CN118094397A - Crown base height prediction method and device, electronic equipment and storage medium - Google Patents
Crown base height prediction method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN118094397A CN118094397A CN202410504507.4A CN202410504507A CN118094397A CN 118094397 A CN118094397 A CN 118094397A CN 202410504507 A CN202410504507 A CN 202410504507A CN 118094397 A CN118094397 A CN 118094397A
- Authority
- CN
- China
- Prior art keywords
- height
- sample
- canopy base
- base height
- canopy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000003860 storage Methods 0.000 title claims abstract description 17
- 238000007621 cluster analysis Methods 0.000 claims abstract description 22
- 238000004364 calculation method Methods 0.000 claims description 40
- 230000006870 function Effects 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 24
- 230000003287 optical effect Effects 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 238000009795 derivation Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 description 36
- 238000005259 measurement Methods 0.000 description 13
- 238000011160 research Methods 0.000 description 9
- 238000012795 verification Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 238000001228 spectrum Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 239000002023 wood Substances 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000003908 quality control method Methods 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 244000025254 Cannabis sativa Species 0.000 description 1
- 241000196324 Embryophyta Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention provides a canopy base height prediction method, a device, electronic equipment and a storage medium, belonging to the technical field of satellite remote sensing, wherein the method comprises the following steps: acquiring a plurality of echo energies of a forest area to be detected; respectively carrying out height value matching on each echo energy to obtain a plurality of height information; performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions; and inputting all the height clustering clusters into a canopy basal height prediction model to obtain a canopy basal height prediction value of the forest region to be detected. According to the canopy base height prediction method provided by the invention, the plurality of height clustering clusters at different tree parts are obtained through clustering, so that the model can be convenient to determine the tree structure, the canopy base height prediction value of the forest region to be detected is further output through the canopy base height prediction model, the laser radar has a larger observation range, the fine prediction in the region range can be realized, and the accuracy of canopy base height prediction in the region range is improved.
Description
Technical Field
The invention relates to the technical field of satellite remote sensing, in particular to a canopy base height prediction method, a canopy base height prediction device, electronic equipment and a storage medium.
Background
Crown base height generally refers to the vertical distance between the ground surface and the base of a continuous living crown (i.e., the lowest branch of the tree crown with living leaves), and is one of the most important forest structure parameters in a variety of forestry applications, both at the stand level and at the individual plant forest level. The method can be used as an index of tree leaf quantity and seedling growth, is used for wood quality evaluation, crown condition and forest health evaluation, and is a key input parameter of various fire behavior prediction models.
The traditional forestry extraction canopy base height needs manual in-situ single wood measurement, and the detailed single wood measurement is time-consuming and labor-consuming. Remote sensing techniques provide another solution for computing forest parameters and are more time and cost efficient. However, different from calculation of the height of the canopy of the forest, the height of the canopy base needs to be obtained to calculate, but the remote sensing technology shoots the ground from high altitude and is shielded by upper branches and leaves, and the optical data cannot directly calculate the height of the canopy base. The laser radar signal can penetrate through the forest canopy, but the acquired reflected signal needs to be analyzed to find the position of the echo of the crown base. At present, the calculation research of the canopy base height based on the laser radar is based on laser radar point cloud and echo data acquired by a foundation and an unmanned aerial vehicle platform, and the direct estimation of the point cloud and waveforms is realized through regression analysis. Another method is to directly analyze the vertical section of the laser radar point cloud, and adopts a voxel method, namely, analyzing the point frequency in the vertical section and determining the rising speed of the point frequency in the section. The canopy base height calculation method based on the foundation and the unmanned airborne laser radar provides an informationized means, but has a limited observation range, so that the accuracy of canopy base height prediction in a regional range is low.
Disclosure of Invention
The invention provides a canopy base height prediction method, a device, electronic equipment and a storage medium, which are used for solving the problem of low canopy base height prediction accuracy in a regional range in the prior art.
In a first aspect, the present invention provides a method for canopy base height prediction, including:
Acquiring satellite-borne laser radar data of a forest area to be detected;
Determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
respectively carrying out height value matching on each echo energy to obtain a plurality of height information;
performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
Inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height prediction value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
In one embodiment, the canopy base height prediction model is obtained by training as follows:
acquiring a satellite optical data sample, a topographic data sample and a meteorological data sample of a forest total area sample;
Performing feature extraction based on the satellite optical data sample, the topographic data sample and the meteorological data sample to obtain a plurality of forest feature samples;
determining all height cluster samples corresponding to the forest area samples respectively and canopy base height labels corresponding to the forest area samples respectively; the forest area sample is a preset area of the forest total area sample;
Model training is carried out based on all the height cluster samples corresponding to each forest feature sample, all the height cluster samples corresponding to each forest region sample and all the canopy base height labels corresponding to each forest region sample, and a canopy base height prediction model is obtained.
In one embodiment, when determining all of the highly clustered samples respectively corresponding to the plurality of forest area samples and the canopy base high labels respectively corresponding to the plurality of forest area samples, the following steps are performed for each forest area sample:
Acquiring a plurality of echo data samples of a current forest area sample; the echo data samples include echo energy samples and height information samples;
Performing cluster analysis on each height information sample to obtain a plurality of height cluster samples at different tree parts; the plurality of high-level cluster samples at least comprise crown cluster samples and ground cluster samples;
and carrying out canopy base height calculation based on each high clustering sample and/or each echo data sample to obtain a canopy base height label of the current forest area sample.
In one embodiment, the canopy base height calculation is performed based on each highly clustered sample to obtain a canopy base height tag of the current forest area sample, including:
determining a high cluster threshold between the crown cluster sample and the ground cluster sample;
and determining the height information sample corresponding to the height cluster threshold as a canopy base height label of the current forest area sample.
In one embodiment, performing canopy base height calculation based on each echo data sample to obtain a canopy base height tag of a current forest area sample, including:
constructing a piecewise function between the echo energy samples and the height information samples based on the echo energy samples and the height information samples;
performing first-order derivation on the piecewise function to obtain a derivative of the piecewise function;
Constructing a graph according to the derivative of the piecewise function;
And determining the height information sample corresponding to the jump position in the graph as a canopy base height label of the current forest area sample.
