论文部分内容阅读
针对云量自动评估算法难以检测Landsat图像中的半透明云问题,提出一种云量自动评估和加权支持向量机相结合的云检测算法。首先根据云在不同波段中的大气辐射特点,结合陆地卫星ETM+图像数据的光谱特性,利用云量自动评估算法将图像像元初步分成云像元、非云像元和待定像元,再以云的光谱特性构造特征向量,利用加权支持向量机算法进行待定像元的云层检测,最终获得全部图像的云检测结果。仿真试验结果表明,该方法既具有云量自动评估算法的云检测优势,还对云量自动评估算法难以识别的半透明云有较好的检测效果。
Aiming at the difficulty in detecting the translucent cloud in the Landsat image based on the automatic cloud-based assessment algorithm, a cloud detection algorithm based on cloud-based automatic assessment and weighted support vector machine is proposed. Firstly, according to the characteristics of cloud radiation in different wave bands and the spectral characteristics of land satellite ETM + image data, the image pixels are preliminarily divided into cloud pixels, non-cloud pixels and undetermined pixels using the cloud-based automatic evaluation algorithm. The feature vectors are constructed, and the cloud layer of undetermined pixels is detected by weighted support vector machine (SVM) algorithm. Finally, the cloud detection results of all the images are obtained. The simulation results show that this method not only has the advantages of cloud detection, but also can detect the translucent cloud which is difficult to identify by cloud automatic evaluation algorithm.