An open-source initiative applying AI, sensing, and edge computing to wildlife research and conservation.
Wildlife researchers collect images, video, audio, tracking signals, and environmental measurements in difficult field conditions. Much of this work still depends on manual observation, fragmented tools, and project-specific data formats.
Open Datalogger For Interspecies aims to connect real field research questions with reusable open-source AI, sensing, and edge computing patterns. The first stage of this repository is intentionally documentation-driven: define the research themes, contribution model, data schemas, and project templates before adding heavier software modules.
A reCamera Pro detects a stray cat as a safe proxy target, triggers a GPIO pulse, and sends a Cat detected alert through a Meshtastic network to a Wio Tracker L1 Pro receiver.
This validates a field workflow that can later be adapted for wildlife or risk-event alerts, such as tiger, wolf, livestock-conflict, or potential poacher detection. It is a pilot workflow demo, not a claim that a production wildlife detector is complete.
| Start Here | Purpose |
|---|---|
| Read the full demo guide | One-page project story, hardware list, wiring, Meshtastic setup, reCamera setup, runtime evidence, and limitations |
| Use the reference implementation | Camera-side trigger scripts, runtime logger, snapshot tools, and Meshtastic node setup notes |
| View hardware and setup images | Photos of the deployment, wiring, power setup, receiver alert, and configuration screens |
| Theme | Core Question | AI Capabilities |
|---|---|---|
| Individual Identification & Lifetime Tracking | Which individual is this, and what happens across its life? | Re-ID, embeddings, tracking, life-history records |
| Acoustic Monitoring & Stress Signals | What sounds are animals producing, and do they indicate stress or abnormal behavior? | Bioacoustics, event detection, classification |
| Population Monitoring | How many animals are present, and how is the population changing? | Detection, counting, occupancy modeling |
| Social Behavior Networks | Who interacts with whom, and how is the group organized? | Multi-object tracking, association analysis, graph analytics |
| Habitat Health & Carrying Capacity | Can the habitat support the population over time? | Sensor fusion, time-series analysis, geospatial modeling |
| Cognitive Aging & Behavioral Change | Are behavior patterns changing with age, disease, or stress? | Longitudinal analytics, anomaly detection, multimodal AI |
Reproduction can be studied across these themes through individual life histories, population recruitment, social relationships, and habitat conditions.
Field devices
cameras, acoustic nodes, trackers, environmental sensors
|
v
Field data layer
observation records, device metadata, deployment sites
|
v
AI capability layer
vision, bioacoustics, time-series, edge AI, LLM-assisted analysis
|
v
Research projects
species-specific studies, conservation programs, open datasets
See System Architecture for the intended structure.
| Path | Purpose |
|---|---|
docs/ |
Vision, architecture, field deployment, data governance, research themes, hardware notes, and AI capability notes |
projects/edge-ai-camera-meshtastic-alert/ |
First pilot demo: edge AI camera detection to Meshtastic alert |
reference-implementations/recamera-meshtastic-trigger/ |
Reusable scripts and setup notes for the first pilot demo |
projects/ |
Project proposal guidance, reusable project template, and pilot examples |
reference-implementations/ |
Lightweight reusable code modules validated by pilot projects |
schemas/ |
JSON Schemas for common wildlife research records |
.github/ |
Issue templates, pull request template, and documentation checks |
- Bring a real research question or field deployment problem.
- Propose a project using the project template.
- Improve research theme documentation with domain knowledge.
- Contribute datasets only when permissions, sensitivity, and species risk have been reviewed.
- Add reference implementations once a pattern has been used in a real project.
Start with CONTRIBUTING.md and RESPONSIBLE_RESEARCH.md.
This repository is in the foundation stage. We are defining shared language, project structure, governance, and minimal data schemas before promoting model code or production pipelines.
If this initiative helps your research, please cite the repository using CITATION.cff.