一种识别NOAA/AVHRR数据异常高温点方法的研究

Method for Recognising the Unusual High Temperature Points through NOAA/AVHRR Data

近年来,极轨气象卫星(NOAA)的应用范围已由气象扩展到海水温度、海冰、水涝灾害、森林火灾和草地火灾等方面的监测,植被宏观监测及城市热岛现象研究等广阔领域,应用技术正向高水平发展,其目标是:定量化应用研究的开展,如农作物估产研究等;提高宏观监测精度和准确率的各种方法的研究和应用。

In this paper, for the error differentiation of applying neural network to recognize the unusual high temperature point by using NOAA data, on the base of KL transformation, the sample description is redefined for recognizing unusual high temperature problem with data supplied by the NOAA satellite. And the old Toptlogical structure of the network is modified. A new simulated experiment shows that the modified network is better than the old one in the recognizing precision.