基于KI算法的多时相ASAR影像水面信息变化检测

Change Detection of Water Surface in Multitemporal ASAR Images Based on KI Algorithm

合成孔径雷达(SAR)已成为国际对地观测最重要的前沿技术之一,独具全天时、全天候观测能力和对地表穿透性等特征,使之成为快速、准确获取水面信息,宏观监测水资源的重要手段。提出了一种基于KI算法的水面信息自动变化检测方法,以高级合成孔径雷达(ASAR)影像数据为实验数据,分析了水体和其它地物在影像上的差异。采用log-ratio算子生成差异影像,在差异影像服从泛高斯分析的前提下,探讨了水面变化信息的自动提取。试验结果验证了该方法能够快速、自动、精确获取水面变化信息。表明多时相ASAR数据能够广泛用于多云多雨地区的水面信息变化获取,水资源管理等领域。

SAR（synthetic aperture radar） remote sensing has the capabilities of 24-hour and all-weather observation and the ability to penetrate the terrain surface. Especially, with the recent development of new imaging radar technologies, SAR has become a useful and efficient method on monitoring the water surface information changing and on investigating water resources. By using multitemporal ASAR data, this paper proposes a fast water surface information changing detection approach integrating log-ratio operator and KI algorithm. Firstly, after removing the speckle noise of original image data, the log-ratio operator is used to represent the image in- formation as changed and unchanged classes. Secondly, by using KI algorithm, the threshold between these two classes can he detected quickly and exactly, then the changing information of water surface obtained automatically. Finally, the result of extraction of water surface information changing is evaluated by a method of comparison. Experimental results on ASAR data prove the effectiveness of the proposed automatic method in detecting the change information of water surface and the approach in this paper is approved to be an effective method.