In one embodiment, the canopy base height calculation is performed based on each highly clustered sample and each echo data sample to obtain a canopy base height tag of the current forest area sample, including:
determining the maximum echo energy sample as the total height of the tree;
Determining a crown cluster center of the crown cluster sample;
Determining a height information sample corresponding to the crown cluster center as the crown center height;
and calculating the canopy base height based on the total height of the tree and the central height of the crown, and obtaining a canopy base height label of the current forest area sample.
In one embodiment, the calculating the canopy base height based on the total height of the tree and the central height of the crown to obtain a canopy base height tag of the current forest area sample includes:
Calculating the difference between the total height of the tree and the central height of the crown to obtain a first difference;
performing double product calculation on the first difference value to obtain a product value;
Calculating the difference value between the total height of the tree and the product value to obtain a second difference value;
and determining the second difference value as a canopy base high label of the current forest area sample.
In a second aspect, the present invention further provides a canopy base height prediction apparatus, including:
The acquisition module is used for acquiring satellite-borne laser radar data of the forest area to be detected;
The determining module is used for determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
The matching module is used for respectively carrying out height value matching on the echo energy to obtain a plurality of height information;
The cluster analysis module is used for carrying out cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
The canopy base height prediction module is used for inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height predicted value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of any one of the canopy base height prediction methods described above when the program is executed.
In a fourth aspect, the present invention also provides a storage medium comprising a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of any of the canopy base height prediction methods described above.
According to the canopy base height prediction method, the device, the electronic equipment and the storage medium, the plurality of echo energies of the forest area to be detected are obtained, the corresponding height information is matched based on each echo energy, the plurality of height information is further subjected to clustering analysis to obtain the plurality of height clustering clusters at different tree positions, the tree structure can be conveniently determined by a model, all the height clustering clusters are further input into the canopy base height prediction model, the canopy base height prediction value of the forest area to be detected, which is output by the model, is obtained, the automation level of the canopy base height prediction mode is realized, the time and the economic cost consumed by the canopy base height prediction are effectively reduced, in addition, the satellite-borne laser radar data are observed through the laser radar, the satellite-borne radar has a larger observation range, the fine canopy base height prediction in the larger area range can be realized, and the accuracy of the canopy base height prediction in the area range is improved.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a canopy base height prediction method provided by the invention;
FIG. 2 is a schematic view of a high degree of clustering provided by the present invention;
FIG. 3 is a graph of the first derivative provided by the present invention;
FIG. 4 is a schematic structural diagram of a canopy-based height prediction device provided by the present invention;
fig. 5 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The terms "first," "second," and the like in this specification are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged, as appropriate, such that embodiments of the present invention may be implemented in sequences other than those illustrated or described herein, and that the objects identified by "first," "second," etc. are generally of a type, and are not limited to the number of objects, such as the first object may be one or more.
The following describes a method, a device, an electronic device and a storage medium for predicting the base height of a canopy according to the present invention with reference to fig. 1 to 5.
FIG. 1 is a schematic flow chart of a canopy base height prediction method provided by the invention.
As shown in fig. 1, the method for predicting the canopy base height provided by the present invention includes, but is not limited to, the following steps:
Step 100: acquiring satellite-borne laser radar data of a forest area to be detected;
step 200: determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
Step 300: respectively carrying out height value matching on each echo energy to obtain a plurality of height information;
step 400: performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
Step 500: and inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height predicted value of the forest area to be detected, which is output by the canopy base height prediction model.
It should be noted that, the canopy base height prediction method provided by the embodiment of the invention is realized based on the canopy base height prediction device, and the canopy base height prediction method is mainly applied to satellite large-scale observation, fills the technical blank that the existing algorithm obtains samples in a large area, and realizes small-scale observation from field measurement to unmanned plane area and then satellite large-scale observation. Therefore, the embodiment of the invention describes the canopy base height prediction method by taking the canopy base height prediction device as an execution main body as an example.
Specifically, the canopy base height prediction device acquires satellite-borne laser radar data of a forest area to be detected, and determines a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data, wherein the forest area to be detected is set according to actual conditions.
It should be noted that, the open earth observation system data portal may provide satellite-borne laser radar data, where the satellite-borne laser radar data is data of ground points acquired by a laser radar apparatus, and these data are usually presented in a sparse point cloud form, where each point represents a sampling point of a ground or vegetation structure, so that the forest area to be detected is a satellite-borne laser radar data point, and a satellite-borne laser radar data point may include a large area in a space of 25 meters×25 meters. In addition, after the laser radar apparatus transmits a pulse, the echo signal is reflected to generate echo energy at different heights and returned to the laser radar apparatus. By extracting 101 height values representing different echo energy for the preprocessed and corrected telemetry data (i.e., L2A data products), height information corresponding to echo energy at different heights can be determined.
Further, the canopy base height prediction device respectively performs height value matching on each echo energy to obtain a plurality of height information.
Further, the canopy base height prediction device performs cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions, wherein the cluster analysis mode mainly comprises a K-Means height cluster algorithm (K-Means algorithm), a mean shift height cluster algorithm (MEAN SHIFT algorithm) and a Density height cluster algorithm (Density-based spatial height cluster and noise application (DBSCAN) algorithm).
It should be noted that the K-Means algorithm is an iterative high clustering algorithm, whose goal is to divide data points into K different clusters, where K is specified in advance by the user. The algorithm assigns each data point to the cluster closest to it by continually iteratively updating the center points of the clusters to minimize the squared distance between the data point and the cluster center to which it belongs. The K-Means algorithm is a commonly used highly clustered algorithm, applicable to most data sets, but sensitive to the shape and size of the cluster.
It should be further noted that MEAN SHIFT algorithm is a density estimation and high clustering algorithm, which does not require the user to pre-specify the number of clusters. The algorithm repeatedly performs the process by finding the density gradient direction of the data point in the feature space, moving the data point to the region with the highest density until the convergence condition is reached. The MEAN SHIFT algorithm has better robustness to irregularly shaped clusters and noise data, and is generally used in the fields of image segmentation, object tracking and the like.
