Tree diameter at breast height extraction method and system based on mobile laser scanning point cloud

Abstract

The invention discloses a tree diameter at breast height extraction method and system based on mobile laser scanning point cloud, and relates to the technical field of forestry resource investigation. The method comprises the following steps: acquiring forest site cloud data based on mobile laser scanning; filtering the forest point cloud data to obtain tree point cloud; performing elevation normalization processing on the tree point cloud to obtain an elevation normalized point cloud; carrying out individual tree instance segmentation on the elevation normalization point cloud to obtain an individual tree instance point set; slicing the single tree instance point set to obtain a breast diameter slicing point set; and performing least square circle fitting based on intensity weighting on the DBH slice point set, and determining the DBH parameter of the individual tree according to a circle model parameter obtained by fitting. According to the method, the tree breast diameter parameters can be rapidly, conveniently and accurately obtained, and basic data is provided for forest operation management, forest resource investigation and carbon cycle model establishment.

Description

Technical Field

[0001] The present invention relates to the technical field of forest resource surveys, and particularly to a method and system for extracting tree breast diameters based on mobile laser scanning point clouds. Background Art

[0002] The breast diameter, also known as the Diameter At Breast Height (DBH), refers to the diameter of a forest tree at breast height (1.3 m above the ground), and is one of the most important measurement factors in forest resource surveys. Accurately obtaining the breast diameter of forest trees is an important prerequisite for monitoring the forest land environmental conditions and the growth status of forest trees. In the inventory of forest land resources, the traditional methods for measuring the breast diameter of standing trees mainly use primitive contact measurement tools such as breast diameter tapes and measuring tapes, which are widely used in the forestry field. However, this method consumes a large amount of human resources, and at the same time, the labor intensity and time cost are huge. Moreover, the manual measurement method is backward, has a single function, and poor accuracy. Manual reading and recording of data are prone to errors, seriously restricting the measurement efficiency. Traditional manual surveying and mapping are difficult to apply to large-scale forest area surveys and are increasingly unable to meet the development requirements of modern forestry measurement, survey, and monitoring. Therefore, it is very necessary to introduce automation technology in breast diameter measurement to reduce the manual labor cost and improve the measurement efficiency.

[0003] Light Laser Detection and Ranging (LiDAR), as an active remote sensing technology, has been widely used in the study of forest structure parameters. Static Terrestrial Laser Scanning (TLS) can obtain high-quality point cloud data with millimeter-level accuracy, but it has limitations in measurement efficiency and data integrity. During single-station scanning, information loss caused by occlusion leads to a low tree detection rate; when using multi-station measurement, it seriously affects the data acquisition efficiency, and subsequent multi-station data registration is required, with complex data processing. The data acquisition efficiency of Mobile Laser Scanning (MLS) is higher than that of TLS, and the occlusion effect is low, but precise positioning is required, usually relying on the Global Navigation Satellite System (GNSS) and Inertial Measurement Unit (IMU) to provide the required data. GNSS has the problem of intermittent access under tree canopy coverage, and new MLS systems use Simultaneous Localization and Mapping (SLAM) technology to determine relative position and orientation. Although the point cloud data collected by MLS is more complete than that of TLS, there is no significant advantage in the extraction accuracy of tree diameter at breast height. The accuracy of the point cloud data obtained by using SLAM-based MLS is relatively low, only reaching the centimeter level, and the noise is also significantly higher than that of TLS. Although SLAM technology can reduce the error caused by IMU drift, inaccurate registration is inevitable, and there is currently no SLAM algorithm that can make the accuracy of the point cloud data collected by MLS reach the TLS level. Poorly registered point cloud segments and mixed point noise data affect the extraction accuracy of tree diameter at breast height, and general point cloud preprocessing algorithms are difficult to remove this type of low-quality points. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for extracting tree diameter at breast height based on mobile laser scanning point cloud, so as to quickly, conveniently and accurately obtain tree diameter parameters, and provide basic data for forest management, forest resource survey and carbon cycle model establishment.

