| Leaf Area Index(LAI)is an input variable of ecological physiological model and ecosystem function of global change model,which is one of the important indexes in the field of ecology.Based on the practical problems of forest canopy LAI inversion in ecological research,this article in order to solve the LAI underestimated problem on optical image method.Some specific research work is as follows:(1)Analysis of the basic principle of the leaf area index inversed using the optical model combined with the feature of optical parameters in the model,the key factors affecting the LAI inversion were studied.In order to guarantee the timeliness of the research,this article compared the LAI measurement results under the same conditions beween using LAI-2000(radiation measuring principle)and Wincanopy2012a(image measuring principle)respectively,measuring locations is in Heilongjiang water national nature reserve 30 sampling point.The results shown that LAI measurement results of Wincanopy2012a are obvious less than measured results of LAI-2000.By comparing the two kinds of optical devices,found that the spectral measurement range of photoelectric sensor cause the main measurement difference;(2)LAI correction model has been proposed which is applicable to two-dimensional image on the basis of the calculation principle of the LAI-2000 and Wincanopy2012a.Using the area weighted method to revise the original weighted way of inversion model that improve the mathematical meaning.As the same time the the extinction coefficient of optical model is analyzed.Now the leaf Angle projection function G = 0.5 of the zenith angle 57.3 °is applied to the zenith Angle 0-90 degrees.And on this basis to build segmented fitting equation by polynomial on extinction coefficient of 6 kinds of typical leaf angle distribution and leaf angle of elliptic distribution,while summed up discrete values of extinction coefficient per 10 degrees zenith Angle,that modified single point inversion problem of the original model,so improve the physical meaning of the model;(3)The edge characteristics of effective hemisphere imaging region have been analyzed,the results show that the boundary linear existing scatter points to light source.On this basis,morphology algorithm has been used to extract the boundary and the least squares has been used to fitting the center of the circle and the radius value in effective imaging region.The effectiveness of this method has been verified by handmade standard template.The method has nothing to do with hardware acquisition device parameters,but affected by light intensity.The experimental results show that when the photon flux in the range of 55-140 ?mol m-2 s-1,the camera's sensitivity in the range of ISO 100-ISO 250,using the method to extract the centre point coordinates has the minimum error,in the meantime,the study also supply the quantified light conditions to image acquisition;(4)Research on effects of the segmentation algorithm and the segmentation number for gap fraction.The different threshold segmentation algorithms have been used to extract the gap fraction,such as minimum entropy algorithm,minimum error algorithm,Winscanopy algorithm and K-mean algorithm and the results shown that different segmentation algorithms causes the different value of gap fraction,among them,minimums error method has good inhibitory performance for big size gaps.Then the effective region has been divided into 6 rings or 9 rings or 18 rings.The values of the gap fractions contrastive result indicate that the effective region divided into 18 rings can effectively decrease the difference which caused by different threshold segmentation algorithm;(5)Analysis the performance of clumping index.The Lang and Xiang clumping index and Chen and Cihlar clumping index and Leblanc clumping index have been calculated respectively using the gap fraction data which are calculated by minimum entropy algorithm,minimum error algorithm,Winscanopy algorithm and K-mean algorithm.The results indicate that Leblanc clumping index has high sensitivity for difference of gap fractions,which can effective correct the difference of LAI inversion caused by image segmentation algorithm;(6)The LAI of the 30 sample points have been calculated using the proposed correction model without considering clumping index.When using the extinction coefficient discrete data of leaf angle elliptical distribution and using the minimum error algorithm to extract the gap fractions,the inverted results of LAI are close to the measured results of LAI-2000,that verify the validity of the model.This article solve the hemisphere image region extraction problem based on the research in the process of LAI inversion method,that create the conditions for autonomous hemisphere image acquisition equipment used in the vegetation canopy analysis or plant canopy.The openness LAI inversion method for forest canopy has been proposed that based on hemisphere image can effectively improve the LAI underestimate problem.The result of this research provides the necessary technical support to improve the controllability and the objectivity for forest canopy LAI inversion. |
PDF Full Text Request