基于特征提取的遥感影像异常图斑识别方法  

Anomaly Spot Recognition Method of Remote Sensing Image Based on Feature Extraction

作　　者：崔玉莹 Cui Yuying(Shanghai Shengtu Remote Sensing Engineering Technology Co.,Ltd.,Shanghai,China)

机构地区：[1]上海盛图遥感工程技术有限公司,上海

出　　处：《科学技术创新》2024年第21期71-74,共4页Scientific and Technological Innovation

摘　　要：为发挥遥感影像更高的市场价值,实现对影像中异常图斑的精准识别,引进特征提取技术,开展遥感影像异常图斑识别方法的研究。遥感影像获取及预处理;引进特征提取技术对遥感影像进行关键区域分割;对分割后的遥感图像进行图斑纹理特征抽取,根据不同图斑的纹理特征变化,实现对图斑中异常信息的识别。将文献[1]、[3]方法作为参照,设计对比实验,实验结果证明:提出的基于特征提取的方法在应用中的效果良好,该方法可以精准辨识遥感影像中异常图斑,并实现对异常图斑的标注。In order to give full play to the higher market value of remote sensing image and realize the accurate recognition of abnormal spots in the image,the feature extraction technology is introduced to carry out the research on the abnormal spot recognition method of remote sensing image.Remote sensing image acquisition and preprocessing;Feature extraction technology is introduced to segment key regions of remote sensing images.The pattern texture feature is extracted from the segmented remote sensing image,and the abnormal information in the pattern is recognized according to the texture feature changes of different pattern spots.The methods in literature [1] and [3] were used as references to design comparative experiments.The experimental results proved that the proposed method based on feature extraction had good effect in application,and the method could accurately identify abnormal spots in remote sensing images and realize the annotation of abnormal spots.

关 键 词：特征提取 纹理 影像分割 识别方法 异常图斑 遥感影像 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]