[0005] To achieve the above purpose, the present invention provides the following solutions:

[0006] A method for extracting tree diameter at breast height based on mobile laser scanning point cloud, comprising:

[0007] Obtaining forest site point cloud data based on mobile laser scanning;

[0008] Performing filtering processing on the forest site point cloud data to obtain tree point cloud;

[0009] Perform elevation normalization on the tree point cloud to obtain an elevation-normalized point cloud;

[0010] Perform single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set;

[0011] Perform slicing on the single-tree instance point set to obtain a DBH slice point set;

[0012] Perform least-squares circle fitting based on intensity weighting on the DBH slice point set, and determine the single-tree DBH parameter according to the circle model parameters obtained by the fitting.

[0013] Optionally, perform filtering on the forest site point cloud data to obtain a tree point cloud, specifically including:

[0014] Use the cloth simulation filtering algorithm to perform filtering on the forest site point cloud data to remove the ground point cloud in the forest site point cloud data and obtain a tree point cloud.

[0015] Optionally, perform elevation normalization on the tree point cloud to obtain an elevation-normalized point cloud, and the specific formula is:

[0016]

[0017] where p = (x p , y p , z p ) is any point in the tree point cloud, p' = (x' p , y' p , z' p ) is the point corresponding to point p after elevation normalization, z pDEM is the pixel value of pDEM, pDEM is the pixel point in the DEM closest to point p, and the DEM is a digital elevation model generated based on the ground point cloud.

[0018] Optionally, perform single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set, specifically including:

[0019] Use the fast Euclidean clustering algorithm to perform single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set.

[0020] Optionally, perform slicing on the single-tree instance point set to obtain a DBH slice point set, specifically including:

[0021] Perform slicing on the single-tree instance point set to extract a point cloud slice with a set thickness at a height of 1.3 meters from the trunk in the single-tree instance point set to obtain a DBH slice point set.

[0022] Optionally, perform least - squares circle fitting based on intensity weighting on the set of breast - height diameter slice points, and determine the single - tree breast - height diameter parameter according to the circle model parameters obtained by fitting, specifically including:

[0023] Construct an intensity - weighted coefficient matrix according to the point - cloud reflection intensity information of the set of breast - height diameter slice points;

[0024] Perform least - squares circle fitting on the horizontal - plane projection point - cloud data of the set of breast - height diameter slice points according to the intensity - weighted coefficient matrix to obtain circle model parameters;

[0025] Determine the single - tree breast - height diameter parameter according to the circle model parameters.

[0026] Optionally, the expression of the circle model parameters is:

[0027] X = [a b c] T =(C T WC) -1 (C T WD)

[0028] The expression of the single - tree breast - height diameter parameter is:

[0029]

[0030] Wherein, is the intensity - weighted coefficient matrix, is the first - order term matrix, is the second - order term matrix, w i，i is the weight of the breast - height diameter slice point p i (x i , y i ) is the horizontal - plane projection point coordinate of the breast - height diameter slice point p i , i = 1, 2,..., n, n is the number of points in the set of breast - height diameter slice points, a, b, c are all general equation parameters of the circle model, A = a / 2 is the abscissa of the center of the circle model, B = b / 2 is the ordinate of the center of the circle model, R is the radius of the circle model, and d is the diameter of the circle model.

[0031] A tree breast - height diameter extraction system based on mobile laser scanning point cloud includes:

[0032] A data acquisition module, which is used to acquire forest site point cloud data based on mobile laser scanning;

[0033] A filtering module, which is used to perform filtering processing on the forest site point cloud data to obtain tree point cloud;

[0034] An elevation normalization module, which is used to perform elevation normalization processing on the tree point cloud to obtain elevation - normalized point cloud;

[0035] A single-tree instance segmentation module for performing single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set;

[0036] A slicing module for slicing the single-tree instance point set to obtain a breast-height diameter slicing point set;

[0037] A breast-height diameter extraction module for performing least-squares circle fitting based on intensity weighting on the breast-height diameter slicing point set and determining the single-tree breast-height diameter parameter according to the circle model parameters obtained by the fitting.

[0038] An electronic device includes a memory and a processor. The memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud.

[0039] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud.

