摘要
由于在利用原有方法进行SAR图像舰船检测数据挖掘时,受噪声较大的影响而无法进行舰船检测数据特征提取,在观测距离为500~1 300 m的范围内,存在特征提取率较低的问题,因此提出一种大数据分析下的SAR图像舰船检测数据挖掘方法。基于大数据分析,通过5个步骤对SAR图像数据实施规范化处理,包括分析SAR图像数据、对工作流与转换规则进行定义、验证SAR图像数据、规范化处理执行以及回流干净数据。通过经验模态分解法实施舰船检测数据特征提取。通过LIBSVM开发包构建支持向量机数据挖掘模型,实现SAR图像舰船检测数据挖掘。为证明该方法的特征提取率更高,在观测距离为500~1 300 m的范围内进行该方法与原有方法的对比实验,实验结果证明该方法的特征提取率高于其他方法,实现了挖掘性能的提升。
When using the original method to mine ship detection data from SAR image, it is difficult to extract the features of ship detection data due to the large noise. In the range of observation distance of 500~1 300 m, the feature extraction rate is low. Therefore, a ship detection data mining method based on big data analysis is proposed. Based on big data analysis, five steps are adopted to standardize SAR image data, including analyzing SAR image data, defining workflow and conversion rules, verifying SAR image data, implementing standardized processing and returning clean data. The empirical mode decomposition(EMD) method is used to extract the features of ship detection data. The data mining model of support vector machine(SVM) is constructed by libsvm development package to realize the data mining of ship detection in SAR image. In order to prove that the feature extraction rate of this method is higher, a comparative experiment is carried out between the method and the original method in the observation distance of 500~1 300 m. The experimental results show that the feature extraction rate of this method is higher than other methods, and the mining performance is improved.
作者
耿德志
GENG De-zhi(Department of Information Technology and Engineering,Jinzhong College,Jinzhong 030600,China)
出处
《舰船科学技术》
北大核心
2020年第22期61-63,共3页
Ship Science and Technology
