摘要
针对无人机受到气流振动及高温蠕变的影响,螺栓结构产生滑动、分离或者脱落问题,传统图像监测方法在对螺栓环固定区域进行图像采集时,螺栓环"环带"与杆之间由于色彩信息比较类似,会产生一定的像素干扰,造成监测过程的抗干扰能力较差等问题。提出基于机器视觉图像处理与卡尔曼滤波相结合的无人机螺栓松动监测方法,引入机器视觉识别原理,对无人机上螺栓环固定区域特征进行一定频率的图像采集,对无人机上的螺栓环区域异常位移特征采用图像帧差的方式进行判断,对螺栓环图像包含的色彩信息中过滤干扰,并进行螺栓环"环带"位移进行识别,为了保证识别准确性,引入了卡尔曼滤波算法对螺栓冒和杆的相对位置进行预测,实时监控无人机螺栓松动。仿真结果说明,提出的方法能够精确识别无人机上螺栓环面脱离的情况,具有较高的监测精度。
A monitoring method of loose bolts in unmanned aerial vehicle (uav) is put forward based on machine vision image processing combined with kalman filter. The regional image characteristics of bolt ring in the uav are ac- quisited at a certain frequency, the regional anomaly displacement characteristics of the bolt on the ring are judged according to the difference of image frames, the interference of colour information in the bolt ring image is filtered, and the "ring - band" displacement of bolt rings is recognized. In order to guarantee the recognition accuracy, kalman filtering algorithm is introduced to predict relative position of the bolt and the rod, and the loose boh is real - time monitored. Simulation result shows that the proposed method can accurately recognize the loose bolt, and has high accuracy of monitoring.
出处
《计算机仿真》
CSCD
北大核心
2015年第10期94-96,359,共4页
Computer Simulation
基金
内蒙古自治区高等学校科学研究项目(NJZY286)