旋转电弧传感器信号消噪的研究  

The research of rotating arc sensor signal de-noising

作　　者：熊勇[1] 张华[1] 贾剑平[1] 李志刚[1] 

机构地区：[1]南昌大学江西省机器人与焊接自动化重点实验室,江西南昌330031

出　　处：《传感器与微系统》2010年第5期41-43,共3页Transducer and Microsystem Technologies

基　　金：国家"863"计划资助项目(2007AA04Z242);江西省科技攻关计划资助项目(2007BG09100)

摘　　要：旋转电弧传感器是焊接自动跟踪中的重要装置,为了获得水下焊接坡口的特征信息,必须对采集的焊接电流信号进行消噪,以去除干扰信息。利用平面中值滤波器对8个方向噪声点进行处理,平滑滤波器对信号中的奇异点和毛刺处理,以及小波变换能在小波域实现信噪分离,以便去除整个频域范围内的噪声,设计了一种中值平滑小波改进型滤波器。Matlab仿真结果表明:此方法能有效消除焊接过程中的各种干扰,且经过处理后的信号表现非常明显的正弦变化规律。通过焊接实验,获得良好的跟踪效果。Rotating arc sensor is an important sensor for automatic weld seam tracking. In order to obtain the characteristics information of groove weld of underwater welding, the welding current signal must be denoised so as to remove interference information. Plane median filter is used to deal with the noisy points of eight directions, and smooth filter is used to process the singular point and the burr of current signal, as well as wavelet transform can achieve separation of signal and noise so as to clearing up the noisy signal in the entire range of frequency domain. A new filtering method is designed, which is integrated by plane median filter, mean filter, smoothing filter, morphological mean filter and improved wavelet filter. The result of simulation using Matlab indicates that this method is very effective to eliminate various kinds of interference signal which is produced in the welding process, and the signal manifests the law of sinusoidal variation after signal processing. Through the welding experiments, the effect of weld seam tracking is great.

关 键 词：旋转电弧传感器 平面中值滤波 平滑滤波 形态均值滤波 改进型小波滤波 

分 类 号：TP409[自动化与计算机技术] TN911[电子电信—通信与信息系统]