[0040] According to the specific embodiments provided by the present invention, the following technical effects are disclosed:

[0041] The method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud proposed by the present invention adopts a series of preprocessing operations such as point cloud filtering and elevation normalization, uses a clustering algorithm to achieve single-tree instance segmentation, and uses the least-squares circle fitting technology based on intensity weighting to extract the breast-height diameter parameter of the tree, realizing non-contact and non-destructive automatic extraction of the breast-height diameter parameter of the tree, and can quickly, conveniently and accurately obtain the breast-height diameter parameter of the tree, providing basic data for forest management, forest resource survey and carbon cycle model establishment, and having important significance for forest ecosystem monitoring. Description of the Drawings

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0043] Figure 1 It is a flowchart of the method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud of the present invention;

[0044] Figure 2 It is an absolute error distribution diagram of the breast-height diameter extraction values obtained by different methods in the embodiments of the present invention;

[0045] Figure 3The linear regression scatter plot of the DBH extraction values and the measured values obtained by different methods in the embodiments of the present invention. Detailed implementation manners

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0047] The purpose of the present invention is to provide a method and system for extracting the tree diameter at breast height based on mobile laser scanning point cloud to quickly, conveniently and accurately obtain the tree diameter at breast height parameters, and provide basic data for forest management, forest resource survey and carbon cycle model establishment.

[0048] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0049] The present invention provides a method for extracting the tree diameter at breast height based on mobile laser scanning point cloud. As Figure 1 shown, the method includes:

[0050] Step S1: Obtain the forest site point cloud data based on mobile laser scanning.

[0051] The forest site point cloud data is collected by using an unmanned vehicle-mounted laser scanning system built independently. In order to obtain relatively complete forest site point cloud data, the operator manually controls the movement of the mobile platform to ensure that the movement path covers the forest land to the greatest extent and can improve the data collection efficiency. After the collection is completed, the corresponding PCD point cloud map of the forest land can be obtained, and offline cropping processing is performed to retain the point cloud data of the area of interest in the forest land.

[0052] Step S2: Perform filtering processing on the forest site point cloud data to obtain the tree point cloud.

[0053] For the obtained forest site point cloud data, it not only contains the tree point cloud, but also a large number of ground points, which will hinder the detection and extraction of the tree point cloud. Ground filtering is a necessary preprocessing process for single-tree instance segmentation and is a prerequisite for separating the tree trunk from the ground. It is easier to implement single-tree instance segmentation based on non-ground point clouds. In this step, the cloth simulation filtering (CSF) algorithm is used to remove the ground point cloud.

[0054] Step S3: Perform elevation normalization processing on the tree point cloud to obtain the elevation-normalized point cloud.

[0055] The forest floor is not a standard plane, and there are height differences in the ground point cloud collected. Therefore, the bottoms of the trees are not on the same horizontal plane. To ensure that the point cloud of the breast diameter slice of each tree comes from the trunk point cloud at a height of 1.3 m from the ground, the elevation of the tree point cloud is normalized, and the tree point cloud is transformed to a horizontal plane at the same height. A Digital Elevation Model (DEM) is generated from the ground point cloud, and the elevation normalization is achieved by means of the pixel value recording the point elevation in the DEM, as shown in the following formula:

[0056]

[0057] where p = (x p , y p , z p ) is any point in the tree point cloud, p' = (x' p , y' p , z' p ) is the point corresponding to point p after elevation normalization, z pDEM is the pixel value of pDEM, representing the elevation of this point, pDEM is the pixel point closest to point p in the DEM, and DEM is the digital elevation model generated based on the ground point cloud.

[0058] Step S4: Perform single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set.

[0059] A single tree is the basic measurement unit, and the single-tree point cloud is the basis for automatically extracting tree structure parameters. The tree point cloud data needs to be segmented to extract the single-tree instance point set, so as to extract the interested part based on the single-tree instance point set to form the breast diameter slice point set of the single tree. This step uses the Fast Euclidean Clustering (FEC) algorithm to achieve single-tree instance segmentation. The FEC algorithm uses the Euclidean (L2) distance metric to measure the degree of closeness of unordered points and groups the close points into the same cluster, which can be described as:

[0060] min||P i -P' i ||2≥D th

[0061] where C i ={P i ∈P} is a cluster different from C' i ={P' i ∈P}, and D th is the maximum distance threshold.

[0062] Step S5: Perform slicing processing on the single-tree instance point set to obtain a breast diameter slice point set.