It should be further noted that the DBSCAN algorithm is a density-based high clustering algorithm, and the number of clusters does not need to be specified in advance. The algorithm divides the data points into clusters of core points and noise points by defining a concept of core points, direct density reachable points and density reachable points. The DBSCAN algorithm can automatically identify clusters with different shapes and sizes, and has good robustness for high-dimensional data and data sets containing noise.
Further, the canopy base height prediction device inputs all the height clustering clusters into a canopy base height prediction model to obtain a canopy base height prediction value of a forest area to be detected, which is output by the canopy base height prediction model, wherein the canopy base height prediction model is obtained by model training based on a forest feature sample and a canopy base height label.
The canopy base height prediction value of each to-be-detected forest area in the preset forest total area can be obtained through the canopy base height prediction model, a canopy base height distribution diagram of the preset forest total area can be further output, and the canopy base height of each to-be-detected forest area is displayed in the canopy base height distribution diagram. In addition, in the canopy base height distribution map, different height values can be distinguished by adopting color depth so as to present the canopy base height distribution condition of the whole preset forest total area.
According to the canopy base height prediction method provided by the invention, the plurality of echo energies of the forest area to be detected are obtained, the corresponding height information is matched based on each echo energy, the plurality of height information is further subjected to clustering analysis to obtain the plurality of height clustering clusters at different tree positions, the model can be convenient to determine the tree structure, all the height clustering clusters are further input into the canopy base height prediction model, the canopy base height prediction value of the forest area to be detected, which is output by the model, is obtained, the automation level of the canopy base height prediction mode is realized, the time and economic cost consumed by the canopy base height prediction are effectively reduced, the satellite-borne laser radar data are observed through the laser radar, and the satellite-borne radar has a larger observation range, so that the fine canopy base height prediction in the larger area range can be realized, and the accuracy of the canopy base height prediction in the area range is improved.
Further, the canopy base height prediction model is obtained through the following training mode:
acquiring a satellite optical data sample, a topographic data sample and a meteorological data sample of a forest total area sample;
Performing feature extraction based on the satellite optical data sample, the topographic data sample and the meteorological data sample to obtain a plurality of forest feature samples;
determining all height cluster samples corresponding to the forest area samples respectively and canopy base height labels corresponding to the forest area samples respectively; the forest area sample is a preset area of the forest total area sample;
Model training is carried out based on all the height cluster samples corresponding to each forest feature sample, all the height cluster samples corresponding to each forest region sample and all the canopy base height labels corresponding to each forest region sample, and a canopy base height prediction model is obtained.
Specifically, the canopy base height prediction device acquires a satellite optical data sample, a terrain data sample, a meteorological data sample, a forest height sample and a land cover data sample of a forest total area sample.
It should be noted that, satellite optical data is mainly obtained by an optical remote sensing satellite, and is used for obtaining images and spectrum information of the earth surface. The terrain data typically comes from radar altimeter satellites or lidar satellites that may measure elevation, slope, etc. of the earth's surface. Meteorological data is generally obtained by meteorological satellites for monitoring air temperature, precipitation, wind speed, etc.
Further, the canopy base height prediction device performs feature extraction on the satellite optical data sample to obtain a spectrum feature sample, a vegetation index feature sample and a texture feature sample, further, the canopy base height prediction device performs feature extraction on the terrain data sample to obtain a terrain feature sample, further, the canopy base height prediction device performs feature extraction on the meteorological data sample to obtain a meteorological feature sample, further, the canopy base height prediction device performs feature extraction on the forest height sample to obtain a forest height feature sample, and further, the canopy base height prediction device performs feature extraction on the land coverage data sample to obtain a land coverage feature sample. Thus, the spectral feature sample, the vegetation index feature sample, the texture feature sample, the topography feature sample, the meteorological feature sample, the forest height feature sample, and the land cover feature sample are all collectively referred to as a forest feature sample.
In one embodiment, as shown in table 1, table 1 is a forest feature sample table. The spectrum characteristic sample comprises 12 wave bands of the optical remote sensing satellite, the vegetation index characteristic sample comprises 17 vegetation indexes calculated by the optical remote sensing satellite, the texture characteristic sample comprises 9 texture characteristics calculated by a gray level co-occurrence matrix, a kanus Conners and the like for each of NIR, NDVI, BI parameters, and a calculation window is 3 multiplied by 3. The spectrum characteristic sample, the vegetation index characteristic sample and the texture characteristic sample are synthesized according to the median value of every two months in the whole year, and the total period is 6. The terrain feature samples include elevation, grade, and slope. The weather feature samples were synthesized from the mean of nearly five years and included 2 meters air temperature, total annual precipitation, mean sea level air pressure, surface air pressure, 10 meters u wind component, and 10 meters v wind component. The forest height characteristic sample adopts forest height distribution data, which is the vertical distance from the ground to the top of the crown. The land cover characteristic sample is GLC_FCS30-2020, and GLC_FCS30-2020 is a global land cover 30 m resolution product, and can provide land cover classification information with 30 m resolution in the global scope.
Table 1 forest characteristic sample table
Further, the canopy base height prediction device determines all the height cluster samples corresponding to the plurality of forest area samples respectively and canopy base height labels corresponding to the plurality of forest area samples respectively, wherein the forest area sample is a preset area of the total forest area sample.
It should be noted that, the forest area sample is actually a satellite-borne laser radar data point, and a plurality of satellite-borne laser radar data points are adopted, and quality problems may exist in part in the acquisition process, so that quality control is required to be performed on the forest area sample, and the satellite-borne laser radar data points with good waveform load states based on energy, sensitivity, amplitude and real-time surface tracking quality standards and positioning information are reserved. Thus, the plurality of forest area samples are actually the retained satellite-borne lidar data points after quality control.
Further, the canopy base height prediction device performs model training based on all the height clustering cluster samples corresponding to each forest feature sample and each forest region sample and the canopy base height labels corresponding to each forest region sample to obtain a canopy base height prediction model.
In one embodiment, the model training process may be trained using gradient-lifted trees. Gradient lifting trees are a powerful machine learning model for solving regression and classification problems. The method is an integrated learning method, and gradually improves model performance by combining a plurality of decision tree models, so that the method is excellent in various tasks. The gradient lifting tree model calculation mainly comprises the steps of initializing a model, calculating residual errors, training a new tree, updating the model and repeating iteration until a certain stopping condition is met, such as the number of iterations reaches a preset value or the residual errors are small enough. The main advantages of gradient-lifted trees include: anti-overfitting, automatic handling of different types of features, robustness.
