摘要
小波变换模型算法中的双高斯模型拟合干湿雪分类的最优阈值时,受初始值影响且典型样本区的选取比较费时,针对这些缺点提出了自动阈值分割的改进的小波变换算法,即用广义高斯模型自动拟合干湿雪分类的最优阈值.该算法继承和发展了冰盖冻融探测无需依赖于实测数据的优点,更好地实现了南极地区冰盖冻融监测系统建设的业务化运行目标.通过对改进前后的结果对比分析表明:改进后的方法与原方法相比,不仅提高了冰盖冻融探测方法的计算效率、实用性和可操作性,而且还在一定程度上提高了冰盖冻融探测的精度.
When the donble-Gaussian model of the current wavelet-transform algorithm fits the optimal threshold value of the dry and wet snow classification,it is easily affected by the initial value and has the shortcomings of consuming more time in selecting the typical sample zones.According to the fact,this paper proposed the improved wavelet-lransforrn algorithm for the automatic threshold segmentation, that is, generalized Gaussian model automalically fits the optimal wet and dry snow classification threshold. The algorithm inherits and develops the advantage of ice-sheet freeze-thaw detection, which does not rely on the measured data and achieves the goal of Antarctic ice-sheet monitoring system business more effectively.By comparing and analyzing of the two algo-ithms, we can see that the improved algorithm improves the computational efficiency, usability and otmability in the ice-sheet freeze-thaw detection as well as the accuracy of detection to some extent.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2013年第2期402-406,共5页
Acta Electronica Sinica
基金
国家自然科学基金(No.41076129)
国家863高技术研究发展计划(No.2008AA121702)
关键词
冰盖冻融探测
最优阈值
改进的小波变换
广义高斯模型
ice-sheet freeze-thaw detection
optimal threshold
improved wavelet transform
generalized Gaussian model
