机器学习方法下的SAR图像分类系统建构及其识别分析

Construction and Recognition Analysis of SAR Image Classification System Based on Machine Learning Method

本文在对图像分类与识别过程中运用机器学习方法,进行不同机器算法下的SAR-ATR系统构建,运用Adaboost算法、PCA算法、NMF算法等进行研究,对SAR图像进行分类对比实验,研究可见Adaboost算法检测识别效果有限,可能陷入局部最优的困境之中。NMF算法、PCA算法在对SAR图像分析处理过程中不需要进行图像的预处理,两者联合在图图像特征的提取层面具有良好的代表性。

In this paper,the machine learning method is used in the process of image classification and recognition,and the SAR-ATR system under different machine algorithms is constructed.The Adaboost algorithm,PCA algorithm,NMF algorithm,etc.are used for research.The SAR images are classified and compared.It can be seen that the Adaboost algorithm has limited detection and recognition effect,and may fall into the dilemma of local optimization.The NMF algorithm and the PCA algorithm do not require image preprocessing during SAR image analysis and processing.The combination of the two has a good representation in the extraction of image features.