基于Kalman滤波器的红外标志点快速识别与跟踪

在军事上，利用移动机器人跟踪前方的引导员，在野外复杂环境中完成后勤运输与救援任务有重要的意义．针对野外复杂环境这个应用背景，提出了一个基于红外标志点的视觉跟踪系统，该系统有效的解决了传统的视觉跟踪方法受周围环境影响而不能够准确的识别跟踪对象的缺陷；针对传统的标志点识别采用全局搜索的方法计算量大、实时性差的缺陷，提出了一种基于Kalman滤波器预测的红外标志点的识别方法，该方法较传统的标志点识别方法识别速度提高了93％．

Real--time Identification and Tracking of Infrared Markers Based on Kalman Filter

It is important that transportation robot tracking its front guider transports logistics materials to the destination in the wild complex terrain circumstances in the military. According to the application background of wild complex circumstances, a visual tracking system based on infrared markers is proposed in this thesis, which can effectively solve the defects of the traditional visual tracking methods affected by the surrounding environment to not accurately identify the tracking object. In order to solve complex com- puting and poor real--time problems of traditional marker identification method using global search, an in- frared identification method based on Kalman filter was proposed in the thesis. The proposed method im- proved 93% identification speed than traditional markers＇ identification methods using global search.