為了對(duì)隨機(jī)擾動(dòng)環(huán)境中得到的管道壁厚測(cè)量數(shù)據(jù)進(jìn)行穩(wěn)健融合估計(jì), 提出了一種多傳感器測(cè)量數(shù)據(jù)的切尾加權(quán)融合算法。利用數(shù)據(jù)探測(cè)技術(shù)中的切尾均值概念構(gòu)造最優(yōu)加權(quán)因子, 從而求得多傳感器測(cè)量數(shù)據(jù)的穩(wěn)健融合估計(jì)值。融合實(shí)例表明, 該算法能有效地提高系統(tǒng)測(cè)量的穩(wěn)健性, 且具有精度高、運(yùn)算簡單的特點(diǎn)。
關(guān)鍵詞　多傳感器　數(shù)據(jù)融合　信號(hào)處理　管道　無損檢測(cè)
Abstract　A t rimmed mean- based weigh ted fusion algo rithm is developed fo r the robust fusion of the wall lo ss measurement data w ith a stochast ic no ise from mult isenso r1 The t rimmed mean method in exp lo rato ry data anal2 ysis is used to build the op t imal weigh ted facto rs, and the robust fused data of mult isenso r data is obtained as a result1 The fusion result s show that the robustness of themeasurement system is evident ly enhanced, and the al2 go rithm has specific advantages over some convent ional algo rithm s such as h igh- accuracy, and h igh- efficiency1
Key words　M ult isenso r　Data fusion　Signal p rocessing　P ipeline　Nondest ruct ive test ing