利用鱼眼影像技术反演不同株型水稻的冠层结构参数  

Inversion of Rice Canopy Construction Parameters from the Hemispherical Photograph

作　　者：胡凝[1] 吕川根[2] 姚克敏[1] 张晓翠[3] 

机构地区：[1]南京信息工程大学应用气象学院/江苏省农业气象重点实验室,江苏南京210044 [2]江苏省农业科学院粮食作物研究所,江苏南京210014 [3]扬州大学农学院,江苏扬州225009

出　　处：《作物学报》2014年第8期1443-1451,共9页Acta Agronomica Sinica

基　　金：国家自然科学基金项目(30871479);江苏省农业气象重点实验室开放课题(KYQ1203)资助

摘　　要：快速、可靠、精确地评估植被冠层结构参数在大气-植被相互作用的研究中起着举足轻重的作用。为探明鱼眼影像在水稻冠层结构研究中的应用前景,本研究选择3种不同冠层结构的水稻品种作为研究对象,利用带有鱼眼镜头的数码相机在冠层的8个不同高度分别拍摄冠层影像,通过对影像的预处理提取冠层间隙度参数,根据冠层内辐射环境与冠层结构之间的定量化关系,利用Beer-Lambert定律反演水稻冠层的叶面积指数（leaf area index, LAI）和平均叶倾角（mean leaf angle, MLA）。研究结果表明,鱼眼影像反演的LAI均方根误差（root mean square error, RMSE）为1.2～1.5,相对误差（relative error, RE）为18.6%～22.5%,仅比人工测定结果低估7.6%～13.1%,优于 Sunscan 的测定结果。反演的MLA与人工测定结果之间有较好的一致性,相关系数为0.9205＊＊, RMSE为11.7°176;, RE为16.1%。研究结果表明,鱼眼影像反演水稻冠层结构是可行的方法。It is important to find a fast, reliable and accurate method for estimating canopy construction parameters in the study of vegetation-climate interaction. In the present study, three rice varieties with different canopy structures were chosen as experi-mental materials. A digital camera with a fisheye lens was used to take photos in eight heights of rice canopy to develop a new approach for concluding rice canopy. Canopy gaps were extracted from those photos, and then leaf area index （LAI） and mean leaf angle （MLA） could be inversed by Beer-Lambert theory, based on the quantitative relationship between the radiation condi-tion and the canopy structure. Results showed that LAI inverted from the hemispherical photograph was 7.6%-13.1%less than that measured manually, and the root mean square error （RMSE） between them was 1.2-1.5. The data showed that hemispherical photography was better than Sunscan canopy system. Moreover, MLA inverted from the hemispherical photograph was related to that measured manually with 0.9205＊＊of correlation coefficient and 11.7°; of RMSE. Therefore, hemispherical photography is a feasible technology to indirectly measure the rice canopy structure.

关 键 词：鱼眼影像 水稻 冠层结构参数 叶面积指数 平均叶倾角 

分 类 号：S511[农业科学—作物学]