UAV LiDAR Tree Segmentation via Canopy Morphology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current individual-tree segmentation methods for UAV LiDAR point cloud data face challenges in accurately identifying and refining wrong segmentation trees, leading to inefficiencies and information loss.
Innovation Solution
An individual-tree segmentation method based on canopy morphology, utilizing a grid-based region growing algorithm and density-based information guidance, which includes preprocessing to obtain canopy height and density models, extracting tree tops, determining correct and wrong segmentation canopies, and performing fine segmentation based on canopy morphology.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If grid-based method with Gaussian filter and CHM interpolation is used, then computational efficiency is improved, but information loss occurs and accuracy depends on filter window size
Solution Approach 1:
The method segments the segmentation process into two stages: coarse segmentation using grid-based CHM and fine segmentation using point cloud data. This segmentation allows the efficient grid-based method to handle the bulk of segmentation while a subsequent refinement stage corrects information loss, thus maintaining computational efficiency while reducing accuracy degradation.
Solution Approach 2:
The method performs preliminary coarse segmentation using the computationally efficient grid-based CHM method to obtain initial tree segmentation results. This preliminary action quickly reduces the problem space, and then a refinement stage corrects errors, thus maintaining overall efficiency while compensating for information loss in the preliminary stage.
2Measurement precision
If point-based method with clustering algorithms is used, then accuracy is improved and information loss is avoided, but calculation and time consumption increase hugely
Solution Approach 1:
Instead of applying the computationally intensive point-based method to all trees, the method applies it partially only to trees identified as incorrectly segmented by the grid-based method. This partial application maintains high accuracy for problematic cases while avoiding the huge time consumption that would result from processing all trees with the point-based method.
Solution Approach 2:
The grid-based CHM method serves as an intermediary that quickly identifies potentially incorrect segmentations, which then become the target for more accurate point-based refinement. This intermediary approach avoids the need to apply the time-consuming point-based method universally while still achieving high accuracy where needed.
3Productivity
If existing combined grid and point method is used, then initial segmentation is obtained, but wrong segmentation trees cannot be accurately identified and refined
Solution Approach 1:
The method implements a feedback mechanism where the grid-based segmentation results are evaluated, and trees identified as incorrectly segmented are fed back for refined processing using point cloud data. This feedback loop continuously improves segmentation accuracy by targeting and correcting specific errors rather than relying on a single-pass approach.
Solution Approach 2:
The method replaces the mechanical, uniform processing approach with an intelligent, adaptive system that uses density analysis and morphological features to identify and selectively refine problematic segmentations. This substitution of rigid mechanical processing with adaptive intelligence-based filtering enables accurate identification of wrong segmentations without sacrificing the speed benefits of grid-based initial segmentation.
Data Source
AI summary
Provided is an individual-tree segmentation method of UAV LiDAR point cloud based on canopy morphology. The method uses woodland data obtained by a UAV LiDAR to: extract initial canopies from a CHM based on a region growing algorithm, determine whether each initial canopy is a correct segmentation canopy according to the number of local density maximum points in each initial canopy, finely segment each wrong segmentation canopy according to canopy morphology to obtain an updated set of tree tops, and finally use each of the updated set of tree tops as a seed point for performing the region growing algorithm to thereby obtain a final individual-tree segmentation result. The method make full use of height information and density information contained in a tree point cloud; and under the guidance of the density information, a wrong segmentation tree can be more accurately identified.