基于仿生机器人的狭小空间裂纹检测技术OA
Narrow Space Crack Detection Technology Based on Bionic Robot
飞机裂纹的检查贯穿飞机的整个生命周期,若不及时发现可能会带来严重的后果,尤其是飞机盒段等狭小空间部位,人不可达,微小裂纹难以快速检查,给飞机服役、地面强度及疲劳试验等带来重大安全隐患.本文研究了可自由进入狭小空间的仿生机器人平台,涉及仿生机器人机构、运动行为及步态规划、环境感知和避障等.开展了大量图像文本的数字增强、图像预处理、基于深度学习的图像裂纹算法等研究,在具有狭小空间特征的飞机盒段样件上开展了试验验证,通过盒段内部预设裂纹,有效地集成图像采集模块和仿生机器人平台,并结合图像采集环境调试,机器人路径规划、裂纹算法优化研究等,成功实现了对狭小空间内部的裂纹长度和位置的识别,为后续开展飞机裂纹全覆盖快速检查和维修策略提供技术支撑.
The inspection of cracks runs through the entire life cycle of aircraft,and it may bring serious consequences with crack not found in time,especially in narrow space parts such as aircraft box segment,inaccessible to person,small cracks are difficult to be inspected quickly,which brings major safety risks to the aircraft service,ground strength and fatigue tests.A bionic robot platform that can freely enter a narrow space is studied,involving bionic robot mechanism,motion behavior and gait planning,environment perception and obstacle avoidance.Digital enhancement on a large number of images,image processing and image crack algorithm based on deep machine learning have been carried out.The sample of aircraft box segment with narrow space characteristics has been experimented.By means of presetting cracks inside the box segment,effectively integrating image acquisition module and the bionic robot platform,combining with the image acquisition environment regulation,robot path planning and crack algorithm optimization,the crack length and location in a narrow space is successfully detected,which provides technical support for the follow-up rapid inspection of full crack coverage of aircraft and maintenance strategy.
王文娟;张梦杰;刘元博;孙介;薛景锋;段晋军;刘中
中国航空研究院,北京 100012南京航空航天大学,江苏 南京 210016北京航空航天大学,北京 100191
狭小空间仿生机器人图像处理深度机器学习裂纹识别
narrow spacebionic robotimage processingdeep machine learningcrack detection
《航空科学技术》 2024 (007)
104-110 / 7
评论