基于支持向量机的云图自动识别和提取方法

Automatic Identification and Extraction Clouds from Astronomical Images Based on Support Vector Machine

在时域天文研究中,中小口径光学望远镜的自动巡天是获取观测数据的重要来源.受台站观测条件影响,图像中会有云的干扰,给测光和暗弱目标的自动提取带来很大的困难.为此有必要对天文图像中的云进行识别和提取,建立云的指标图,为后续的信息提取提供参考.为了实现这一目标,建立了天文图像筛选系统,该系统将图像灰度不一致度与纹理差异作为图像的多维特征指标,并与支持向量机(Support Vector Machine,SVM)相结合,对有云的数据进行筛选.在筛选的基础上,利用直方图变换和特征提取,进一步提取了云的指标图.实验结果表明:纹理识别结合SVM对天文图像进行实时筛选,分类准确率大于98%.利用直方图变换和特征提取可以初步提取云的指标图,给测光和暗弱目标的提取提供参考.

For the time-domain astronomical observations,optical telescopes with small and medium size of aperture can get huge amount of data through automatic sky surveying.A certain proportion of automatically acquired data are interfered by clouds,which makes it very difficult for automatically extracting dim objects and photometry.Therefore,it is necessary to identify and extract clouds from these images as the index figures for reference during post-processing.In this paper,an astronomical image selection system based on support vector machine is proposed,which sets the grayscale inconsistency and texture difference as reference to select images interfered by clouds.Based on the classification results,the clouds can be extracted from the images through histogram transformation and feature selection.The results show that our method can achieve real-time selection for astronomical images with a classification accuracy of more than 98%.The histogram transformation and feature selection can preliminarily extract the images of clouds as references for photometry and dim object extraction.