According to the embodiment of the invention, the forest feature samples provided by different satellites in the forest total area sample are extracted in a feature extraction mode, all the highly clustered samples corresponding to the forest area samples respectively and the canopy base height labels corresponding to the forest area samples respectively are further determined, and then model training is carried out by combining the forest feature samples and the canopy base height labels, so that various environmental factors of the forest area are integrated, and complex features of the forest area and the variation rule of the canopy base height are better captured.
Further, when determining all the height cluster samples corresponding to the plurality of forest area samples and the canopy base height labels corresponding to the plurality of forest area samples, the following steps are executed for each forest area sample:
Acquiring a plurality of echo data samples of a current forest area sample; the echo data samples include echo energy samples and height information samples;
Performing cluster analysis on each height information sample to obtain a plurality of height cluster samples at different tree parts; the plurality of high-level cluster samples at least comprise crown cluster samples and ground cluster samples;
and carrying out canopy base height calculation based on each high clustering sample and/or each echo data sample to obtain a canopy base height label of the current forest area sample.
Specifically, the canopy base height prediction device acquires satellite-borne laser radar data of a current forest area sample, and determines a plurality of echo energy samples of the current forest area sample according to the satellite-borne laser radar data.
Further, the canopy base height prediction device respectively performs height value matching on each echo energy sample to obtain a plurality of height information samples.
Further, the canopy base height prediction device performs cluster analysis on each height information sample to obtain a plurality of height cluster samples at different tree positions, wherein the height cluster samples at least comprise crown cluster samples and ground cluster samples.
It should be noted that, in the model training process, the clustering mode also uses the K-Means algorithm, MEAN SHIFT algorithm and DBSCAN algorithm. And part of data is randomly selected for algorithm test, and according to a test result, a cluster mode which is the best in terms of vertical height clustering of the forest height information samples is MEAN SHIFT algorithm, so that crown cluster samples and ground cluster samples can be effectively clustered.
It should be further noted that crowns typically have complex structures, including leaves and branches, which are capable of causing reflections of radar waves, thereby producing a significant signal in the echo signal. The ground can reflect under the action of radar waves, different earth surface types (such as soil, water body and the like) have different echo characteristics, and the reflection of radar signals can also cause obvious change of echo energy. The trunk is generally relatively slender and less dense than the crown and ground, and its structure is less reflective of radar waves and therefore contributes less to the echo signal. Thus, the echo signals reflect mainly both the crown and the ground, that is to say the echo energy samples and the height information samples reflect mainly both the crown and the ground. Therefore, the highly clustered samples obtained by the cluster analysis are crown clustered samples and ground clustered samples.
Further, the canopy base height prediction device performs canopy base height calculation based on each high clustering cluster sample and/or each echo data sample to obtain a canopy base height label of the current forest area sample.
According to the embodiment of the invention, the height information samples of the forest region samples are subjected to cluster analysis to obtain a plurality of height cluster samples at different tree parts, and the canopy base height calculation is further performed based on each height cluster sample and/or each echo data sample to obtain the canopy base height label of the forest region samples, so that accurate canopy base height calculation aiming at the tree structure can be realized, further, the model generalization capability can be higher as model training data, and the model prediction accuracy is improved.
Further, performing canopy base height calculation based on each highly clustered sample to obtain a canopy base height label of the current forest area sample, including:
determining a high cluster threshold between the crown cluster sample and the ground cluster sample;
and determining the height information sample corresponding to the height cluster threshold as a canopy base height label of the current forest area sample.
Specifically, the canopy base height prediction device determines a high cluster threshold between the crown cluster samples and the ground cluster samples.
Further, the canopy base height prediction device determines a height information sample corresponding to the height clustering threshold value as a canopy base height label of the current forest area sample.
It should be noted that, as shown in fig. 2, fig. 2 is a schematic diagram of the high-level cluster provided by the present invention, circles in fig. 2 represent sample data in a ground cluster sample, triangles represent sample data in a crown cluster sample, lines "—" represent crown cluster centers, lines "- -" represent ground cluster centers, and lines "- - - -" represent high-level cluster thresholds. The location of the high cluster threshold may be considered as the intermediate position of the crown and the ground, i.e. the threshold that distinguishes the crown cluster sample and the ground cluster sample into different high clusters. And because the grass and shrubs on the surface are clustered into ground types, the position of the high clustering threshold value is positioned at the position of the canopy base height, and therefore the height information sample corresponding to the high clustering threshold value can be determined as the canopy base height. This way of determining the canopy-based high label can be considered a highly clustered thresholding method.
According to the embodiment of the invention, the crown base height is determined by determining the height cluster threshold value between the crown cluster sample and the ground cluster sample and determining the height information sample corresponding to the height cluster threshold value as the crown base height label of the forest area sample, so that the crown and ground two parts based on the satellite-borne laser radar data point are effectively distinguished, and the crown base height is determined, and therefore, the fine prediction of the crown base height in a larger area range is realized.
Further, performing canopy base height calculation based on each echo data sample to obtain a canopy base height label of the current forest area sample, including:
constructing a piecewise function between the echo energy samples and the height information samples based on the echo energy samples and the height information samples;
performing first-order derivation on the piecewise function to obtain a derivative of the piecewise function;
Constructing a graph according to the derivative of the piecewise function;
And determining the height information sample corresponding to the jump position in the graph as a canopy base height label of the current forest area sample.
Specifically, the canopy-based height prediction device constructs a piecewise function between the echo energy samples and the height information samples based on each echo energy sample and each height information sample.
Further, the canopy base height prediction device conducts first-order derivation on the piecewise function to obtain the derivative of the piecewise function.
Further, the canopy base height prediction device constructs a graph according to the derivative of the piecewise function.
Further, the canopy base height prediction device determines a height information sample corresponding to the jump position in the graph as a canopy base height label of the current forest area sample.
