基于半球摄影法的LAI自动组网测量系统设计与应用

Design and application of LAI automatic network measurement system based on hemispherical photography

针对叶面积指数地面测量工作量大、效率低等问题,研制基于半球摄影法的叶面积指数组网观测系统。该系统基于半球摄影技术和无线传感网络技术,改进图像自适应分割算法,对复杂条件下植被冠层图像进行准确分割,从而实时获取LAI等冠层结构信息;针对5 min采集频率原始数据波动的问题,提出“最稳定窗口”精校正算法,提取每日LAI代表值。与LAI-2200C对比试验结果表明,两者LAI测量值具有极显著相关性,玉米、林地和草地植被类型下相关系数分别为0.95、0.69和0.87;在江苏溧阳、吉林长春多植被类型试验中两种仪器测量结果的相关系数分别达到85%和81%,均为极显著水平;与MODIS MOD15A2H长时间序列产品对比试验也表明,两者的相关系数达到0.98,准确反映小麦从返青到成熟期叶面积动态变化趋势。系统具有快速准确、经济方便、使用范围广、可实现多站点长期监测的优点,已在多个观测站得到应用,其研制将对遥感产品真实性检验及农业、生态、气候变化等领域的科研与实践提供有力支撑。

Aiming at the problems of large workload and low efficiency of ground-based measurements of LAI,a network observation system for LAI based on hemispheric photography was developed.Based on hemispherical photography and Wireless Sensor Network(WSN)technology,the system improved the image adaptive segmentation algorithm to accurately segment vegetation canopy images under complex environment,so as to obtain canopy structure information such as LAI in real time;for the issue of fluctuation of raw data at 5-minute collection frequency,the“most stable window”precision correction algorithm was proposed to extract daily LAI representative values.The results of comparison experiments with LAI-2200C showed that the two LAI measurements were highly significantly correlated,with the coefficients of 0.95,0.69 and 0.87 for maize,woodland and grassland vegetation types,respectively.The correlation coefficients of the measurement results of the two instruments in the multi-vegetation type experiments in Liyang,Jiangsu and Changchun,Jilin also reached 85% and 81%,respectively,both of which were highly significant levels.The comparison experiment with MODIS MOD15A2H long-term series products also showed that the R^(2) of the two reached 0.98,which accurately reflected the dynamic change trend of wheat leaf area from green to maturity.The system has the advantages of fast and accurate,economical and convenient,wide range of use and multi-site long-term monitoring.The instrument has been applied in many observatories,and its development will provide strong support to the remote sensing products authenticity validation and scientific research and practice in the fields of agriculture,ecology and climate change.