SDZWA Conservation Technology Lab

Western US

Open-source tooling for camera-trap imagery: ingestion, annotation, automated detection and species classification, and reproducible analysis pipelines.

Classes

Work focuses on robust ingestion from diverse camera systems, annotation workflows that speed labeling, and models that detect animals and estimate abundance. We also prioritize reproducible notebooks and deployment-ready detection pipelines.

Savanna

Projects focused on physical sensors and deployments: low-power environmental monitoring, data ingest and telemetry, calibration, and integration with analysis pipelines.

We produce firmware examples, ingest connectors, and best-practice guides for deployments in remote locations. Emphasis is on low-power design, robust telemetry, and reproducible downstream analysis.

Amazon

Tools and workflows for acoustic monitoring: sound preprocessing, automated detection/classification, and scalable pipelines for long-term acoustic datasets.

We develop reproducible pipelines for long-duration audio, tools for training detectors and classifiers, and integrations to index and search acoustic archives at scale. Typical outputs include event detectors, spectrogram-based embeddings, and lightweight ingestion tools for field recordings.

Andes

Research and tools for individual re-identification from images or acoustic signatures, including embedding models, matching workflows, and evaluation scripts.

Projects include model training recipes for embeddings, example matching services, and evaluation suites for comparing approaches. We provide guidance for dataset preparation, cross-validation strategies, and deployment considerations for field use.

Vietnam

Coming soon...