HJ-1B星海表温度产品业务化云检测算法研究

Study inoperational multi-spectral threshold method cloud detection for HJ-1B sea surface temperature products

可见光-红外遥感手段自身的缺陷,使得云层的存在直接影响海洋背景下卫星遥感定量化产品的精度。目前,尚未见到有关HJ-1B卫星红外相机IRS(InfraRed Scanner)传感器可业务化运行的便捷云检测方法研究的报道。针对HJ-1B卫星IRS传感器波段特征,对云与海洋下垫面反射及发射特性进行分析,提出了一种可业务化应用的海洋背景卫星图像云检测方法。选取渤海及黄海海域2012年5景不同时相、不同云被特征条件下IRS卫星影像数据进行实例验证分析,与准同步的MODIS云检测产品(MOD35)进行比对检验,发现该业务化算法具有不同时相条件下的普适性,能够有效检测全部的厚云、冷云及大部分的薄云。算法高效准确,适于业务化的应用。

For the defects of visible-infrared remote sensing technique,cloud over sea obstructs the accuracy of remote sensing products with ocean background. So far,there are no reports about HJ-1B / IRS operational cloud detection. Based on the characteristics of HJ-1B infrared scanner,this paper analyzed the reflection and emission characteristics of cloud and sea waters and proposed a new kind of operational cloud detection method with the situation of ocean background. Take the Bohai sea and northern Yellow sea for example,5 scenes of images,which were collected at different time in 2012 and under different underlying conditions,were applied with the new method. Comparison with MODIS cloud detection products-MOD35 by visual method to test the performance of the method,it was found that the method was universally applicable for different time phase and could effectively detect all thick and cold cloud and most of thin cloud. The method is concise and simple to operational application.