Abstract: Microwave imaging belongs to nonlinear and non-adaptive problem. Many reconstruction algorithms proposed is not optimal. In order to get better image, a trust region reconstruction(TRR) is proposed. It transforms microwave imaging into restrainable least square algorithm, and then it transforms least square problem into LCP(linear complemental problem). The implementation shows that reconstruction image quality of TRR is better than L-M(Levenberg-Marquardt).

Key words: microwave imaging, reconstruction algorithm, trust region, least square algorithm

摘要： 微波成像问题属于非线性和非适应性问题，目前很多重建算法得到的图像质量并不是最优，为了得到更好的图像，文章提出了信赖域的重建算法(TRR)，该算法将断层成像问题转化为具有约束的最小二乘问题，逐步将非适应性的最小二乘问题转化为线性互补问题求解。模拟结果表明，应用该算法可以得到比L-M算法更好的重建图像。

关键词: 微波成像, 重建算法, 信赖域, 最小二乘算法

CLC Number: 

• TP338