It should be noted that, as shown in fig. 3, fig. 3 is a first derivative graph provided by the present invention, the line "—" in fig. 3 represents a piecewise function graph, the line "-" represents a first derivative graph, the line "-" characterizes the point of maximum of the first derivative, i.e., the jump position in the graph of the first derivative, as the position where the piecewise function changes most rapidly. Because the vegetation at the trunk is sparse, the vegetation at the crown is denser, and for equal height intervals, the laser radar instrument obtains less echo energy from the trunk and more echo energy from the crown. Therefore, a piecewise function between the echo energy sample and the height information sample is constructed, the piecewise function is subjected to first-order derivation to obtain a derivative of the piecewise function, and the position of the fastest change of the height information sample corresponding to the echo energy sample can be determined, wherein the position is the position of the tree trunk converted into the tree crown, so that the crown base height label is determined. This way of determining the cap base height label can be considered as a first derivative method.
According to the embodiment of the invention, the first derivative is obtained by constructing the piecewise function between the echo energy sample and the height information sample, the graph is constructed according to the first derivative, the canopy base height label of the forest area sample is determined through the height information sample corresponding to the jump position in the graph, the position of the canopy base is effectively identified based on the maximum change position of the height value of the satellite-borne laser radar data point, and the canopy base height is determined, so that fine canopy base height prediction in a larger area range is realized.
Further, performing canopy base height calculation based on each highly clustered sample and each echo data sample to obtain a canopy base height label of the current forest area sample, including:
determining the maximum echo energy sample as the total height of the tree;
Determining a crown cluster center of the crown cluster sample;
Determining a height information sample corresponding to the crown cluster center as the crown center height;
and calculating the canopy base height based on the total height of the tree and the central height of the crown, and obtaining a canopy base height label of the current forest area sample.
Specifically, the canopy base height prediction device determines the maximum echo energy sample as the total height of the tree, wherein the total height of the tree is the height value from the top end of the tree to the ground.
It should be noted that in the on-board lidar data, the maximum peak of the reflected echo generally corresponds to the position of the crown or the crown of the tree, because the reflection area encountered at these places is the largest.
Further, the canopy base height prediction device determines a crown cluster center of the crown cluster sample, and further, the canopy base height prediction device determines a height information sample corresponding to the crown cluster center of the crown cluster sample as a crown center height, wherein the crown center height is a height value from the crown center position to the ground.
Further, the canopy base height prediction device calculates the canopy base height based on the total height of the tree and the central height of the tree crown, and a canopy base height label of the current forest area sample is obtained.
According to the embodiment of the invention, the maximum echo energy sample is determined as the total height of the tree, the height information sample corresponding to the crown cluster center of the crown cluster sample is determined as the crown center height, and further, the crown base height calculation is carried out based on the total height of the tree and the crown center height to obtain the crown base height label of the forest area sample, so that the crown base height of the satellite-borne laser radar data point is logically determined according to the formula calculation of the tree structure, and fine crown base height prediction in a larger area range can be realized.
Further, the calculating the canopy base height based on the total height of the tree and the central height of the crown to obtain a canopy base height tag of the current forest area sample comprises:
Calculating the difference between the total height of the tree and the central height of the crown to obtain a first difference;
performing double product calculation on the first difference value to obtain a product value;
Calculating the difference value between the total height of the tree and the product value to obtain a second difference value;
and determining the second difference value as a canopy base high label of the current forest area sample.
Specifically, the canopy base height prediction device calculates the difference between the total height of the tree and the central height of the crown to obtain a first difference.
Further, the canopy base height prediction device performs double product calculation on the first difference value to obtain a product value.
Further, the canopy base height prediction device calculates the difference between the total height of the tree and the product value to obtain a second difference.
Further, the canopy base height prediction device determines the second difference value as a canopy base height label of the current forest area sample.
It should be noted that, the calculation mode of the canopy base height can be expressed as the total height of the tree minus the length of the canopy. The central position of the crown cluster sample obtained by clustering can be expressed as the crown central height, the height of the crown center from the top of the tree can be obtained by subtracting the crown central height from the total height of the tree, the crown length can be calculated by multiplying multiples, and the crown base height can be calculated. This way of determining the cap base height label can be considered a formula calculation.
In one embodiment, the calculation of the canopy-based high tag is as follows:
；
Wherein CBH represents the crown base height label, CH represents the total height of the tree, and CC represents the crown center height.
According to the embodiment of the invention, formula calculation is carried out according to the tree structure, and the canopy base height of the satellite-borne laser radar data point is logically determined, so that fine canopy base height prediction in a larger area range is realized.
The following description is one verification example of a canopy base height prediction method provided by an embodiment of the present invention.
In the prior art, a ground measurement mode and an unmanned aerial vehicle prediction mode are adopted for predicting the canopy base height. For the ground measurement mode, a researcher performs manual measurement through various measuring instruments. For the unmanned aerial vehicle prediction mode, the following steps are executed:
And for the collected unmanned plane laser radar data, carrying out data reconstruction by using three-dimensional reconstruction software, ignoring laser points which are 250 meters away from the laser radar sensor, and reconstructing to obtain laser radar point clouds of a research area. And carrying out point cloud classification on the obtained laser radar point cloud through ENVI LIDAR software to obtain a ground point cloud, generating a digital elevation model through the ground point cloud, normalizing the original point cloud through Python software by utilizing the digital elevation model, and obtaining the research area point cloud data corrected by the topography. And drawing the range of the designated calculation region by using ArcGIS software, and storing the range as a Shapefile format file. And cutting the point cloud data corrected by the terrain by using python software to obtain laser radar point cloud data of a designated calculation area, and using the laser radar point cloud data as canopy base height calculation.
For the preprocessed lidar data, the canopy base height was calculated by two algorithms using python software:
① Point cloud layering algorithm
The main ideas of the point cloud layering algorithm are that the densities of different layers of the forest are different, the vegetation density of the crown position is high, and the trunk position is sparse, so that the position where the density of the point cloud is changed from sparse to dense can be regarded as the position where the crown base height is located. Based on the idea, the python software is utilized to layer the point cloud data of the calculation area, and the point number of the point cloud of each layer is calculated according to 0.5 meter as one layer. And comparing the number of the point clouds of each layer with 1% of the number of the point clouds of more than 2 meters, if the number of the point clouds is smaller than the number of the point clouds of each layer, assigning 0, and if the number of the point clouds is larger than the number of the point clouds of each layer, assigning 1. And (3) converting the first position of the base height from a sparse position to a dense position from 0 to 1, and obtaining the crown base height.
