BP(Back-propagation)神经网络以其非线性映射的特点在遥感影像的分类中得到了应用,采用LM(Leven-berg-Marquardt)算法的BP网络收敛速度明显加快。本文以TM影像为例,讨论了基于LM算法改进的BP神经网络在遥感影像分类中的应用,分类结果同最大似然法进行对比,并在此基础上,将主成分变换方法与LM算法结合,进行有益尝试。试验结果证明LM算法改进的BP神经网络分类优于常用的最大似然法分类。 BP (Back-propagation) neural network has been applied to the classification of remote sensing images with its nonlinear mapping characteristics. The convergence speed of BP network using LM (Leven-berg-Marquardt) algorithm is obviously faster. This paper takes TM image as an example, discusses the application of improved BP neural network based on LM algorithm in remote sensing image classification, the classification result is compared with the maximum likelihood method, and on this basis, the principal component transformation method is combined with the LM algorithm , Make useful attempts. The experimental results show that the BP neural network improved by LM algorithm is superior to the commonly used maximum likelihood method.