1999—2013年中国东北植被物候信息遥感监测

随着3S技术的迅猛发展,用于地表遥感监测的卫星影像数据浩如烟海,而影像上地表植被长势等隐含信息的提取通常工作量大、耗费时间。论文提出了一种计算机自动提取地表植被物候信息的方法,主要用于对地表植被生长季及其年内长势进行快速提取。方法的基本原理是基于NDVI序列数据集,构建时间点对的时长跨度与该点对半方差的函数关系。选择东北松嫩平原地区作为重点试验区,计算了该区域1999—2013年地表植被生长季长度和长势特点,并选择东北地区物候观测数据进行验证分析。结果表明：1）东北地区农耕作物的生长季持续期在107~126 d左右。计算得到的结果与实测数据的最大误差在5 d值域范围内,沼泽植被在160 d以上（误差10 d左右）,草地为120~139 d;2）研究区地表植被的生长盛期峰值出现在第150天前后。结果较为真实、合理地反映了研究区域地表植被的物候信息。

Remote Sensing Monitoring of Vegetation Phonological Features in Songnen Plain of Northeast China during 1999-2013

With the rapid development of 3S technology, there are vast volume of satellite image data can be used in monitoring the surface vegetation coverage, moreover, the extraction of the implied information in the images, mainly referred to the vegetation ecological features such as the annual semi-variance of surface vegetation growth and the duration length of the vegetation-growing season, are often heavy work and time-consuming. A new extraction method using the variation function is proposed to quickly and accurately extract vegetation ecological features and study the vegetation change in the growing season. The method uses normalized difference vegetation index (NDVI) dataset, and builds the relational function of the time span and the semi-variance of each point pair based on the NDVI dataset. The Songnen Plain is selected as the experimental zone, where the surface vegetation is analyzed to extract the duration and the semi-variance of the vegetation growing season. The extracted data were statistically analyzed and verified. Eventually, the calculated result was compared to the actual local phonology. The results are as follows: 1) The largest differences (the maximum of semi-variance function) of the surface vegetation growth status appeared at the time-interval of 150 d. 2) The duration of the rain-fed crops in the growing season is 107-126 d, while the deviation of the crops between the fitted values and the measured values of the duration in the growing season is less than 5 d. Similarly, for grassland the range is 120-139 d (deviation, 5-10 d), for marsh vegetation it is 160-170 d (deviation, 10 d). So the results resembled the ecological state of the surface vegetation.