② Point cloud height derivation algorithm
The main idea of the point cloud height derivation algorithm is that the point cloud heights of the research area point clouds with increased percentile values are different, namely the height difference corresponding to the percentile of the trunk position is obviously higher than the height difference corresponding to the percentile of the tree crown position with equal interval. And deriving a height list corresponding to every 5 percentiles of the point cloud by using python software, and calculating a maximum value of the first derivative, namely, a position of the maximum value of the jump of the height corresponding to the percentile, which is used as the position of the canopy base height, so that the canopy base height can be obtained.
Further, for the accuracy verification of the canopy base height prediction model, the verification of ground measurement and unmanned aerial vehicle prediction results, the verification of unmanned aerial vehicle and satellite-borne laser radar data point prediction canopy base height and the verification of the base height results of area scale calculation by using the satellite-borne laser radar are respectively carried out, and the verification indexes adopt R 2 and EA, MAE, RMSE, RMSE%. Wherein, R 2, EA and RMSE% are relative error indexes, and MAE and RMSE are absolute error indexes. The formula is as follows:
/>
wherein n is the number of sample points, Predicted value output for model,/>Is the true value of the sample point,/>Is the average of the true values of the sample points.
1) Unmanned aerial vehicle estimation accuracy
The ground measurement method is accurate, but because of subjectivity of different measurement personnel, consistency is not high, and the connection measurement of the surface scale cannot be realized. And calculating the canopy base height by using the ground measurement data of the research area, cutting unmanned plane laser radar data at the position corresponding to the ground sampling point, and comparing the calculated canopy base height with the ground measurement result by using a point cloud layering algorithm and a point cloud height deriving algorithm. The results show that the crown base height obtained by the point cloud layering algorithm is measured with the ground by r2=0.69, mae=1.08 meters, rmse=1.51 meters, RMSE% =29.46%. The crown layer base height obtained by the point cloud height derivative algorithm and the ground measured r2=0.44, mae=1.57 meters, rmse=2.02 meters, and RMSE% = 39.60%.
Both algorithms can realize calculation of canopy base height based on data acquired by the unmanned aerial vehicle, and a layered calculation method obtains higher estimation precision, and both methods realize spatially continuous canopy base height measurement. In the subsequent accuracy verification, a point cloud layering algorithm is adopted to calculate the canopy base height of the research area.
2) Crown base height of satellite-borne laser radar data point
The problem of discontinuous measurement space of the canopy base height is solved based on the canopy base height prediction of the unmanned aerial vehicle, but the method is limited by the acquisition range of the unmanned aerial vehicle data, and large-scale data acquisition is difficult to realize, so that the acquisition of regional scale canopy base height sample points based on the satellite-borne laser radar data points is adopted.
The crown base height of the overlapping point of the research area and the unmanned aerial vehicle laser radar data is calculated through a first derivative method, a high clustering threshold method and a formula calculation method, and is compared with the crown base height calculated by the unmanned aerial vehicle, and the result shows that the estimation accuracy of the crown base height obtained by the high clustering threshold method is optimal.
3) Crown base high model prediction accuracy
And carrying out canopy base height prediction on the satellite-borne laser radar data points acquired in the research area by adopting a high clustering threshold method, and acquiring 7836 canopy base height samples of the satellite-borne laser radar data points in the research area. Based on the obtained sample data, carrying out gradient lifting tree regression analysis by combining 348 features of spectrum, vegetation index, texture, terrain, weather, forest height and land coverage subjected to data pretreatment. Through the feature contribution evaluation, the first 100 features are selected as feature inputs. Training 90% of 7836 samples obtained by using satellite-borne laser radar data points, optimizing model parameters of the gradient lifting tree, training a model by using 490 trees, verifying the obtained model result and the rest 10% samples, and finally obtaining a high verification result accuracy.
Finally, using the trained model, a canopy base height distribution map of the study area can be obtained, wherein the canopy base height distribution map shows the canopy base height in each 25 m spatial resolution area.
Furthermore, the invention also provides a device for predicting the base height of the canopy.
Referring to fig. 4, fig. 4 is a schematic structural diagram of a canopy-based height prediction device provided by the present invention.
The canopy base height prediction device comprises:
an acquisition module 410, configured to acquire satellite-borne laser radar data of a forest area to be detected;
A determining module 420, configured to determine a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
a matching module 430, configured to match the height values of the echo energies respectively, so as to obtain a plurality of height information;
The cluster analysis module 440 is configured to perform cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
The canopy base height prediction module 450 is configured to input the plurality of height clusters into a canopy base height prediction model, and obtain a canopy base height predicted value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
According to the canopy base height prediction device provided by the invention, the plurality of echo energies of the forest area to be detected are obtained, the corresponding height information is matched based on each echo energy, the plurality of height information is further subjected to clustering analysis to obtain the plurality of height clustering clusters at different tree positions, the model can be convenient to determine the tree structure, all the height clustering clusters are further input into the canopy base height prediction model, the canopy base height prediction value of the forest area to be detected, which is output by the model, is obtained, the automation level of the canopy base height prediction mode is realized, the time and economic cost consumed by the canopy base height prediction are effectively reduced, the satellite-borne laser radar data are observed through the laser radar, and the satellite-borne radar has a larger observation range, so that the fine canopy base height prediction in the larger area range can be realized, and the accuracy of the canopy base height prediction in the area range is improved.
Further, the canopy base height prediction device is further configured to:
acquiring a satellite optical data sample, a topographic data sample and a meteorological data sample of a forest total area sample;
Performing feature extraction based on the satellite optical data sample, the topographic data sample and the meteorological data sample to obtain a plurality of forest feature samples;
determining all height cluster samples corresponding to the forest area samples respectively and canopy base height labels corresponding to the forest area samples respectively; the forest area sample is a preset area of the forest total area sample;
Model training is carried out based on all the height cluster samples corresponding to each forest feature sample, all the height cluster samples corresponding to each forest region sample and all the canopy base height labels corresponding to each forest region sample, and a canopy base height prediction model is obtained.