[0063] After the individual tree instance segmentation process, the point cloud data of each tree is obtained. The basic data for breast diameter extraction is the point cloud slice at a height of 1.3 m on the tree trunk. Therefore, in this step, the individual tree instance point set is sliced to extract the data of the region of interest. Specifically, a point cloud slice with a set thickness at a height of 1.3 m on the tree trunk in the individual tree instance point set is extracted. The point cloud slice can be described as:

[0064] P 1.3 ={p i =(x i , y i , z i ) ∈ P, z i ∈ |1.3 ± ∈|}

[0065] where P 1.3 is the point cloud set with a certain thickness at a height of 1.3 m on the tree trunk; p i is any point in the individual tree instance point set P; ∈ is the parameter related to the slice thickness (unit: m).

[0066] Step S6: Perform least squares circle fitting based on intensity weighting on the breast diameter slice point set, and determine the individual tree breast diameter parameter according to the circle model parameters obtained by fitting.

[0067] This step specifically includes: constructing an intensity weighting coefficient matrix according to the point cloud reflection intensity information of the breast diameter slice point set; performing least squares circle fitting on the horizontal plane projection point cloud data of the breast diameter slice point set according to the intensity weighting coefficient matrix to obtain the circle model parameters; and determining the individual tree breast diameter parameter according to the circle model parameters.

[0068] The specific calculation process of the individual tree breast diameter parameter is as follows:

[0069] Regarding the breast diameter slice projection point cloud data as a circle, the best fitting model parameters are determined using circle fitting technology based on this data. The mathematical representation of the fitting model is the general equation of a circle:

[0070] f(x, y) = x 2 + y 2 - ax - by - c = 0

[0071] According to the least squares principle, the sum of the squares of the minimized errors of the least squares circle fitting can be expressed as:

[0072]

[0073] The goal is to determine the optimal values of a, b, and c to minimize E(a, b, c). The general equation is written in matrix form for least squares solution:

[0074] CX = D

[0075]

[0076] Among them, C is the linear term matrix, D is the quadratic term matrix, (x i , y i ) is the horizontal plane projection point coordinates of the diameter at breast height slice point p i . i = 1, 2,..., n, where n is the number of points in the diameter at breast height slice point set, X is the circle model parameter matrix, and a, b, and c are all the general equation parameters of the circle model.

[0077] The intensity weighted coefficient matrix W is introduced by using the point cloud reflection intensity information, and the intensity weighted least squares circle fitting method for extracting the diameter at breast height is proposed in combination with the least squares method. The solution of the optimal circle model parameters can be expressed as:

[0078] X = (C T WC) -1 (C T WD)

[0079]

[0080] Among them, w i,i is the weight of the diameter at breast height slice point p i . By observing the distribution of the intensity value histogram, the intensity values of the slice point cloud are evenly divided into four intervals, and the proportion of the number of points in each interval to the total number of points is calculated. This proportion value is the weight of the points in the interval. The intensity weighted coefficient matrix is constructed to adjust the importance of different data points, reduce the contribution of mixed point noise data and low-quality registration data to the fitting result, and make the fitting result more accurate. According to the solved optimal circle model parameters X = [a b c] T , the center coordinates (A, B) and radius R of the fitted circle can be calculated. The calculation formulas are: A = a / 2, B = b / 2, The diameter d = 2R of the circle model is used as the diameter at breast height value of the tree. Based on the least squares circle fitting method, the present invention introduces the point cloud reflection intensity information and constructs a method for extracting the diameter at breast height based on intensity weighted least squares (IWLS) circle fitting, which can reduce the influence of mixed point noise data and low-quality registration data on the accuracy of diameter at breast height extraction.

[0081] Furthermore, this method further includes:

[0082] Step S7: Using the measured value data of the diameter at breast height of a single tree as the standard reference data, comparing the diameter at breast height parameter data of the single tree extracted from the forest land with the standard reference data to evaluate the accuracy of diameter at breast height extraction.

[0083] Select four evaluation indicators, namely Root Mean Square Error (RMSE), Absolute Error (AE), coefficient of determination R 2 and relative accuracy, to evaluate the extraction accuracy of diameter at breast height (DBH). The calculation formulas for the evaluation indicators are as follows:

[0084]

[0085] AE = |DBH k - DBH refk |

[0086]

[0087]

[0088] where K is the number of trees in the forest land; DBH k is the extracted DBH value; DBH refk is the measured DBH value; is the average value of the extracted DBH; is the average value of the measured DBH.

