光学层析重建算法改进的研究

光学层析技术在温度、密度等流场的非接触测试中有着广泛的应用前景，然而光学层析的重建是非完全数据的投影重建，通常采用的代数迭代法不能很好地解决重建精度这一难题。为此提出了一种新的光学层析技术的代数迭代重建算法，在算法中引入了包含先验知识的属性矩阵，并摒弃了通常所采用的对超松弛系数人为的确定取法，采用了变超松弛系数。实验计算结果表明，引入属性矩阵和变超松弛系数的代数迭代法能够较好的重建非完全数据投影的待测场，极大地提高重建精度，较好地与实测结果吻合。

Improvement of The Reconstruction Algorithm of Optical Computerized Tomography

Optical computerized tomography (OCT) has found wide applications in the noncontact detection of the flow field such as temperature filed, density field etc. However, the reconstruction of OCT is a projection reconstruction of incomplete data, and the algebra reconstruction technique usually adopted cannot solve the problem of reconstruction precision very well. A new kind of algebra reconstruction technique is presented. The property matrix including the prior knowledge is introduced. The fixed ultra -relaxation coefficient usually determined artificially is abandoned and an unfixed ultra -relaxation coefficient is adopted in the algorithm. The reconstruction results indicate that the algorithm with property matrix and unfixed ultra -relaxation coefficient can reconstruct the flow field of incomplete data satisfactorily, improve the reconstruction precision greatly, and match the measured results.