GelPoLight: A Novel Visual-Tactile Sensor Based on Photometric Stereo With Point Lighting

被引:0
作者
Lu, Chuang [1 ]
Liang, Ziting [1 ]
Stoyanov, Danail [1 ]
Stilli, Agostino [1 ]
机构
[1] UCL, UCL Hawkes Inst, London WC1E 6BT, England
基金
英国工程与自然科学研究理事会; 英国惠康基金;
关键词
Lighting; Sensors; Elastomers; Light emitting diodes; Tactile sensors; Shape; Cameras; Visualization; Reflectivity; Deformation; Cast shadowing; depth maps; diffuse reflectance; directional lighting; GelSight; light refraction; photometric stereo; point lighting; surface normal vector; visual-tactile sensor (VTS);
D O I
10.1109/JSEN.2025.3535080
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Visual-tactile sensors based on photometric stereo (VTS-PS) have attracted much attention due to their sensitivity to subtle contact forces and their capability to provide real-time, high-resolution depth maps of contact geometry at a low cost. Conventional VTS-PS, such as GelSight, typically assume directional and uniform illumination to simplify the algorithm of photometric stereo. However, this assumption poses stringent requirements on the hardware side and limits the sensors in shape, size, and flexibility for various applications. Additionally, the configuration of directional lighting exacerbates the issue of cast shadowing. In this article, we analyze the benefits of point lighting over conventional directional lighting for VTS-PS and propose GelPoLight, a novel VTS-PS with point lighting, which can help handle or alleviate the aforementioned problems. GelPoLight does not assume ideal diffuse reflectance for the reflective membrane and considers the varying influencing factors of light intensity, light directions, and reflectance factors in the estimation of surface normal vectors. Accordingly, we developed a new calibration method for this approach. We also corrected image distortions caused by light refraction, which is a prevalent issue in visual-tactile sensors (VTSs). The final results demonstrate that GelPoLight can effectively generate the contact depth map, with errors of approximately 5% in the planar direction and 8% in the depth direction.
引用
收藏
页码:9575 / 9584
页数:10
相关论文
共 31 条
[1]   Minsight: A Fingertip-Sized Vision-Based Tactile Sensor for Robotic Manipulation [J].
Andrussow, Iris ;
Sun, Huanbo ;
Kuchenbecker, Katherine J. J. ;
Martius, Georg .
ADVANCED INTELLIGENT SYSTEMS, 2023, 5 (08)
[2]   MEMS-Based Tactile Sensors: Materials, Processes and Applications in Robotics [J].
Bayer, Ilker S. .
MICROMACHINES, 2022, 13 (12)
[3]  
Dong SY, 2017, IEEE INT C INT ROBOT, P137, DOI [10.1109/CIIS.2017.29, 10.1109/IROS.2017.8202149]
[4]  
Donlon E, 2018, IEEE INT C INT ROBOT, P1927, DOI 10.1109/IROS.2018.8593661
[5]   GelTip: A Finger-shaped Optical Tactile Sensor for Robotic Manipulation [J].
Gomes, Daniel Fernandes ;
Lin, Zhonglin ;
Luo, Shan .
2020 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2020, :9903-9909
[6]  
Jia XD, 2013, 2013 WORLD HAPTICS CONFERENCE (WHC), P175, DOI 10.1109/WHC.2013.6548404
[7]   A 3-D Surface Reconstruction with Shadow Processing for Optical Tactile Sensors [J].
Jiang, Hanjun ;
Yan, Yan ;
Zhu, Xiyang ;
Zhang, Chun .
SENSORS, 2018, 18 (09)
[8]   Microgeometry Capture using an Elastomeric Sensor [J].
Johnson, Micah K. ;
Cole, Forrester ;
Raj, Alvin ;
Adelson, Edward H. .
ACM TRANSACTIONS ON GRAPHICS, 2011, 30 (04)
[9]  
Johnson MK, 2009, PROC CVPR IEEE, P1070, DOI 10.1109/CVPRW.2009.5206534
[10]   A Soft and Inflatable Vision-Based Tactile Sensor for Inspection of Constrained and Confined Spaces [J].
Kim, Han Soul ;
Kara, Ozdemir Can ;
Alambeigi, Farshid .
IEEE SENSORS JOURNAL, 2023, 23 (23) :29605-29618