KnowFlow

A Unified Knowledge-and-Flow-fused Agent for Comprehensive Remote Sensing Applications

Procedural Knowledge Base: 1,008 curated workflows and action graphs to ground remote-sensing agents.
Dynamic Workflow Adjustment: Adaptive re-planning that aligns tools, data sources, and constraints per task.
Evolutionary Memory: Iterative refinement using feedback signals to stabilize long-horizon executions.
KnowFlow-Bench: Comprehensive evaluation across 162 tasks and 324 groundtruth workflows for reliability.
Earth Observation Ready: Built for satellite imagery, geospatial tools, and real-world decision pipelines.
KnowFlow Overview

Zhengchao Chen1, Haoran Wang1,2, Jing Yao1, Pedram Ghamisi3,4, Jun Zhou5, Peter M. Atkinson4, Bing Zhang1
1. State Key Laboratory of Remote Sensing and Digital Earth, AIRCAS, Beijing, China
2. University of Chinese Academy of Sciences, Beijing, China
3. Helmholtz-Zentrum Dresden-Rossendorf, Freiberg, Germany
4. Faculty of Science and Technology, Lancaster University, UK
5. School of Information and Communication Technology, Griffith University, Australia
Research intention

Research intention: anchor on knowledge-driven workflow templates, pair with dynamic adjustment and evolutionary memory to build verifiable, recoverable, and evolving remote sensing agents (architectures of a: existing agents and b: our proposed Cangling-KnowFlow).

Abstract

The automated and intelligent processing of massive remote sensing (RS) datasets is critical in Earth observation (EO). Existing automated systems are normally task-specific, lacking a unified framework to manage diverse, end-to-end workflows--from data preprocessing to advanced interpretation--across diverse RS applications. To address this gap, this paper introduces Cangling-KnowFlow, a unified intelligent agent framework that integrates a Procedural Knowledge Base (PKB), Dynamic Workflow Adjustment, and an Evolutionary Memory Module. The PKB, comprising 1,008 expert-validated workflow cases across 162 practical RS tasks, guides planning and substantially reduces hallucinations common in general-purpose agents. During runtime failures, the Dynamic Workflow Adjustment autonomously diagnoses and replans recovery strategies, while the Evolutionary Memory Module continuously learns from these events, iteratively enhancing the agent’s knowledge and performance. We evaluated Cangling-KnowFlow on the KnowFlow-Bench, a novel benchmark of 324 workflows inspired by real-world applications, testing its performance across 13 top LLM backbones from open-source to commercial. Across all complex tasks, Cangling-KnowFlow surpassed the Reflexion baseline by at least 4% in Task Success Rate. As the first comprehensive validation along this emerging field, this work demonstrates the great potential of Cangling-KnowFlow as a robust, efficient, and scalable automated solution for complex EO challenges by leveraging expert knowledge (Knowledge) into adaptive and verifiable procedures (Flow).

Framework

Framework: Orchestrator Agent + Procedural Knowledge Base + Dynamic Execution Engine + Evolutionary Memory Module.

Method

Results

BibTeX

@article{CangLingKnowFlow2025,
  title={CangLing-KnowFlow: A Unified Knowledge-and-Flow-fused Agent for Comprehensive Remote Sensing Applications},
  author={Chen, Zhengchao and Wang, Haoran and Yao, Jing and Ghamisi, Pedram and Zhou, Jun and Atkinson, Peter M and Zhang, Bing},
  journal={arXiv preprint arXiv:2512.15231},
  year={2025}
}