Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a novel anomaly-detection framework for identifying artifacts in satellite-based rainfall data, addressing the urgent need for data validation in the burgeoning private satellite sector. By leveraging pre-trained computer vision models and integrating limited human-labeled data, the system effectively distinguishes between regular and artifact-containing orbits in real-time. The results indicate that this approach achieves performance on par with existing state-of-the-art methods while providing explainability and the ability to iteratively refine its classifications.
Anomaly detection in satellite rainfall data can now be performed in near-real-time, ensuring higher data integrity before public dissemination.
Historically, retrieving rainfall data from satellite imagery has been the domain of space agencies. However, in recent years, the development of cheaper, more compact satellites (SmallSats) capable of detecting rainfall proxies has led to a significant increase in private-sector initiatives for satellite launch and surface precipitation products. This rapid growth has yet to be matched by data validation efforts. Consequently, the need for a robust tool to detect anomalies in near-real-time data before it is disseminated to the public has become critical. In this paper, we present the development of an anomaly-detection system to identify artifacts in global satellite-based rainfall products. The developed framework leverages pre-trained computer vision models and incorporates scarce human-labeled data to detect specific anomalies. Our proposed anomaly detection strategy is tested on data from the Special Sensor Microwave Imager (SSMI) and the Special Sensor Microwave Imager/Sounder (SSMIS). Results demonstrate the efficacy of our approach at separating regular orbits from artifact-containing orbits for each satellite, with performance comparable to state-of-the-art in-place methods. Additionally, the framework offers explainability and the capacity for iterative refinement following false-positive or false-negative classifications.