Research Article
A Novel and Efficient Distance Detection Based on Monocular Images for Grasp and Handover
@INPROCEEDINGS{10.1007/978-3-030-92635-9_37, author={Dianwen Liu and Pengfei Yi and Dongsheng Zhou and Qiang Zhang and Xiaopeng Wei and Rui Liu and Jing Dong}, title={A Novel and Efficient Distance Detection Based on Monocular Images for Grasp and Handover}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 17th EAI International Conference, CollaborateCom 2021, Virtual Event, October 16-18, 2021, Proceedings, Part I}, proceedings_a={COLLABORATECOM}, year={2022}, month={1}, keywords={Monocular RGB image Distance detection Grasping Human-robot handover}, doi={10.1007/978-3-030-92635-9_37} }- Dianwen Liu
Pengfei Yi
Dongsheng Zhou
Qiang Zhang
Xiaopeng Wei
Rui Liu
Jing Dong
Year: 2022
A Novel and Efficient Distance Detection Based on Monocular Images for Grasp and Handover
COLLABORATECOM
Springer
DOI: 10.1007/978-3-030-92635-9_37
Abstract
Robot grasping and human-robot handover (HRH) tasks can significantly facilitate people’s production and life. In these tasks, robots need to obtain the real-time 3D position of the object, and the distance from the object to the camera plane is the critical information to get the object position. Currently, depth camera-based distance detection methods always need additional equipment, which results in more complexity and cost. In contrast, RGB camera-based methods often assume that the object’s size is known or the object is at a fixed height. To make distance detection more adaptive and with low cost, a novel and efficient distance detection method based on monocular RGB images is proposed in this paper. With a simple marker, the method can estimate the object’s distance in real-time from the pixel information obtained by a general, lightweight target detector. Experiments on the Baxter robot platform show the effectiveness of the proposed method, where the success rate of the grasping test reaches 87.5%, and the success rate of the HRH test goes 84.7%.
