基于时间序列MODIS NDVI的农作物物候信息提取

Extraction of Crop Phenological Information Based on Time Series MODIS NDVI

物候是植被长期适应环境的周期性变化而形成的生长发育规律,农作物物候研究对长势监测和产量评估以及气候变化都具有重要价值。针对现有研究的问题,改进物候信息遥感提取方法,并利用MODIS NDVI时间序列,提取伊洛河流域主要农业区的农作物物候信息。首先将8天间隔的NDVI时间序列进行Savitzky-Golay滤波去除噪音信息;再根据冬小麦和夏玉米的生物生理特性,利用与关键形态特征点的相对位置来界定物候期;最后提取多年关键物候期信息。监测结果与往年观测资料对比,研究结果客观可信。整体上物候期相对稳定,但当农作物受气候变化以及异常天气的影响,年际间会存在明显的差异。其中,出苗期相对变异最小,而冬小麦的抽穗期和夏玉米的抽雄期相对变异最大。研究发现10年间同一物候期相差最高达20天左右。结果证明了本文研究方法的可行性和有效性,及时掌握农作物物候信息对农业生产与管理,以及农业遥感的深入研究具有重要意义。

Phenology is the formed growth and development law of vegetation to adapt to the long-term cyclicalchange of environment, it has great value in growth monitor, yield assessment and climate change. Aiming atproblems in the existing researches, this study improved the remote sensing extraction method of phenologicalinformation, and extracted the crop phenology information by using the MODIS NDVI time series data in mainagricultural areas of Yiluo River Basin. Firstly, the 8-day interval NDVI time series data were processed bySavitzky-Golay filter to remove noise information. After that, the phenological dates were defined by therelative position to the key morphological feature points according to the biological physiologicalcharacteristics of winter wheat and summer maize. Finally, the key phenological information was extracted. Themonitored phonological information was objective and credible compared with previous observations. Thephenological period was relatively stable, however, there were clear interannual variations when crops wereaffected by climate change and abnormal weather. Among them, the relative variation of emergence stage wasthe smallest, while that of the heading stage of winter wheat and the tasseling stage of summer maize were themost obvious. It was also found that the maximum variation of phenology of the same plant was about 20 days.The results demonstrated that this study method was feasible and effective, it was of certain significance totimely grasping phenological information for agricultural production and management, as well as in-depthagricultural remote sensing research.