[0089] The IWLS DBH extraction method proposed in the present invention can effectively reduce the influence of mixed point noise data and low-quality registration data on the extraction accuracy of DBH and improve the extraction accuracy of DBH parameters by introducing point cloud reflection intensity information. Table 1 shows the extraction errors and accuracies of different methods. The results show that compared with the Random Sample Consensus (RANSAC) algorithm and the Ordinary Least Square (OLS) algorithm, the IWLS extraction method proposed in the present invention performs the best among the three methods. The two accuracy evaluation indicators (AE and RMSE) used to reflect the error both decrease, and the relative accuracy is the highest.

[0090] Table 1 Extraction errors and accuracies of different methods

[0091]

[0092]

[0093] It can be seen from the analysis of the data in Table 1 that the maximum AE values of the three extraction methods vary greatly. Among them, the maximum AE value of the RANSAC extraction method is 16.75 cm, and the maximum AE value of the IWLS extraction method is 2.89 cm. The minimum AE values of the three extraction methods vary slightly. Except that the minimum AE value of the RANSAC extraction method is 0.72 cm, the others do not exceed 0.23 cm. Further analyzing the absolute error values of the three extraction methods and presenting them in descending order, asFigure 2 As shown, where the four parts (a), (b), (c), and (d) correspond to forestlands 1, 2, 3, and 4 respectively. The absolute error of the RANSAC extraction method is significantly larger than that of the OLS and IWLS methods. Overall, the absolute error between the extracted value of the DBH by the IWLS method for fitting the circle model and the measured value is smaller and shows a decrease. By further analyzing and comparing the absolute errors of the three methods, it can be proven that the IWLS extraction method performs better than the RANSAC and OLS methods.

[0094] The scatter plot of the extracted value of the forestland DBH and the measured value of the DBH is as Figure 3 shown, where the four parts (a), (b), (c), and (d) correspond to forestlands 1, 2, 3, and 4 respectively. Comparing the linear regression scatter plots of different methods for the same forestland, the scatter distribution of the IWLS extraction method is closer to the scatter regression line, and the R 2 is closer to 1, with the highest goodness of fit, indicating that the extraction result of this method has a significant linear correlation with the measured value. The scatter plot shows that by adjusting the importance of data points through intensity information, the influence of low-quality point cloud data on the extraction accuracy of the DBH can be reduced, and the accuracy of the DBH fitting result can be improved.

[0095] To implement the above method to achieve the corresponding functions and technical effects, the following provides a tree DBH extraction system based on mobile laser scanning point cloud, including:

[0096] A data acquisition module for acquiring forestland point cloud data based on mobile laser scanning.

[0097] A filtering module for filtering the forestland point cloud data to obtain tree point cloud.

[0098] An elevation normalization module for performing elevation normalization processing on the tree point cloud to obtain elevation-normalized point cloud.

[0099] A single-tree instance segmentation module for performing single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set.

[0100] A slicing module for slicing the single-tree instance point set to obtain a DBH slice point set.

[0101] A DBH extraction module for performing least squares circle fitting with intensity weighting on the DBH slice point set and determining the single-tree DBH parameter according to the circle model parameters obtained by the fitting.

[0102] The present invention also provides an electronic device, including a memory and a processor. The memory is used to store a computer program, and the processor is used to run the computer program so that the electronic device executes the above-mentioned tree DBH extraction method based on mobile laser scanning point cloud. The electronic device can be a server.

[0103] In addition, the present invention further provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the above-mentioned method for extracting the diameter at breast height of trees based on mobile laser scanning point cloud.

[0104] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts among the various embodiments can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0105] Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A method for extracting the diameter at breast height of trees based on mobile laser scanning point clouds, characterized in that, Including: Obtaining forest point cloud data based on mobile laser scanning; Filtering the forest point cloud data to obtain tree point cloud; Performing elevation normalization processing on the tree point cloud to obtain elevation-normalized point cloud; Performing single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set; Performing slicing processing on the single-tree instance point set to obtain a breast-height diameter slicing point set; Performing least squares circle fitting with intensity weighting on the breast-height diameter slicing point set, and determining the single-tree breast-height diameter parameter according to the circle model parameters obtained by fitting.

