基于Landsat8的南极内陆考察车辙印自动提取

分析了内陆考察车辙印的Landsat 8卫星影像特征,利用目视解译以及统计量分析车辙印与背景雪面的差异,得出近红外和短波红外波段是车辙印自动提取的优选波段,高斯增强方法为提升车辙印与背景之间的对比度的优选方案;同时运用动感模糊、canny边缘提取算子、膨胀腐蚀算子和噪声线段去除技术,实现车辙印在不同环境下的高精度自动提取.通过对中国第33次南极考察期间的内陆考察车辙印进行自动提取试验验证,此算法直接实现54%路线上车辙印的自动提取,32%的路线略需人工干预,仅14%的路线难以实现自动识别.另外,本文结合雪积累率、雪粒径大小、冰流速以及地形数据分别分析了气候、冰雪物理特性、冰川动力特征和地形特征对车辙印识别的影响.     

Antarctic accumulation seasonality

The resemblance of the orbitally filtered isotope signal from the past 340 kyr in Antarctic ice cores to Northern Hemisphere summer insolation intensity has been used to suggest that the northern hemisphere may drive orbital-scale global climate changes. A recent Letter by Laepple et al. suggests that, contrary to this interpretation, this semblance may instead be explained by weighting the orbitally controlled Antarctic seasonal insolation cycle with a static (present-day) estimate of the seasonal cycle of accumulation. We suggest, however, that both time variability in accumulation seasonality and alternative stable seasonality can markedly alter the weighted insolation signal. This indicates that, if the last 340 kyr of Antarctic accumulation has not always looked like the estimate of precipitation and accumulation seasonality made by Laepple et al., this particular accumulation weighting explanation of the Antarctic orbital-scale isotopic signal might not be robust.