Estimation Of Forest Volume Based On Unmanned Aerial Vehicle Remote Sensing

Forest resource management and survey methods are being combined with more and more new technologies.In this paper,the UAV is used as the image acquisition platform,and the high-resolution digital surface model(DSM)and digital Orthophoto Map(DOM)are obtained by software.The image is segmented by edge-penalty Hierarchical region merging.The data of crown width,tree number and tree height were obtained by GIS spatial analysis.The canopy density and stand volume were estimated.Finally,the accuracy of the extracted data was evaluated by field survey data.The main contents are as follows:1.High-resolution aerial image acquisition and generationWith the help of UAV remote sensing system,the image was collected by the small class in the high-Hong Kong work area of Baiyun Mountain State-owned Forest Park.The results of the collected images were processed by Pix4D Mapper software.After distortion correction,connection point extraction and adjustment,empty three encryptions,Match,orthophoto correction,get 0.06m resolution DOM and 0.16m resolution DSM.2.The edge penalty and hierarchical region merging methodIn the aspect of the extraction of canopy information,We used the edge length penalty as a index to segment the image,with three steps.Firstly,we use the direction edge information to construct the edge penalty image segmentation model,and the initial over-segmentation result is obtained by using MDRED and watershed transform.The edge of the edge is extracted by the contour of the polygon approximation,and the edge information of the image is extracted by MDRED.Using RAG and its NNG to represent image segmentation,the speed of regional merging is improved.The accuracy of the crown extraction was 83%and the accuracy of the number of factors was 80%.The results showed that the canopy density of the experimental area was consistent with the field estimation.3.Using block statistics to generate DEM and obtain tree height dataThe high and low canopy area has always been difficult to extract trees.According to the characteristics of forest gap,try to improve the DEM acquisition method in the high and low canopy area.The CHM(Canopy height model,CHM)is generated by DSM subtracting DEM,and the maximum value of CHM region is extracted by block statistic method by using the DSM value of the bare area and the DSM minimum of the block statistic.Point as the crown of the potential vertex,and its extraction process,get the crown of the crown.The precision of tree height extraction was 78%,79%and 81%,respectively,by comparing the accuracy with the measured data.4.Establish the crown-breast diameter model,and estimate stocking rate of Chinese fir.The high and low canopy area has always been difficult to extract trees.According to the characteristics of forest gap,try to improve the DEM acquisition method in the high and low canopy area.In this paper,we use the automatic extraction of Chinese fir crown and the diameter of Chinese fir to establish the regression model,and then extract the diameter of each tree in the study area.The total volume of experimental area is obtained by the binary volume equation,the volume of the stock collected from the experimental area is analyzed by using the high-diameter data of each tree.The overall accuracy is 59%,among which the Chinese fir the precision is 69%.In summary,this paper uses the UAV aerial remote sensing technology to estimate the forest volume,the extracted crown width,high tree canopy closure and tree number and other factors in the accuracy of 70%to 80%,the extraction of Chinese fir volume accuracy of 69%.This study helped to improve the information and equipment improvement of forest survey.