The Research Of Single Tree Parameters Extraction Based On Airborne LiDAR Data And Aerial Photography

The forest is one of the largest land resources, can provide the timber and forest products, the soil conservation, windbreak and sand, regulating climate, wildlife habitat and ecological functions. Changes in forest resources is not only related to social and economic development, but also have a huge impact on the environment and ecological. With the human environmental awareness in recent years, the importances of forest resources are more widely perception. Therefore a timely and efficient access to forest resources became a necessity. While the forest is composed of individual trees, the accuracy single tree parameters of the forest are necessary to ensure that detailed forest resources, so accurate single tree parameters to obtain high-precision far-reaching.LiDAR system as an active remote sensing technology, earth observation, especially the unparalleled advantage of other means of remote sensing to detect surface features height. Access by virtue of its woodland digital elevation model and detection of forest height advantage, the application of remote sensing in forestry climbed to a new level. Scholars at home and abroad to obtain the parameters of the single tree in the application of LiDAR data to do a lot of research, the research focused on two aspects of the methods and algorithms. Study found that although the advantages of LiDAR extract a single wood tree is significantly weaker abilities to describe the individual tree crown edge, we can not get individual tree crown information also affects the extraction accuracy of the single tree height.In this study, Yichun City, Heilongjiang Province, LiangShui Nature Reserve, airborne LiDAR point cloud data of the small spot data and to synchronize access to the large-scale aerial photographs for the data source, a high-precision single tree parameters extraction. Filtering and classification of airborne LiDAR data for the DEM and DSM, and both for the poor to get the CHM; Individual tree has been segmented in the DOM to extract a single tree crown polygon and vector format output, calculations, crown size and field measurement data inversion, the percent accuracy is More than85%; CHM and crown polygon overlay, search in the CHM the maximum elevation in the crown of each plant stumpage polygon, as the single wood tree; Select83sample trees in the field measurements using60strains as the training sample to estimate tree height and measured height correlation analysis to establish the regression equation, the use of23the remaining sample trees inversion using LIDAR to test the equation accuracy,86.037%. The results prove that large-scale aerial photographs combined with LiDAR data can be extracted out of the high-precision single tree parameters.