# US 2026/0187309 A1 · Method for assessing wind loads on trees based on point cloud models

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0187309 A1
- Document ID: US20260187309A1
- Status: Published application
- Kind: A1
- Application number: 19/433,051
- Filed: 2025-12-25
- Published: 2026-07-02
- Earliest claimed: 2024-12-27
- Assignees: Tongji University
- Applicants: Tongji University
- Inventors: Yi Rui; Yu Xin Zhang; Xiao Jun LI; Hehua ZHU; Pei Nan LI; Mei YIN; Pei WANG; Ting Chen FANG; Jian Rong CHEN; Tong WANG
- Web page: https://patentgrep.com/patent/US20260187309A1

## CPC

- G06F 30/23: using finite element methods [FEM] or finite difference methods [FDM]
- G06T 7/0004: {Industrial image inspection}
- G06T 7/55: from multiple images
- G06T 7/75: {involving models}
- G06T 7/77: using statistical methods
- G06V 10/30: Noise filtering
- G06V 10/46: Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features (colour feature extraction G06V10/56)
- G06V 20/17: taken from planes or by drones
- G06V 20/188: {Vegetation}
- G06V 20/647: {by matching two-dimensional images to three-dimensional objects}
- G06V 20/70: Labelling scene content, e.g. deriving syntactic or semantic representations
- G06T 2207/10028: Range image; Depth image; 3D point clouds
- G06T 2207/10032: Satellite or aerial image; Remote sensing
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2210/56: Particle system, point based geometry or rendering
- Y02A 90/10: Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- Y02A: TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE

## Abstract

A method for assessing wind loads on trees based on point cloud models. This is achieved by performing semantic segmentation on images captured by unmanned aerial vehicles (UAVs), extracting images of the tree crowns for three-dimensional reconstruction, generating point cloud models, extracting point cloud data for both the crowns and trunks, and ultimately producing three-dimensional point cloud models for the target trees. Calculate the wind load impact area of tree crowns using the Alpha Shape algorithm. Combining wind speed, wind direction and tree geometric characteristics, simulate wind load distribution under varying wind direction and speed conditions using ANSYS Fluent software, and compute wind load results under applied wind loads. Finally, utilizing finite element analysis methods in conjunction with wind load data, structural stress analysis of trees is conducted to evaluate their fatigue lifespan and failure risk under wind loads.

## Claims

### Claim 1

A method for assessing wind loads on trees based on point cloud models, comprising the following steps:
employing unmanned aerial vehicles to capture multi-angle imagery of target tree, thereby acquiring visual representations of the subject trees;
applying image semantic segmentation algorithms to annotate the imagery, generating labels for each pixel within the images, these labels comprise tree crowns, tree trunks, and background elements;
constructing a preliminary three-dimensional point cloud model based on the annotated imagery, each point cloud data point within this preliminary model incorporates corresponding label information,
generating a three-dimensional point cloud model of the target tree based on the preliminary three-dimensional point cloud model and its label information;
calculating the wind load impact area of the tree crowns based on the aforementioned three-dimensional point cloud model of the target tree;
simulating multi-directional wind loads based on the wind load impact area of the canopy, wind speed, and the target tree's geometric characteristics to obtain wind load simulation results, wherein the target tree's geometric characteristics are derived from the target tree's 3D point cloud model;
conducting structural stress analysis on the tree in conjunction with the wind load simulation results to predict the risk of tree failure under wind forces,
the calculation of wind load impact area on tree crowns based on the target tree's three-dimensional point cloud model comprises the following steps:
extracting point cloud data labelled as the tree crowns from the target tree's three-dimensional point cloud model to generate a three-dimensional point cloud model of the tree crowns;
projecting the canopy point cloud data from the three-dimensional canopy point cloud model along a predetermined direction, converting the three-dimensional canopy point cloud model into two-dimensional planar data;
converting the three-dimensional coordinates within the three-dimensional canopy point cloud model into corresponding two-dimensional coordinates according to the predetermined direction, thereby obtaining a two-dimensional point set for the canopy,
employing the Alpha Shape algorithm to compute the boundary envelope of the two-dimensional point set representing tree crown yields the wind load impact area of the crown;
based on the wind load impact area of the crown, wind speed, and geometric characteristics of the target tree, simulate multi-directional wind loads to obtain wind load simulation results, comprising the following steps:
calculating wind resistance coefficients at the tree crowns and the tree trunk based on the target tree's geometric characteristics; generating a three-dimensional CAD model from the tree crown's point cloud data and import it into Fluent; inputting the wind resistance coefficients at the tree crowns and tree trunks into ANSYS Fluent software; using Fluent to simulate wind velocity and pressure distributions generated by winds from multiple directions; analyzing the resulting flow field simulation outcomes to obtain wind load simulation results, calculated using the formula:
F
=
1
2
·
C
d
·
A
·
ρ
·
v
2
where F denotes wind load, Cd represents the drag coefficient, A signifies the wind load-affected area of the canopy, ρ denotes air density, and v indicates wind velocity.

### Claim 2 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the image annotation performed via an image semantic segmentation algorithm generates labels for each pixel within the image, comprising the following steps:
performing preprocessing operations on captured images, including denoising and color correction;
inputting the preprocessed images into a pre-trained convolutional neural network model for semantic segmentation, outputting semantic segmentation maps for each image, each pixel within the semantic segmentation map is assigned a label, the convolutional neural network model includes the Mask R-CNN model.

