Greenbelt, MD – January 2026
GAMA-1 Technologies is proud to announce its participation in the 2026 Annual Meeting of the American Meteorological Society (AMS) through a technical poster presentation highlighting cutting-edge advances in satellite data quality and artificial intelligence.
The work will be featured during the OESS / SATMOC / JCSDA Poster Session III, chaired by Kathryn Shontz of National Oceanic and Atmospheric Administration (NOAA).
Poster Session: OESS / SATMOC / JCSDA Poster Session III
Date: Wednesday, January 28, 2026
Time: 3:00 PM – 4:30 PM (CT)
Location: Hall B3, George R. Brown Convention Center
Poster #634: Advancing GOES ABI Anomaly Detection with Deep Learning
GAMA-1’s contribution to this session—Poster #634, “Progress in GOES ABI Observation Anomaly Detection and Classification With Deep Learning”—showcases collaborative research advancing real-time quality assurance for satellite radiance data critical to weather forecasting and environmental monitoring.
Developed within NOAA/NESDIS/OCS, the presented multi-label deep learning classification model is designed to automatically detect and classify image-level anomalies in GOES Advanced Baseline Imager (ABI) radiance observations, with future extensibility to VIIRS observations aboard NPP, JPSS-1, and JPSS-2. By automating anomaly detection, the model strengthens real-time operational Quality Assurance (QA) monitoring and improves pixel-level QA flags—directly benefiting downstream Level-2 product processing.
From Hard-to-Detect Anomalies to Operational Readiness
In its initial development phase, the model was trained to identify lunar intrusion Type I anomalies, such as latched detector responses and bad detector stripes, that are notoriously difficult to detect through manual or traditional methods. Validation results demonstrate that deep learning significantly enhances detection capability for these subtle issues.
As the training dataset expands, the model is being extended to classify additional complex anomalies, including caterpillar tracks, shark fin anomalies, bright object avoidance effects, solar farm and ringing anomalies, and lunar intrusion Type II. The poster details the methodology behind this work, including training data selection and validation challenges, image conventions, and the use of a ResNet-based model architecture.
Collaboration Driving Satellite Data Integrity
The poster will be presented by GAMA-1’s Haibing Sun supporting National Environmental Satellite, Data, and Information Service (NESDIS), with contributions from a strong, cross-organizational team including Rebekah Esmaili, Elizabeth Kline, Walter Wolf, and Letitia Soulliard.
Strengthening the Public Mission Through AI
High-quality satellite observations are foundational to NOAA’s mission to protect life and property. By applying advanced deep learning techniques to operational quality control, this work enhances the reliability of satellite-derived products that underpin weather forecasting, climate analysis, and environmental intelligence relied upon by communities nationwide.
GAMA-1 Technologies is honored to support this mission-focused innovation and to collaborate alongside NOAA scientists and engineers advancing the future of satellite operations. We invite AMS attendees to visit Hall B3 during Poster Session III to learn more and engage with the team driving these AI-enabled advancements.
About GAMA-1 Technologies – Mission-Ready for Government Transformation
GAMA-1 Technologies is mission-driven, empowering federal agencies to deliver smarter services through secure cloud modernization, AI-driven automation, and adaptive cybersecurity and compliance. We specialize in enabling decision support mission critical data to benefit humanity, leveraging innovation to streamline government and accelerate outcomes that matter.
With deep roots in environmental observation infrastructure and data modernization, we help agencies like NOAA, NASA, and others transform legacy systems, unify mission data, and embrace digital-first operations. Our solutions are engineered for public good, driving efficiency, reducing costs, and amplifying mission success.
Contact:
Stan Coachman, Contracts Manager – 301-982-4262, [email protected]
- Discovery
- Analysis
- Vision Definition and Development
- Documentation
