Research On Multi-temporal Airborne Li Dar Forest Point Cloud Registration And Single-tree Change Detection Method

Forest resources are an important natural and ecological resource.Multi-temporal determination of single trees in the forest to detect changes can provide a data basis for forest monitoring,which is conducive to forest resource management and dynamic monitoring of ecological conditions.The rapid development of light detection and ranging can quickly and accurately obtain three-dimensional coordinates on the surface of trees,and is widely used in forestry investigations.The forest point cloud data obtained by lidar technology can be used to extract forest structure,extract individual tree height,crown diameter and other parameters,so as to achieve efficient and highprecision forest monitoring.Airborne laser scanning technology uses aviation aircraft or drones equipped with lidar to quickly and efficiently scan a large area of forest,which can obtain richer canopy information,which is suitable for canopy reconstruction.However,the point cloud data of airborne laser scanning at a single time cannot reflect the changes of single tree information over time.Using the point cloud data obtained by multiple scans to calculate the difference of the extracted information of the same single tree can intuitively reflect the changes of the single tree,which is conducive to further improving the efficiency and accuracy of forest resource monitoring and realizing continuous forest resources dynamic monitoring.This paper attempts to match multi-phase airborne lidar forest point clouds,and detects single tree changes based on the matching results.The main research contents are as follows:（1）In view of the uneven density of forest point cloud data,the point cloud contains noise,the point cloud of the crown part is blocked,and the point cloud of the trunk part is missing,this paper uses a key point extraction method that combines the center of the trunk and the terrain features to generate the key points are used for point cloud registration.First search for the local maximum value of the canopy height model as the potential apex of tree,extract the coordinates of apexes of trees in the forest point cloud based on the potential apex of tree,use the relative position relationship between the apex of tree and the trunk to obtain the trunk point cloud,calculate the trunk center point and project it Go to the ground to generate key points for matching.（2）This paper proposes a single tree segmentation method combining the canopy height model and point cloud data.The canopy height model is used to determine the seed point,and the point cloud in the seed point neighborhood is clustered based on the threshold to complete the single tree segmentation.The deviation introduced by the interpolation operation in the process of generating the canopy height model reduces the number of point clouds involved in the calculation and maintains the accuracy of the three-dimensional coordinates of the point cloud data.（3）Based on the results of multi-temporal airborne lidar forest point cloud matching and single tree segmentation,the tree height and crown width changes of the corresponding single tree in the multi-temporal data are detected.The comparison between the experimental results and the measurement data in the sample area shows that the change detection effect is good and feasible.This method can effectively realize the dynamic monitoring of forest resources,and is beneficial to the management and utilization of forest resources and ecological environment protection.