Rethinking Air-Ground Collaboration: A Progressive Cross-Task Benchmark and Socialized Learning Framework

2026-06-17

AI Summary

This work addresses the limitations of existing aerial-ground collaborative perception methods, which are often confined to single-task cross-view fusion and struggle with negative transfer caused by geometric, scale, and occlusion discrepancies, while overlooking functional dependencies among localization, object association, and fine-grained parsing. To overcome these challenges, the authors propose a task-conditioned progressive collaboration paradigm, introduce the AGPC benchmark comprising 745,000 spatiotemporally aligned frames, and develop a Socialized Co-Perception framework that enables ordered cross-task cooperation through a coarse-to-fine pipeline. A key component, the Dual-Layer Router, decouples multi-scale expert selection from task modulation, facilitating selective cross-view interaction. Experiments demonstrate an average performance gain of 7.86% on downstream tasks and a collaborative evolution improvement of 3.73%, significantly outperforming conventional unified fusion strategies.

Abstract

Air-ground collaborative perception is crucial for robust visual understanding in real-world dynamic environments. However, existing studies typically formulate collaboration as single-task cross-view fusion, overlooking the functional dependencies among localization, target association, and fine-grained parsing. In addition, the heterogeneous nature of aerial and ground views introduces substantial geometric, scale, and occlusion discrepancies, making uniform feature sharing vulnerable to negative transfer. To tackle these issues, we model air-ground perception as a progressive cross-task collaboration task and construct the Air-Ground Progressive Collaboration (AGPC) benchmark, a spatio-temporally aligned benchmark comprising more than 745K raw video frames. Built upon this benchmark, we propose Socialized Co-Perception (SCP), a coarse-to-fine framework that organizes collaboration progressively from aerial global localization to ground target association and identity-aware parsing. Its core module, the Dual-Layer Router (DLR), decouples input-side multi-scale expert selection from output-side task-conditioned modulation, enabling selective cross-view and cross-task interaction while suppressing harmful interference. Extensive experiments demonstrate the effectiveness of SCP. It achieves a 3.73\% coevolutionary gain and a 7.86\% improvement in average downstream performance. These results show that task-conditioned collaboration is more effective than uniform fusion for heterogeneous air-ground perception. The code is available at https://github.com/g1136639260-spec/AGSCP.

Problem

• air-ground collaboration
• cross-task dependency
• heterogeneous perception
• negative transfer
• collaborative perception

Innovation

• progressive cross-task collaboration
• air-ground co-perception
• Dual-Layer Router
• task-conditioned modulation
• heterogeneous perception