针对涉电安全工作与治理工作的增强现实展示应用,需要在移动平台上实现快速平面识别和位置姿态估算,并且叠加虚拟场景与实际平面联动的需求,提出一种快速而高效的三维平面物体识别和跟踪算法。同时,为了解决物体识别和跟踪过程中的数据抖动问题,采用了Butterworth滤波器构建了位置和姿态数据的后处理平滑功能模块,以极低的系统延迟为代价,将输出的结果数据进行了平滑和防抖动处理。实验表明,该方法可以快速、低延迟、无抖动地将虚拟电站培训场景与摄像头呈现的现实画面叠加渲染,实现虚实结合的展示效果,从而满足了实际使用中的需要。 This paper combines the specific needs of fast plane recognition and poses estimation in augmented reality applications, to achieve a fast and efficient three-dimensional planar object identification and tracking algorithm. In order to solve the jitter problem during tracking, this paper also implements a post-processing module for position and attitude data smoothing, based on the Butter-worth filter, which handles data at the expense of very low system latency, to satisfy the needs of practical uses in virtual power station training scenes to achieve a combination of virtual and reality display.
展开▼
机译:针对涉电安全工作与治理工作的增强现实展示应用,需要在移动平台上实现快速平面识别和位置姿态估算,并且叠加虚拟场景与实际平面联动的需求,提出一种快速而高效的三维平面物体识别和跟踪算法。同时,为了解决物体识别和跟踪过程中的数据抖动问题,采用了Butterworth滤波器构建了位置和姿态数据的后处理平滑功能模块,以极低的系统延迟为代价,将输出的结果数据进行了平滑和防抖动处理。实验表明,该方法可以快速、低延迟、无抖动地将虚拟电站培训场景与摄像头呈现的现实画面叠加渲染,实现虚实结合的展示效果,从而满足了实际使用中的需要。 This paper combines the specific needs of fast plane recognition and poses estimation in augmented reality applications, to achieve a fast and efficient three-dimensional planar object identification and tracking algorithm. In order to solve the jitter problem during tracking, this paper also implements a post-processing module for position and attitude data smoothing, based on the Butter-worth filter, which handles data at the expense of very low system latency, to satisfy the needs of practical uses in virtual power station training scenes to achieve a combination of virtual and reality display.
展开▼