Sparse-Dense Motion Modelling and Tracking for Manipulation without Prior Object Models

TL;DR

This paper introduces a novel method for modeling and tracking unseen objects in robotic manipulation, combining dense and sparse tracking to enable robust, online object detection and grasping without prior models.

Contribution

It presents a new approach that segments, models, and tracks objects using motion cues, improving robustness and enabling manipulation of unknown objects in real-time.

Findings

Enables online detection and grasping of new objects without prior models

Outperforms state-of-the-art methods in robustness

Uses simple geometric representations for effective manipulation

Abstract

This work presents an approach for modelling and tracking previously unseen objects for robotic grasping tasks. Using the motion of objects in a scene, our approach segments rigid entities from the scene and continuously tracks them to create a dense and sparse model of the object and the environment. While the dense tracking enables interaction with these models, the sparse tracking makes this robust against fast movements and allows to redetect already modelled objects. The evaluation on a dual-arm grasping task demonstrates that our approach 1) enables a robot to detect new objects online without a prior model and to grasp these objects using only a simple parameterisable geometric representation, and 2) is much more robust compared to the state of the art methods.