摘要
基于数学形态学的开-闭变换及其组合形式,综合考虑输出信号的均方根误差和信噪比,提出采用最优可变权值的多级广义形态滤波器,避免了自适应广义滤波器算法中最速下降法在迭代接近最优权值时收敛缓慢的缺点;将其应用于MOA在线监测数据处理,以提高复杂电磁环境应力下MOA真实信号提取的精度,保证进一步分析诊断的准确性。仿真分析和实验结果表明,选取最优可变权值的广义滤波器可以抑制孤立的小点、毛刺、小桥和正负脉冲等噪声,运算简单,易于写入硬件(如嵌入式系统等)。
the close-open transforms and their combination mode based on math morphological, considering root-mean-square error and noise ratio, the best variable weight multilevel generalized morphological filter was proposed, and avoiding defect of steepest descent method on the adaptive generalized morphological filter when iterative weight is close to the best one because of slow constriction, It was applied in the on-line monitoring data processing in order to improve the precision of obtaining the real signal of MOA in the stress of complicated electromagnetic environment and ensure the accuracy of further analysis and diagnosis. Results of simulation study and field data processing show that the best generalized morphological filter of variable right value can suppress isolated points, burrs, small bridges, positive and negative impulse etc. And it is simply calculates and easy to write into hardware （embedded system）.
出处
《电瓷避雷器》
CAS
北大核心
2009年第3期34-38,共5页
Insulators and Surge Arresters