### Claim 3 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the construction of a preliminary three-dimensional point cloud model based on annotated images comprises the following steps:
performing multi-view three-dimensional reconstruction of captured images using COLMAP three-dimensional reconstruction software to generate a non-semantic three-dimensional point cloud model;
computing the three-dimensional coordinates of each pixel within the semantic segmentation map of each image in the non-semantic 3D point cloud model, determining the corresponding point cloud data for each pixel based on its three-dimensional coordinates, passing the labels of each pixel to the corresponding point cloud data to generate a preliminary 3D point cloud model.

### Claim 4 (depends on 3)

The method for assessing wind loads on the trees based on the point cloud models according to claim 3, wherein the employing COLMAP three-dimensional reconstruction software to perform multi-view three-dimensional reconstruction on captured images, generating a non-semantic three-dimensional point cloud model, comprising the following steps:
the SIFT algorithm is employed to extract feature points and their descriptors from each image, a nearest neighbor matching algorithm is then applied to match feature points across multiple images based on their respective descriptors, the RANSAC algorithm is utilized to eliminate erroneously matched feature points,
based on the matched feature points, COLMAP is employed to construct a non-semantic 3D point cloud model through an incremental reconstruction method.

### Claim 5 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the generation of a target tree three-dimensional point cloud model from a preliminary three-dimensional point cloud model and its label information comprises the following steps:
by utilizing the semantic segmentation maps of each image and the geometric characteristics of point cloud data within a preliminary 3D point cloud model, non-tree debris is removed through geometric conditions based on color, position, and normal vectors;
point cloud data labelled as tree crowns and tree trunks is extracted from the preliminary 3D point cloud model;
simplifying the extracted point cloud data using a voxel grid algorithm, spatially partitioning the point cloud into a grid to reduce redundant points;
applying a SOR filtering algorithm to remove noise from the point cloud data, generating a three-dimensional point cloud model of the target tree.

### Claim 6 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the predetermined direction is dynamically adjusted according to variations in wind speed and wind direction.

### Claim 7 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the wind load impact area of the tree crowns is calculated using the following formula:
A
=
∑
i
=
1
n
⁢
1
2
⁢
❘
"\[LeftBracketingBar]"
x
i
⁢
y
i
+
1
-
x
i
+
1
⁢
y
i
❘
"\[RightBracketingBar]"
,
where A denotes the wind load-affected area of the tree crowns, (xi, yi) represents the two-dimensional coordinates of the i-th point cloud data within the tree crown's two-dimensional point set, n signifies the total number of point cloud data points in the tree crown's two-dimensional point set, and (xn+1, yn+1) is regarded as forming a closed path with (x1, y1).

### Claim 8 (depends on 1)

The method for assessing wind loads on the trees based on the point cloud models according to claim 1, wherein the combining wind load simulation results with structural stress analysis of the tree to predict the risk of failure under wind forces comprises the following steps:
based on the three-dimensional point cloud model of the target tree, a finite element model of the tree is constructed, the finite element model incorporates the tree's geometric shape, material properties, mechanical characteristics, and boundary conditions,
the results of the wind load simulation are input into the finite element model, acting as external loads applied to each node or element within the finite element model, and the structural stresses within the tree under wind load are calculated using the following formula:
σ
=
F
A
c
⁢
r
⁢
o
⁢
s
⁢
s
,
where σ denotes stress, F denotes wind load, and Across denotes the cross-sectional area of the wind load application zone;
based on the calculated structural stresses, combined with the fatigue strength and fatigue index of the target tree material, the fatigue life of the target tree under long-term wind loading is predicted using the following formula:
N
f
=
(
Δσ
σ
f
)
β
,
where Nf denotes fatigue life, Δσ denotes stress amplitude, obtained from the calculated structural stress, σf denotes the fatigue strength of the target tree material, and β denotes the fatigue index of the target tree material;
assess the risk of failure of target tree under wind loads based on their predicted fatigue life under long-term wind loading.

### Claim 9 (depends on 2)

The method for assessing wind loads on the trees based on the point cloud models according to claim 2, wherein the construction of a preliminary three-dimensional point cloud model based on annotated images comprises the following steps:
performing multi-view three-dimensional reconstruction of captured images using COLMAP three-dimensional reconstruction software to generate a non-semantic three-dimensional point cloud model;
computing the three-dimensional coordinates of each pixel within the semantic segmentation map of each image in the non-semantic 3D point cloud model, determining the corresponding point cloud data for each pixel based on its three-dimensional coordinates, and passing the labels of each pixel to the corresponding point cloud data to generate a preliminary 3D point cloud model.

### Claim 10 (depends on 2)

The method for assessing wind loads on the trees based on the point cloud models according to claim 2, wherein the generation of a target tree three-dimensional point cloud model from a preliminary three-dimensional point cloud model and its label information comprises the following steps:
by utilizing the semantic segmentation maps of each image and the geometric characteristics of point cloud data within a preliminary 3D point cloud model, non-tree debris is removed through geometric conditions based on color, position, and normal vectors;
point cloud data labelled as tree crowns and tree trunks is extracted from the preliminary 3D point cloud model;
simplifying the extracted point cloud data using a voxel grid algorithm, spatially partitioning the point cloud into a grid to reduce redundant points; and
applying a SOR filtering algorithm to remove noise from the point cloud data, generating a three-dimensional point cloud model of the target tree.
