基于SVM决策树的自适应相数ECT图像重建算法

在电容层析成像(ECT)技术的实际应用中，多相流是最普遍存在的流体类型，而被测场域中存在的介质种类常常是未知的。为了解决测量时出现多相流相数未知的情况，利用支持向量机（SVM）决策树算法在分类精度和分类速度上的优点，提出采用基于SVM决策树的电容层析成像图像重建算法对未知相数进行相数预测及图像重建。仿真结果表明，该算法能够实现多相流相数自适应，并提高图像重建精度。

An Adaptive Phase Number Image Reconstruction Algorithm Baesd on DTSVM

Multiphase flow is the most prevalent type of fluid in the practical application of electrical capacitance tomography (ECT) technology,and the types of media in the measured field are often unknown.In order to solve the problem of unknown number of multiphase flow phase during measurement,ECT image reconstruction algorithm based on decision tree SVM that has the advantages of the classification accuracy and speed is proposed for phase prediction and image reconstruction of unknown number of multiphase flow.The simulation results show that the proposed algorithm can realize the phase number adaptation of multiphase flow and improve the image reconstruction accuracy.