混合智能优化算法的SAR图像特征选择

混合智能优化算法的SAR图像特征选择 Feature selection for SAR images using the hybrid intelligent optimization algorithm 张琴 ZHANG Qin 1 first-author 谷雨 GU Yu 1 徐英 XU Ying 2 赖晓平 LAI Xiaoping 1 杭州电子科技大学通信信息传输与融合技术国防重点学科实验室 杭州电子科技大学通信信息传输与融合技术国防重点学科实验室 Fundamental Science on Communication Information Transmission and Fusion Technology Laboratory,Hangzhou Dianzi University Fundamental Science on Communication Information Transmission and Fusion Technology Laboratory,Hangzhou Dianzi University 杭州电子科技大学生命信息与仪器工程学院 杭州电子科技大学生命信息与仪器工程学院 College of life information science & instrument engineering,Hangzhou Dianzi University College of life information science & instrument engineering,Hangzhou Dianzi University 为提高SAR图像自动目标识别的准确率及实时性,提出了一种基于混合智能优化的SAR图像特征选择算法。首先,采用分形特征对SAR图像进行增强,基于分割后的图像提出了一种基于图像矩的方位角估计方法。然后基于未校正和校正后的图像分别提取Zernike矩、Gabor小波系数和灰度共生矩阵构成候选特征集合,使用遗传算法结合二值粒子群的混合优化算法实现SAR图像特征选择。最后,采用MSTAR数据库验证本文算法的有效性。实验结果表明,优化后的特征集合具有一定泛化能力,一方面提高了SAR目标识别的准确率,另一方面减小了SAR图像目标识别的时间。 To improve the automatic target recognition accuracy of SAR images and real-time performance,this study proposes a feature selection algorithm based on hybrid intelligent optimization for such images. First,a fractal feature is used to enhance an SAR image. An azimuth estimation method is then developed based on the image moment after image segmentation. Subsequently,the features of Zernike moment,Gabor wavelet coefficients,and gray level co-occurrence matrix are extracted from the original and the rectified images to form feature candidates. The genetic algorithm and the binary particle swarm optimization algorithm are combined to select features for SAR images. The effectiveness of the proposed algorithm is verified with the MSTAR database. Results demonstrate that the optimal feature sets can be generalized,thereby improving the target recognition rate and reducing recognition time. SAR图像 特征选择 混合智能优化算法 分形特征 Zernike矩 SAR image feature selection hybrid intelligent optimization algorithm fractal feature Zernike moment TN957.52 国家自然科学基金（编号:61174024;61372024）;浙江省自然科学基金（编号:LQ13F050010）~~ 2021-06-09 1 20