无人机观测时间对玉米冠层叶绿素密度估算的影响

Influence of unmanned aerial vehicle observation time on estimation of canopy chlorophyll density of maize

为探讨不同时间获取的无人机多光谱数据对玉米冠层叶绿素密度(canopy chlorophyll density,CCD)估算的影响,分别在玉米抽雄吐丝期、籽粒建成期、乳熟期和蜡熟期选择同一天的10:00—10:59、11:00—11:59、13:00—13:59和14:00—14:59进行无人机多光谱观测试验,并结合PROSAIL模型模拟结果与实测CCD数据,分析一天中不同时刻典型植被指数的变化规律及CCD估算结果的差异。结果表明:在同一天中,无人机玉米冠层反射率和与实测CCD相关性较好的植被指数值均随时间变化,近红外波段的反射率变化最明显,越接近12:00,实测的植被指数值越低,而在一天的不同时间PROSAIL模型模拟的植被指数值几乎没有差异。在同一天,基于不同观测时间获取的同一植被指数与实测CCD的相关性存在较大差异,且不同生育时期和不同指数间的差异不一致;而模拟得到的同一植被指数与CCD的相关性在同一天不同时间的差异不明显。在不同生育时期,基于不同观测时间无人机数据构建的CCD估算模型均可以取得较好的精度,但不同观测时间的估算结果存在差异,决定系数最低的为0.53,最高的为0.80。这些结果表明,在传统的光谱数据获取时间范围内(10:00—14:00),无人机影像获取时间仍对玉米CCD估算有影响,越接近12:00,估算精度越高。研究结果可为后续作物的CCD精准估算提供基础支撑。

In order to explore the effect of multispectral data of unmanned aerial vehicles(UAV)acquired at different time on the estimation of canopy chlorophyll density(CCD)of maize,the UAV multispectral observation experiment was conducted at 10:00—10:59,11:00—11:59,13:00—13:59,14:00—14:59 at the tasseling silking stage,blister stage,milk stage and dough stage,respectively.The PROSAIL model simulation results and the measured CCD data were combined to analyze the changes of the typical vegetation indexes at different time of the day and the difference between the CCD estimation results.The results showed that both the canopy reflectance and the value of vegetation indexes strongly correlated with the measured CCD changed with the time on the same day.The most obvious change of reflectance was found in the near-infrared band on the same day.The vegetation index value decreased as it approached 12:00 in the noon,whereas the value of vegetation indexes simulated by PROSAIL model at different time of the day showed little difference.The correlation between the same vegetation index obtained at different observation time and the measured CCD was different on the same day,and the difference between growth stages and indexes was not consistent;while the correlation between the same simulated vegetation index and CCD was not obvious at different time on the same day.CCD estimation models based on UAV data at different observation time could achieve good accuracy at different growth stages,but the estimation results at different observation time were different,with the lowest determination coefficient(R~2)of 0.53 and the highest R~2 of 0.80.These results showed that the traditional spectral data acquisition time range(10:00—14:00)had an impact on CCD estimation of maize,and the closer to the noon(12:00),the higher the estimation accuracy.These results provided basic support for the accurate CCD estimation of crops.