2. The method for extracting the tree diameter at breast height based on the mobile laser scanning point cloud according to claim 1, characterized in that Filtering the forest point cloud data to obtain tree point cloud, specifically including: Using a cloth simulation filtering algorithm to filter the forest point cloud data to remove the ground point cloud in the forest point cloud data and obtain tree point cloud.

3. The method for extracting the tree diameter at breast height based on mobile laser scanning point cloud according to claim 1, wherein Performing elevation normalization processing on the tree point cloud to obtain elevation-normalized point cloud, and the specific formula is: where p = (x p , y p , z p ) is any point in the tree point cloud, p' = (x' p , y' p , z' p ) is the point after elevation normalization corresponding to point p, z pDEM is the pixel value of pDEM, pDEM is the pixel point in the DEM closest to point p, and DEM is the digital elevation model generated based on the ground point cloud.

4. The method for extracting the tree diameter at breast height based on the mobile laser scanning point cloud according to claim 1, characterized in that Performing single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set, specifically including: Using a fast Euclidean clustering algorithm to perform single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set.

5. The method for extracting the tree diameter at breast height based on mobile laser scanning point cloud according to claim 1, characterized in that, Performing slicing processing on the single-tree instance point set to obtain a breast-height diameter slicing point set, specifically including: Performing slicing processing on the single-tree instance point set to extract a point cloud slice with a set thickness at a height of 1.3 meters of the tree trunk in the single-tree instance point set, and obtaining a breast-height diameter slicing point set.

6. The method for extracting the tree diameter at breast height based on the mobile laser scanning point cloud according to claim 1, wherein Performing least squares circle fitting with intensity weighting on the breast-height diameter slicing point set, and determining the single-tree breast-height diameter parameter according to the circle model parameters obtained by fitting, specifically including: Constructing an intensity weighting coefficient matrix according to the point cloud reflection intensity information of the breast-height diameter slicing point set; Performing least squares circle fitting on the horizontal plane projection point cloud data of the breast-height diameter slicing point set according to the intensity weighting coefficient matrix to obtain circle model parameters; Determining the single-tree breast-height diameter parameter according to the circle model parameters.

7. The method for extracting the tree diameter at breast height based on the mobile laser scanning point cloud according to claim 6, characterized in that, The expression of the circle model parameters is: X = [a b c] T = (C T WC) -1 (C T WD) The expression of the single-tree breast-height diameter parameter is: Among them, is the intensity weighting coefficient matrix, is the linear term matrix, is the quadratic term matrix, w i,i is the weight of the breast diameter slice point p i . The weight of (x i , y i ) is the horizontal plane projection point coordinate of the breast diameter slice point p i . i = 1, 2,..., n, where n is the number of points in the breast diameter slice point set. a, b, and c are all parameters of the general equation of the circle model. A = a / 2 is the abscissa of the center of the circle model, B = b / 2 is the ordinate of the center of the circle model, R is the radius of the circle model, and d is the diameter of the circle model.

8. A tree diameter at breast height extraction system based on mobile laser scanning point cloud, characterized in that, Including: A data acquisition module for obtaining forest point cloud data based on mobile laser scanning; A filtering module for filtering the forest point cloud data to obtain tree point cloud; An elevation normalization module for performing elevation normalization processing on the tree point cloud to obtain elevation-normalized point cloud; A single-tree instance segmentation module for performing single-tree instance segmentation on the elevation-normalized point cloud to obtain a single-tree instance point set; A slicing module for performing slicing processing on the single-tree instance point set to obtain a breast-height diameter slicing point set; A breast-height diameter extraction module for performing least squares circle fitting with intensity weighting on the breast-height diameter slicing point set, and determining the single-tree breast-height diameter parameter according to the circle model parameters obtained by fitting.

9. An electronic device, characterized in that, Including a memory and a processor, the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, and when the computer program is executed by a processor, it implements the method for extracting the breast-height diameter of a tree based on mobile laser scanning point cloud as described in any one of claims 1 to 7.

Images