include:

  • accepted and rejected measurements
  • unstable sensor states
  • low-quality observations
  • diagnostic information

This file supports review, debugging, and future reprocessing.

mapXX.csv

This is the filtered mapping dataset. It contains only the observations that pass the acceptance logic and are intended for surface reconstruction.

Why the Two-File Design Matters

This structure allows the system to:

  • preserve the raw record
  • keep filtering transparent
  • support alternate reprocessing later
  • simplify mapping workflows
  • prevent noisy data from dominating reconstruction

The raw log preserves everything the system saw. The mapping file preserves what the system judged suitable for map generation.


Depth Processing

Depth is treated as a measured signal that may require refinement before reconstruction.

Smoothing

A moving average filter is used to reduce short-term noise and isolated spikes in the depth series. This improves stability and reduces the visual impact of transient sensor noise.

Temperature-Based Sound Speed Correction

Because acoustic depth sensing depends on the speed of sound in water, and that speed changes with temperature, the system applies a temperature-based correction using measured water temperature. This improves depth consistency across varying conditions.


Reconstruction Method

Reconstruction Input

The reconstruction stage begins with filtered spatial samples:

(lat, lng, depth)

These samples are irregularly spaced and do not directly define a continuous terrain surface.

Why Interpolation Is Needed

A moving survey platform only measures specific locations along its path. The bottom surface must therefore be estimated between recorded points.

Inverse Distance Weighting (IDW)

OMEGA reconstructs the surface using Inverse Distance Weighting (IDW). In this method:

  • nearby points influence the estimate more strongly
  • distant points influence it less
  • each output location is computed from surrounding accepted samples

IDW is a practical choice because it works well with irregular point spacing and is easy to interpret.

Strengths of the Method

  • simple and efficient
  • stable for irregularly spaced field points
  • well suited to lightweight exploratory mapping
  • easy to use across small and medium datasets

Limitations of the Method

  • assumes local smoothness
  • may oversimplify abrupt terrain changes
  • estimates unsampled regions rather than measuring them
  • does not express uncertainty on its own

The reconstructed surface should therefore be interpreted as a modeled estimate constrained by accepted field samples.


Output Products

The system supports multiple output types.

Raw and Filtered CSV Files

These preserve the complete acquisition history and the filtered mapping dataset.

Scatter Maps

Scatter maps display accepted points directly. They are useful for:

  • checking survey coverage
  • reviewing point spacing
  • diagnosing acquisition behavior

Interpolated Surface Maps

These produce a continuous estimated depth surface from accepted spatial samples.

Contour Maps

Contour maps show depth zones and transitions, making bottom structure easier to read.

Geographic Exports

These include:

  • point KML exports
  • colored point KML exports
  • georeferenced overlays for Google Earth

Coordinate System and Units

Coordinate Reference System

  • WGS84

Units

  • depth — centimeters in logging and processed depth values as applicable
  • temperature — degrees Celsius
  • speed — kilometers per hour

Sources of Error

The system is affected by several interacting forms of uncertainty.

Sensor Uncertainty

Depth sensing may be influenced by:

  • acoustic noise
  • beam spread
  • range limitations
  • unstable returns
  • bottom reflectivity
  • sloped terrain

GPS Uncertainty

Position may be affected by:

  • horizontal error
  • poor satellite geometry
  • stale fixes
  • degraded reception

Motion-Induced Error

Motion affects both geometry and sample distribution through:

  • pitch
  • roll
  • speed
  • platform instability

Environmental Variability

Measurements may also be affected by:

  • water temperature changes
  • acoustic interference
  • turbulence
  • changing field conditions

Reconstruction Uncertainty

Interpolation introduces assumptions about the behavior of unsampled areas and depends heavily on the spatial distribution of accepted points.


Design Decisions

Several design choices shape the system.

Real-Time Filtering

Filtering in the field improves the quality of the mapping dataset early and reduces later cleanup.

Separate Raw and Filtered Files

This preserves both traceability and flexibility. The raw file keeps the complete field history, while the filtered file remains focused on mapping.

Post-Processed Reconstruction

Map generation is deferred until after acquisition so the logging system remains lightweight and reconstruction settings can be tuned afterward.

IDW Surface Estimation

IDW provides a practical balance between simplicity, interpretability, and computational cost for this kind of irregular mobile survey data.


Known Limitations

The current system has several important limits:

  • single-beam depth sensing rather than swath coverage
  • strong dependence on GPS quality
  • no salinity-based sound speed correction in the current model
  • interpolation of unsampled regions
  • reduced accuracy under unstable motion conditions
  • better suited to exploratory mapping than formal hydrographic survey work

These limitations define the system’s intended scope.


Future Development

Possible future directions include:

Sensor Improvements

  • pressure-based depth sensing
  • improved sonar hardware
  • multibeam or wider-swath sensing

Validation Improvements

  • adaptive thresholds
  • confidence scoring
  • motion-aware acceptance logic
  • stronger sensor fusion

Reconstruction Improvements

  • alternative interpolation methods
  • uncertainty-aware mapping
  • denser and more configurable grid control

System Integration

  • wireless telemetry
  • remote monitoring
  • live preview mapping
  • distributed sensing and networked devices

Environmental Modeling

  • salinity-aware sound speed correction
  • improved acoustic calibration
  • more advanced environmental compensation

Summary

OMEGA is a structured bathymetric acquisition and reconstruction system built around the distinction between measurement and mapping-quality measurement. Its core strength is the integration of real-time filtering, dual-file logging, and post-processed surface generation into a lightweight field workflow.

Rather than treating every observation as equally trustworthy, the system preserves the full record while selectively constructing a cleaner mapping dataset for reconstruction. That separation between acquisition, validation, logging, and surface generation is the core architectural principle of the platform.