Further, the canopy base height prediction device is further configured to:
Acquiring a plurality of echo data samples of a current forest area sample; the echo data samples include echo energy samples and height information samples;
Performing cluster analysis on each height information sample to obtain a plurality of height cluster samples at different tree parts; the plurality of high-level cluster samples at least comprise crown cluster samples and ground cluster samples;
and carrying out canopy base height calculation based on each high clustering sample and/or each echo data sample to obtain a canopy base height label of the current forest area sample.
Further, the canopy base height prediction device is further configured to:
determining a high cluster threshold between the crown cluster sample and the ground cluster sample;
and determining the height information sample corresponding to the height cluster threshold as a canopy base height label of the current forest area sample.
Further, the canopy base height prediction device is further configured to:
constructing a piecewise function between the echo energy samples and the height information samples based on the echo energy samples and the height information samples;
performing first-order derivation on the piecewise function to obtain a derivative of the piecewise function;
Constructing a graph according to the derivative of the piecewise function;
And determining the height information sample corresponding to the jump position in the graph as a canopy base height label of the current forest area sample.
Further, the canopy base height prediction device is further configured to:
determining the maximum echo energy sample as the total height of the tree;
Determining a crown cluster center of the crown cluster sample;
Determining a height information sample corresponding to the crown cluster center as the crown center height;
and calculating the canopy base height based on the total height of the tree and the central height of the crown, and obtaining a canopy base height label of the current forest area sample.
Further, the canopy base height prediction device is further configured to:
Calculating the difference between the total height of the tree and the central height of the crown to obtain a first difference;
performing double product calculation on the first difference value to obtain a product value;
Calculating the difference value between the total height of the tree and the product value to obtain a second difference value;
and determining the second difference value as a canopy base high label of the current forest area sample.
It should be noted that, when the canopy base height prediction device provided by the present invention specifically operates, the canopy base height prediction method described in any one of the above embodiments may be executed, which is not described in detail in this embodiment.
Fig. 5 is a schematic structural diagram of an electronic device according to the present invention, and as shown in fig. 5, the electronic device may include: processor 510, communication interface (Communications Interface) 520, memory 530, and communication bus 540, wherein processor 510, communication interface 520, memory 530 complete communication with each other through communication bus 540. Processor 510 may invoke logic instructions in memory 530 to perform a canopy base height prediction method comprising: acquiring satellite-borne laser radar data of a forest area to be detected; determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data; respectively carrying out height value matching on each echo energy to obtain a plurality of height information; performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions; inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height prediction value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
Further, the logic instructions in the memory 530 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, are capable of performing the canopy base height prediction method provided by the above embodiments, the method comprising: acquiring satellite-borne laser radar data of a forest area to be detected; determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data; respectively carrying out height value matching on each echo energy to obtain a plurality of height information; performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions; inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height prediction value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
In yet another aspect, the present invention further provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the canopy base height prediction method provided by the above embodiments, the method comprising: acquiring satellite-borne laser radar data of a forest area to be detected; determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data; respectively carrying out height value matching on each echo energy to obtain a plurality of height information; performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions; inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height prediction value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A canopy base height prediction method, comprising:
Acquiring satellite-borne laser radar data of a forest area to be detected;
Determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
respectively carrying out height value matching on each echo energy to obtain a plurality of height information;
performing cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
Inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height prediction value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
2. The canopy base height prediction method according to claim 1, wherein the canopy base height prediction model is obtained by training as follows:
acquiring a satellite optical data sample, a topographic data sample and a meteorological data sample of a forest total area sample;
Performing feature extraction based on the satellite optical data sample, the topographic data sample and the meteorological data sample to obtain a plurality of forest feature samples;
determining all height cluster samples corresponding to the forest area samples respectively and canopy base height labels corresponding to the forest area samples respectively; the forest area sample is a preset area of the forest total area sample;
Model training is carried out based on all the height cluster samples corresponding to each forest feature sample, all the height cluster samples corresponding to each forest region sample and all the canopy base height labels corresponding to each forest region sample, and a canopy base height prediction model is obtained.
3. The canopy base height prediction method according to claim 2, wherein when determining all the highly clustered samples corresponding to the plurality of forest area samples respectively and the canopy base height labels corresponding to the plurality of forest area samples respectively, the following steps are performed for each forest area sample:
Acquiring a plurality of echo data samples of a current forest area sample; the echo data samples include echo energy samples and height information samples;
Performing cluster analysis on each height information sample to obtain a plurality of height cluster samples at different tree parts; the plurality of high-level cluster samples at least comprise crown cluster samples and ground cluster samples;
and carrying out canopy base height calculation based on each high clustering sample and/or each echo data sample to obtain a canopy base height label of the current forest area sample.
4. The canopy base height prediction method of claim 3, wherein the canopy base height calculation is performed based on each highly clustered sample to obtain a canopy base height label of the current forest area sample, comprising:
determining a high cluster threshold between the crown cluster sample and the ground cluster sample;
and determining the height information sample corresponding to the height cluster threshold as a canopy base height label of the current forest area sample.
5. The canopy base height prediction method according to claim 3, wherein the canopy base height calculation is performed based on each echo data sample to obtain a canopy base height label of a current forest area sample, comprising:
constructing a piecewise function between the echo energy samples and the height information samples based on the echo energy samples and the height information samples;
performing first-order derivation on the piecewise function to obtain a derivative of the piecewise function;
Constructing a graph according to the derivative of the piecewise function;
And determining the height information sample corresponding to the jump position in the graph as a canopy base height label of the current forest area sample.
6. The canopy base height prediction method according to claim 3, wherein the performing canopy base height calculation based on each highly clustered sample and each echo data sample to obtain a canopy base height label of the current forest area sample comprises:
determining the maximum echo energy sample as the total height of the tree;
Determining a crown cluster center of the crown cluster sample;
Determining a height information sample corresponding to the crown cluster center as the crown center height;
and calculating the canopy base height based on the total height of the tree and the central height of the crown, and obtaining a canopy base height label of the current forest area sample.
7. The canopy base height prediction method according to claim 6, wherein the performing canopy base height calculation based on the total height of the tree and the crown center height to obtain a canopy base height label of the current forest area sample comprises:
Calculating the difference between the total height of the tree and the central height of the crown to obtain a first difference;
performing double product calculation on the first difference value to obtain a product value;
Calculating the difference value between the total height of the tree and the product value to obtain a second difference value;
and determining the second difference value as a canopy base high label of the current forest area sample.
