Odisha, India · Prototype research

Intelligence at the edge of the sea.

TurtleGuard helps authorised conservation teams understand artificial-light and thermal conditions at olive ridley rookeries - locally, explainably, and with people making the final call.

Local-firstOffline edge processing
ExplainableEvery signal made visible
Human-ledNo autonomous intervention
An adult olive ridley sea turtle resting on the sand near the shoreline at Rushikulya
01 / Rushikulya rookery Photograph by the authors
Scroll to discover

The abstract

Measure more than a single number.

TurtleGuard is an edge-based sensor-fusion and risk-scoring system designed to help authorised conservation teams assess two interconnected conditions at olive ridley rookeries: artificial light that may disrupt hatchlings' seaward orientation, and thermal conditions associated with incubation risk.

The prototype combines directional horizon brightness, light-source visibility, spectral measurements, local weather context, and non-invasive sand-temperature proxies. When authorised passive footage is available, reviewed movement trajectories can help evaluate the optical score. Every result exposes its contributing measurements and data-quality status for human review.

What it observes

One shoreline.
Two signal families.

A single lux reading or surface-temperature snapshot misses direction, spectrum, cloud, shade, moisture, depth, and other context. TurtleGuard brings those signals together without hiding the underlying measurements.

01

Optical context

Directional brightness and the visibility, spectrum, and position of light sources near the seaward horizon.

  • Horizon-sector brightness
  • Visible-light spectrum
  • Optional reviewed trajectories
02

Thermal context

Non-invasive, multi-depth sand-temperature proxies interpreted with air, humidity, weather, and local site context.

  • Multi-depth calibration plots
  • Weather and humidity context
  • Quality-gated proxy indicators
03

Edge intelligence

Local acquisition and analysis keep the workflow usable at field sites while making uncertainty explicit.

  • Offline-capable processing
  • Explainable risk indicators
  • Alerts for human review
1AcquireCapture local signals
2Quality gateFlag weak or missing data
3InterpretBuild transparent indicators
4ReviewSupport expert judgement

Field notes

From bench idea to shoreline prototype.

Know more

The thinking behind the prototype.

Explore the project architecture, scientific context, field setup, limitations, and proposed validation pathway in the team's source documents.

PDF19 pages45 MB

TurtleGuard research paper

A detailed account of the problem framing, sensing architecture, explainable scoring approach, field prototype, and study limitations.

PDF6 pages12 MB

Project synopsis

A concise introduction to the conservation challenge, the proposed solution, implementation flow, and intended use in the field.

These documents present prototype research and a proposed validation framework. They should not be read as evidence of a completed longitudinal field trial.

About us

Two researchers.
One guiding mentor.

TurtleGuard is developed at DAV Public School Unit VIII, Bhubaneswar, with a shared interest in field-ready, explainable conservation technology.

01

Student researcher · Co-author

Suryakanta
Lenka

Student researcher at DAV Public School Unit VIII, Bhubaneswar.
03

Project mentor · ATL Coordinator

Tanmay Kumar
Nayak

Mentor at DAV Public School Unit VIII, Bhubaneswar.

Contact us

Start a conversation from the shoreline.

For research, conservation, collaboration, or media enquiries, write to the TurtleGuard team.