本文選題：電阻抗斷層成像(EIT)　切入點：FCM聚類算法　出處：《北京航空航天大學學報》2017年11期 　論文類型：期刊論文

【摘要】：針對電阻抗斷層成像(EIT)技術(shù)中FCM聚類算法的靈敏度系數(shù)信息缺失以及測量電壓的利用率低兩方面問題,提出一種新的成像算法。在該算法中引入靈敏度系數(shù)矩陣信息修正各個剖分單元的電壓。同時提出了將測量電壓數(shù)據(jù)按照其權(quán)系數(shù)進行處理的方法,該方法可應(yīng)用于所有EIT經(jīng)典反演算法之中。理論和數(shù)值仿真結(jié)果均表明,與已有的FCM聚類算法相比,優(yōu)化后算法對兩相流型的定位準確度更高,得到的重建圖像的空間分辨率與之前相比相對誤差降低了5%~15%,相關(guān)系數(shù)提高了5%~20%。
[Abstract]:Aiming at the lack of sensitivity coefficient information of FCM clustering algorithm and the low utilization rate of measuring voltage in the electrical impedance tomography (EIT) technique, A new imaging algorithm is proposed, in which sensitivity coefficient matrix information is introduced to modify the voltage of each subdivision unit, and a method to process the measured voltage data according to its weight coefficient is proposed. This method can be applied to all classical EIT inversion algorithms. The theoretical and numerical simulation results show that compared with the existing FCM clustering algorithm, the optimized algorithm is more accurate in locating two-phase flow patterns. The spatial resolution of the reconstructed image is 5 / 15 lower than that of the previous one, and the correlation coefficient is increased by 5 / 20.
【作者單位】： 天津大學電氣與自動化工程學院;
【基金】：國家自然科學基金(61174014)~~
【分類號】：TP391.41;V231
，

本文編號：1632255