8. A canopy base height prediction device, comprising:
The acquisition module is used for acquiring satellite-borne laser radar data of the forest area to be detected;
The determining module is used for determining a plurality of echo energies of the forest area to be detected according to the satellite-borne laser radar data;
The matching module is used for respectively carrying out height value matching on the echo energy to obtain a plurality of height information;
The cluster analysis module is used for carrying out cluster analysis on the plurality of height information to obtain a plurality of height clusters at different tree positions;
The canopy base height prediction module is used for inputting the plurality of height clusters into a canopy base height prediction model to obtain a canopy base height predicted value of the forest area to be detected, which is output by the canopy base height prediction model; the canopy base height prediction model is obtained by model training based on the high clustering cluster sample and the canopy base height label corresponding to the high clustering cluster sample.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the canopy base height prediction method of any one of claims 1 to 7 when the computer program is executed by the processor.
10. A storage medium comprising a non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the canopy base height prediction method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410504507.4A CN118094397B (en) | 2024-04-25 | Crown base height prediction method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410504507.4A CN118094397B (en) | 2024-04-25 | Crown base height prediction method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118094397A true CN118094397A (en) | 2024-05-28 |
| CN118094397B CN118094397B (en) | 2024-07-30 |
Family
ID=
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110149267A1 (en) * | 2009-12-22 | 2011-06-23 | Weyerhaeuser Nr Company | METHOD AND APPARATUS FOR ANALYZING TREE CANOPIES WITH LiDAR DATA |
| CN112729130A (en) * | 2020-12-29 | 2021-04-30 | 四川天奥空天信息技术有限公司 | Method for measuring height of tree canopy by satellite remote sensing |
| CN115561773A (en) * | 2022-12-02 | 2023-01-03 | 武汉大学 | Forest carbon reserve inversion method based on ICESat-2 satellite-borne LiDAR data and multispectral data |
| US20230039554A1 (en) * | 2021-08-09 | 2023-02-09 | Institute of Forest Resource Information Techniques CAF | Tree crown extraction method based on unmanned aerial vehicle multi-source remote sensing |
| CN117075138A (en) * | 2023-08-18 | 2023-11-17 | 中南大学 | Remote sensing measurement and calculation method, system and medium for canopy height of 30-meter forest in area |
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110149267A1 (en) * | 2009-12-22 | 2011-06-23 | Weyerhaeuser Nr Company | METHOD AND APPARATUS FOR ANALYZING TREE CANOPIES WITH LiDAR DATA |
| CN112729130A (en) * | 2020-12-29 | 2021-04-30 | 四川天奥空天信息技术有限公司 | Method for measuring height of tree canopy by satellite remote sensing |
| US20230039554A1 (en) * | 2021-08-09 | 2023-02-09 | Institute of Forest Resource Information Techniques CAF | Tree crown extraction method based on unmanned aerial vehicle multi-source remote sensing |
| CN115561773A (en) * | 2022-12-02 | 2023-01-03 | 武汉大学 | Forest carbon reserve inversion method based on ICESat-2 satellite-borne LiDAR data and multispectral data |
| CN117075138A (en) * | 2023-08-18 | 2023-11-17 | 中南大学 | Remote sensing measurement and calculation method, system and medium for canopy height of 30-meter forest in area |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | Lidar boosts 3D ecological observations and modelings: A review and perspective | |
| CN108921885B (en) | Method for jointly inverting forest aboveground biomass by integrating three types of data sources | |
| McRoberts et al. | Using remotely sensed data to construct and assess forest attribute maps and related spatial products | |
| Wulder et al. | The role of LiDAR in sustainable forest management | |
| Sumnall et al. | Comparison of small-footprint discrete return and full waveform airborne lidar data for estimating multiple forest variables | |
| Hosoi et al. | 3-D voxel-based solid modeling of a broad-leaved tree for accurate volume estimation using portable scanning lidar | |
| Næsset et al. | Laser scanning of forest resources: the Nordic experience | |
| Halme et al. | Utility of hyperspectral compared to multispectral remote sensing data in estimating forest biomass and structure variables in Finnish boreal forest | |
| Ørka et al. | Effects of different sensors and leaf-on and leaf-off canopy conditions on echo distributions and individual tree properties derived from airborne laser scanning | |
| Chauve et al. | Advanced full-waveform lidar data echo detection: Assessing quality of derived terrain and tree height models in an alpine coniferous forest | |
| Vauhkonen | Estimating crown base height for Scots pine by means of the 3D geometry of airborne laser scanning data | |
| Meyer et al. | Canopy area of large trees explains aboveground biomass variations across neotropical forest landscapes | |
| CN109031344B (en) | Method for jointly inverting forest structure parameters by full-waveform laser radar and hyperspectral data | |
| Stone et al. | Determining an optimal model for processing lidar data at the plot level: results for a Pinus radiata plantation in New South Wales, Australia | |
| CN108981616B (en) | Method for inverting effective leaf area index of artificial forest by unmanned aerial vehicle laser radar | |
| CN113204998A (en) | Airborne point cloud forest ecological estimation method and system based on single wood scale | |
| CN112861435B (en) | Mangrove quality remote sensing inversion method and intelligent terminal | |
| Pirotti et al. | A comparison of tree segmentation methods using very high density airborne laser scanner data | |
| CN109061601A (en) | A method of based on unmanned plane laser radar inverting artificial forest forest structural variable | |
| CN115700370A (en) | Carbon reserve calculation method, carbon reserve calculation device, electronic device, and storage medium | |
| CN115561181A (en) | Water quality inversion method based on multispectral data of unmanned aerial vehicle | |
| Wang et al. | Hybrid model for estimating forest canopy heights using fused multimodal spaceborne LiDAR data and optical imagery | |
| Chen et al. | An integrated GIS tool for automatic forest inventory estimates of Pinus radiata from LiDAR data | |
| CN113534083B (en) | SAR-based corn stubble mode identification method, device and medium | |
| Holmgren et al. | Tree crown segmentation in three dimensions using density models derived from airborne laser scanning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |