Join the Meetup to hear talks from experts on cutting-edge topics across Raleigh AI, ML, and Computer Vision.

While most AI/ML applications are discussed in contexts such as enterprise and healthcare, there is a potential to use them for public good. The inherent data-intensive nature and abundance of public-benefit issues in urban planning lends itself well to applied AI/ML. The real word challenges presented also necessitate innovations in architecture and processes. In this talk, I will summarize and share the backgrounds, experiences, and findings from some of my past and present applied AI/ML work in urban planning. The projects include identifying Mobile Home Parks from high-resolution aerial imagery using computer vision, identifying occupations to retrain to clean energy occupations using natural language processing combined with unsupervised clustering, comparing large language model and computer vision approaches to identify small scale solar photovoltaic, and combining graphical neural network with flow matching for energy network optimization and simulation. These projects offer insights on real world complexity, potential to contribute to public good, and responding to challenges and limitations with innovative approaches.

Satellite imagery is a rich source of Earth-observation data across spatial, temporal, and spectral dimensions, but its application in precision agriculture is limited due to its low spatial resolution, which ranges from 3 to 40 m per pixel. Drone imagery offers high resolution but it requires a substantial effort to map a field, and the cost could be substantial if bands beyond RGB bands are required. Our work proposes a Super-Resolution model to obtain high-resolution multispectral images using RGB bands from drone imagery combined with commercially available multispectral satellite imagery such as Sentinel-2 or PlanetScope SuperDove. The model is trained with simulated multispectral data from a hyperspectral camera, achieving a PSNR up to 39 dB. This approach reduces the cost of getting a multispectral mapping of a field, and provides relevant data to calculate vegetation indices, estimate biomass and crop health, and build other models on top of it.

Data—not models—is the bottleneck in agricultural computer vision. This talk shares how Precision Sustainable Agriculture (PSA) is tackling that gap with the Agricultural Image Repository (AgIR): a cloud bank of high-resolution, labeled images spanning weeds (40+ species), cover crops, and cash crops across regions, seasons, and sensors. We’ll show how AgIR blends two complementary streams: (1) semi-field, high-throughput data captured by BenchBot, our open-source, modular gantry that autonomously images plants and feeds a semi-automated annotation pipeline; and (2) true field images that capture real environmental variability. Together, they cut labeling cost, accelerate pretraining, and improve robustness in production.
On top of AgIR, we’ve built a data-centric training stack: hierarchical augmentation groups, batch mixers, a stand-alone visualizer for rapid iteration, and a reproducible PyTorch Lightning pipeline. We’ll cover practical lessons from segmentation (crop/weed/residue/water/soil), handling domain shift between semi-field and field scenes, and designing metadata schemas that actually pay off at model time.

Anomaly detection is transforming manufacturing and surveillance, but what about agriculture? Can AI actually detect plant diseases and pest damage early enough to make a difference?

This talk demonstrates how anomaly detection identifies and localizes crop problems using coffee leaf health as our primary example. We'll start with the foundational theory, then examine how these models detect rust and miner damage in leaf imagery.

The session includes a comprehensive hands-on workflow using the open-source FiftyOne computer vision toolkit, covering dataset curation, patch extraction, model training, and result visualization. You'll gain both theoretical understanding of anomaly detection in computer vision and practical experience applying these techniques to agricultural challenges and other domains.