STOPNet: Multiview-based 6-DoF Suction Detection for Transparent Objects on Production Lines

TL;DR

STOPNet is a novel multiview stereo framework that accurately detects 6-DoF suction poses for transparent objects on production lines using only RGB input, overcoming depth sensing limitations.

Contribution

The paper introduces STOPNet, a multiview stereo-based method for 6-DoF suction detection that works with transparent objects and generalizes well to new environments and objects.

Findings

Outperforms existing methods in accuracy and speed.

Successfully detects transparent objects in real industrial settings.

Demonstrates strong generalization in diverse scenarios.

Abstract

In this work, we present STOPNet, a framework for 6-DoF object suction detection on production lines, with a focus on but not limited to transparent objects, which is an important and challenging problem in robotic systems and modern industry. Current methods requiring depth input fail on transparent objects due to depth cameras' deficiency in sensing their geometry, while we proposed a novel framework to reconstruct the scene on the production line depending only on RGB input, based on multiview stereo. Compared to existing works, our method not only reconstructs the whole 3D scene in order to obtain high-quality 6-DoF suction poses in real time but also generalizes to novel environments, novel arrangements and novel objects, including challenging transparent objects, both in simulation and the real world. Extensive experiments in simulation and the real world show that our method significantly surpasses the baselines and has better generalizability, which caters to practical industrial needs.