基于双目视觉的多源信息融合机器人抓取方法研究

Research on Robot Grasping Method for Multi Source Information Fusion Based on Binocular Vision

随着数字化、智能化时代的来临,机器人的应用场景和深度对社会的发展显得尤为重要,这对机器人的安全性、准确性和适应性提出了更高的要求。针对这一情况,基于双目视觉立体匹配的多源信息融合机器人抓取方法研究具有十分重要的意义。多传感器采集目标信息进行不同层面融合,利用双目视觉立体匹配的多源信息融合机器人抓取物体并将其移动到不同的姿势,该策略从视觉数据的角度纠正方向,提高了现代机器人的准确性简化了控制方法,为实现完全由多传感器数据融合控制的工业机器人提供依据,最后研究和分析了多传感器数据融合技术在机器人抓取操作中的应用。通过实验仿真表明,基于双目视觉立体匹配的多源信息融合技术,可以使工业机器人更灵活、更准确地完成抓取操作,效率得到进一步提升。

With the advent of the digital and intelligent era,the application scenarios and depth of robots are particularly important for the development of society,which puts forward higher requirements for the safety,accuracy,and adaptability of robots.In response to this situation,the research on the multi-source information fusion robot grasping method based on binocular vision stereo match-ing is of great significance.Firstly,based on the research of multi-sensor information fusion,stereo matching feature extraction is cal-ibrated using binocular vision cameras to improve the accuracy of robot grasping.Then,a hypothetical approach based on directional arrangement strategy is proposed,which corrects the direction from the perspective of visual data,improves the accuracy of modern robots,simplifies control methods,and provides a basis for achieving industrial robots completely controlled by multi-sensor data fu-sion.Finally,the application of multi-sensor data fusion technology in robot grasping operations is studied and analyzed.Through the experimental simulation,it has been shown that the multi-source information fusion technology based on binocular vision stereo matching can make industrial robots more flexible and accurate in grasping operations,and the accuracy is further improved.