在隨機擾動環境中采用漏磁場法在線測量油管壁厚時,每個磁性傳感器只能得到環境的
部分信息。為了對各傳感器的局部測量值進行穩健融合估計,提出了油管壁厚測量數據的一致性加權融合估計算法。利用改進的數據探測技術中的分布圖方法對各個傳感器的一致性進行檢驗,并采用簡單加權平均法求得一致性傳感器測量數據的穩健融合估計值。融合實例表明,該算法能有效地提高系統測量的穩健性,且具有精度高、運算簡單等特點。該算法適用于多傳感器測量數據的實時高精度融合估計。
關鍵詞:多傳感器;數據融合;信號處理;油管;無損檢測
　　Abstract :In an on - line wall thickness measurement system of oil well tubing using magnetic flux leakage technique , each magnetic sensor can only offer partial information with stochastic noise about it’s environment . The consensus weighted fusion algorithm for the robust fusion of multisensor data is developed. The improved distribution display method
in exploratory data analysis is used to test various sensor’s consensus. Meanwhile ,the simple weighted mean method is in2 troduced to fuse consensus sensor data ,and the robust fused data of multisensor data is obtained as a result . The fusion re2 sults show that the robustness of the measurement system is evidently enhanced ,and the algorithm has specific advantages over some conventional algorithms such as high - accuracy ,and high - efficiency. It is predicted that the algorithm is avail2 able for real - time ,accurate fusion of multisensor data.
Key Words :Multisensor ;Data Fusion ;Signal Processing ;Oil Well Tubing ;Nondestructive Testing