Seg2Grasp: A Robust Modular Suction Grasping in Bin Picking

2026-07-20

AI Summary

This work addresses the limited generalization of end-to-end grasping methods to unknown or complex objects in unstructured environments by proposing a modular, decoupled framework that sequentially performs segmentation, grasping, and open-vocabulary classification. The approach leverages a Transformer to generate category-agnostic object masks, combines surface normal estimation to guide suction-based grasp point selection, and employs a fine-tuned Mask-CLIP model for accurate open-vocabulary recognition. Real-world robotic experiments demonstrate that the proposed method significantly improves grasping success rates and generalization capabilities in cluttered scenes, outperforming existing approaches and showing strong potential for practical applications such as industrial automated sorting.

Abstract

Current bin picking methods that rely heavily on end-to-end learning often falter when confronted with unfamiliar or complex objects in unstructured environments. To overcome these limitations, we introduce Seg2Grasp, a modular pipeline designed for robust suction grasping in dynamic and cluttered bin scenarios. Seg2Grasp is built on a three-step process: Segmentation, Grasping, and Classification. The Segmentation module employs a Transformer-based model to generate class-agnostic object masks from RGB-D images, ensuring accurate detection across various conditions. The Grasping module uses surface normals and mask proposals to determine the optimal suction points, enhancing grasp success. Finally, the Classification module leverages fine-tuned open-vocabulary Mask-CLIP for precise object identification, enabling versatile handling of diverse objects. Real-world robotic experiments demonstrate that Seg2Grasp outperforms existing methods in success rates and adaptability, establishing it as a powerful tool for automated bin picking in industrial settings.

Problem

• bin picking
• suction grasping
• unstructured environments
• complex objects
• robust grasping

Innovation

• modular grasping
• Transformer-based segmentation
• suction grasping
• open-vocabulary classification
• bin picking