| Forest, covering30%of the global land and playing an irreplaceable role in absorbingcarbon dioxide, the flora and fauna, hydrological regulation, sand-fixing and consolidation ofthe soil, is one of the most important aspect of which the planet’s ecosystems constitutes. It iscomposed by the tree as the main surface biomes. It has a complex structure, rich in species, awide variety of functions. Forest and abiotic environment of the forest organically combined,to form a complete ecosystem. Forest which is the largest gene pool, carbon storage repository,reservoir, and energy library on Earth to maintain the ecological balance of the entire planet, isthe human survival, development of resources and the environment, even it is a tremendousand valuable green wealth of Nature.Remote sensing, as an emerging comprehensive detection of Science and Technology, is arelatively young discipline. However, because of its built in computer science, modern physics,mathematics and geoscience basis developing rapidly, it have carried out extensive researchand application, including the fields of environmental science, ecology, geology, geography,atmospheric science and oceanography. After decades of development, remote sensingtechnology has undergone tremendous changes in theory, technical and application. Theemergence and development of hyperspectral remote sensing is undoubtedly a very prominentpart in these changes. At the same time, as human society is facing the pressure of increasingpopulation, resources and the environment, people make unremitting efforts to seek possiblesolutions. Hyperspectral remote sensing can effectively solve the human predicament.Leaf area index refers to the ratio of land area occupied by the plants all the leaves oftheplant-sided area the sum of plant. The leaf area index is an important structural parametersof forest ecosystems and biological parameters that reflect the plant growing and a veryimportant plant ecological study botany parameters; also it is one of the most basic expressionsof the vegetation canopy structure parameters. Leaf area index has become an importantforest-quantitative evaluation. In this paper, using airborne hyperspectral CASI remote sensing data inversing forest leaf area index, aims to explore and analyze the ability of hyperspectralremote sensing invensing forest leaf area index, extracting of spectral bands sensitive,screening the optimal model. In order to obtain little new achievements and experience inforest applications of hyperspectral data from the theoretical and methodological, enrich andstrengthen China’s forestry remote sensing, promote the theoretical research and practicalapplication of Forestry Remote Sensing in china, remote sensing data of vegetation parameterquantitative retrieval can learn from this study. The main contents and conclusions are asfollows:(1) Find the appropriate preprocessing method of CASI data to preprocess it by choosingthe best method by studying deeply the imaging mechanism of the CASI data and imagecharacteristics.(2) It selected the best band combination of LAI inversion using C++programming.Near-infrared bands selected805.6nm, red-band selected724.6nm/729.4nm which can achievethe best inversion results.(3) We reasearched the inversion of forest leaf area index by normalized differencevegetation index, ratio vegetation index，RVI, Modified Soil Adjusted Vegetation Index,modified simple ratio index, and the accuracy of MSAVI is higher under the same conditions.(4) Selected CASI inversion of LAI of the optimal model. We found that using thequadratic polynomial model to estimate LAI results is the best if we selected a variety ofsingle-variablelinear and nonlinear prediction model to estimate of LAI.(5) Making the accuracy assessment and producting the maps of study area LAI levelaccording to the best vegetation index and the optimal model. |
PDF Full Text Request