GraspView: Active Perception Scoring and Best-View Optimization for Robotic Grasping in Cluttered Environments

2025-11-06

AI Summary

To address robotic grasping failures in cluttered scenes caused by occlusion, poor perceptual quality, and inconsistent 3D reconstruction, this paper proposes a monocular RGB-only active perception grasping framework. Methodologically, it integrates single-view geometric reconstruction with multi-view fusion for global scene modeling; introduces a rendering-based grasp scoring mechanism to guide active viewpoint selection; and designs an online metric alignment module to ensure scale-consistent grasp planning. Its key contribution lies in eliminating reliance on depth sensors—instead, it establishes a robust closed-loop active perception pipeline combining VGGT-based pose estimation, GraspNet-based grasp proposal generation, and rendering-driven viewpoint optimization. Experiments demonstrate that the method significantly outperforms both RGB-D and single-view RGB baselines under challenging conditions, including severe occlusion, close-range manipulation, and grasping of transparent or specular objects.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Video Understanding & Activity Analysis
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Robotic grasping is a fundamental capability for autonomous manipulation, yet remains highly challenging in cluttered environments where occlusion, poor perception quality, and inconsistent 3D reconstructions often lead to unstable or failed grasps. Conventional pipelines have widely relied on RGB-D cameras to provide geometric information, which fail on transparent or glossy objects and degrade at close range. We present GraspView, an RGB-only robotic grasping pipeline that achieves accurate manipulation in cluttered environments without depth sensors. Our framework integrates three key components: (i) global perception scene reconstruction, which provides locally consistent, up-to-scale geometry from a single RGB view and fuses multi-view projections into a coherent global 3D scene; (ii) a render-and-score active perception strategy, which dynamically selects next-best-views to reveal occluded regions; and (iii) an online metric alignment module that calibrates VGGT predictions against robot kinematics to ensure physical scale consistency. Building on these tailor-designed modules, GraspView performs best-view global grasping, fusing multi-view reconstructions and leveraging GraspNet for robust execution. Experiments on diverse tabletop objects demonstrate that GraspView significantly outperforms both RGB-D and single-view RGB baselines, especially under heavy occlusion, near-field sensing, and with transparent objects. These results highlight GraspView as a practical and versatile alternative to RGB-D pipelines, enabling reliable grasping in unstructured real-world environments.

Problem

• Addresses robotic grasping challenges in cluttered environments with occlusions
• Overcomes limitations of RGB-D sensors on transparent and glossy objects
• Eliminates dependency on depth sensors through RGB-only perception pipeline

Innovation

• RGB-only grasping pipeline without depth sensors
• Active perception strategy for next-best-view selection
• Online metric alignment for physical scale consistency