Isw agent links natural language to satellite wave image analysis
ISW-Agent: Integrating LLM-Driven Dialogue with n8n Workflow for Internal Solitary Wave Remote Sensing Image Processing
Artificial Intelligence
Summary
Detecting internal solitary waves in satellite images needs several steps that share details like locations and task settings. The authors developed ISW-Agent to connect natural language instructions with a sequence of image processing tasks. This system uses large language models to understand commands and workflows to perform detection and prediction. It was tested on real satellite data and showed high accuracy and reliable error handling. This helps users request complex wave monitoring steps just by talking to the system.
What this means in practice
- •For remote sensing engineers: Automate internal wave detection and prediction in satellite imagery from natural language task requests using ISW-Agent.
- •For geospatial software developers: Incorporate LLM-driven task interpretation and workflow orchestration into image analysis platforms for coastal and ocean monitoring.
Authors
Hao Zhang, Hengrui Liu, Lina Sun, Junmin Meng
Abstract
Internal solitary wave (ISW) remote sensing requires task parameters and spatial information to be transferred across image acquisition, detection, stripe extraction, and propagation prediction. Here, we present ISW-Agent, which integrates large language model (LLM) interpretation with n8n workflows and independent scientific models through structured task constraints, geospatial transformations, and stage-specific output validation. In tests of 120 Chinese instructions per backend, GPT-5.6 Sol and DeepSeek both achieved a strict task-match rate of 94.2%, with all intents and execution decisions matching the reference. Sentinel-1 cases demonstrated continuous processing from dialogue requests to wavefront prediction for the specified scene. Of 50 timed scenes, 39 completed detection, extraction, and prediction with a median turnaround time of 48.04 s; 11 stopped after empty extraction. Across 3,408 controlled fault-injection runs, retries recovered 234 transient-failure conditions and output validation rejected 234 structural defects, while 61 provenance or completeness errors indicated the need for content-consistency checks. Overall, ISW-Agent thus connects natural-language requests to traceable spatial processing chains for multimodel ISW analysis.