CN107705344A - Plant canopy model extracting method in laser scanning environment cloud data - Google Patents
Plant canopy model extracting method in laser scanning environment cloud data Download PDFInfo
- Publication number
- CN107705344A CN107705344A CN201710891872.5A CN201710891872A CN107705344A CN 107705344 A CN107705344 A CN 107705344A CN 201710891872 A CN201710891872 A CN 201710891872A CN 107705344 A CN107705344 A CN 107705344A
- Authority
- CN
- China
- Prior art keywords
- tree crown
- point
- crown
- tree
- radius
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/20—Drawing from basic elements, e.g. lines or circles
- G06T11/203—Drawing of straight lines or curves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Geometry (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses plant canopy model extracting method in a kind of laser scanning environment cloud data, this method is by tree crown classification model construction, point cloud rasterizing, grid filtering, tree crown window calculation, profile are delineated and result statistic of classification forms.When carrying out tree crown type classification with extraction using LiDAR systems, due to characteristics such as the property at random of laser point cloud and big data quantities, need to extract selective vegetation tree crown feature for big data, specific aim filtering process is carried out to tree crown scan data, the operation needs first to carry out rasterizing processing to cloud data, then to raster data carry out tree crown dynamic window calculating, obtain inclined degree, by corner space relation calculate suitable crown outline delineate parameter complete crown outline delineate.The significantly common size vegetation target identification of independent characteristic during being navigated by water available for scientific investigation ship, perceived for scientific investigation ship information system environment and guide for method is provided.
Description
Technical field
The invention belongs to public affair ship acquisition of information and field of target recognition, more particularly to a kind of laser scanning environment point cloud number
According to middle plant canopy model extracting method.
Background technology
With flourishing for China's marine affairs, many types of section in urgent need of strengthening for the detectivity of islands and reefs, Near shore environment
The construction for examining ship provides numerous mobile test platforms.Ship-borne helicopter and unmanned plane are that non-the excellent of ground formula sensor is taken
Medium is carried, and based on the information in a manner of being realized as non-contact sensor by the miniaturization airborne lidar instrument of upper mounting plate
Obtain equipment.
Near shore environment is the vital area of global economic development, particularly utilizes ocean resource, exploitation offshore island in China
Under the background of reef, scientific investigation ship all has been configured with many types of airborne platform mostly, it is possible to achieve the redundant observation of table object over the ground, collection
Abundant information to carry out data acquisition and modeling to surrounding enviroment.Wherein, offshore regional vegetation is important environmental information, its
It is significant for a variety of applications such as terrain environment exploration, Investigation Forestry Resources, carbon remittance calculating, the protections of tide littoral zone.
And the laser acquisition based on boat-carrying airborne platform is relative to the conventional detection methods such as aerial print, satellite image, radar, its product
Laser scanning point cloud is abundant in content true three-dimensional data resource, includes the object geometry for obtaining surface vegetation information and spectrum letter
Breath, these are the important evidences for carrying out subsequent environments modeling, situation judgement and decision-making.Therefore swept, it is necessary to study from airborne laser
The modeling method of different types vegetation tree crown is extracted in described point cloud.
Airborne lidar technology (Light Detection And Ranging, LiDAR) is a kind of contactless space
Information remote sensing active probe acquiring technology means.It is by launching and receiving laser scanning Shu Shixian to object in certain distance
Measurement and sensing work.Its product form mainly has two kinds of discrete type and continuous type.Discrete type laser point cloud passes through to environment pair
The multiecho of elephant is handled, and obtains the strength information of true three-dimensional information and the part surface characteristic of specific object of observation.
LiDAR technologies have the characteristic of multiecho for transparent object, can be formed to object of observation by multiecho
Internal structural information, so as to provide abundant data for comprehensive detection and target identification.
The content of the invention
The technical problem to be solved by the invention is to provide plant canopy model in a kind of laser scanning environment cloud data to extract
Method, the significantly common size of independent characteristic during being navigated by water available for scientific investigation ship (crown canopy covering radius is about 0.5 meter -3 meters)
Vegetation target identification, perceived for scientific investigation ship information system environment and guide for method is provided.
LiDAR systems are mainly to carry out object appearance scanning by the non-visible light laser beam of infrared band, and the wave band swashs
For light in addition to the object such as water body and minute surface, all there is diffusing reflection, refraction effect in most of scanned objects.Therefore, LiDAR is passed through
System can complete the inside and outside geometric properties identification of tree crown destination object, be that scientific investigation ship carries out extra large land environment survey, investigation etc. times
A kind of preferably target identification means of business.Main support boat-carrying airborne platform of the invention and airborne scanning device carry out tree crown object
The modeling and identification of target.
The technical solution adopted for the present invention to solve the technical problems is：There is provided in a kind of laser scanning environment cloud data
Plant canopy model extracting method, comprises the following steps, step 1, and discretization point is carried out to object of observation using airborne LiDAR systems
Cloud is scanned, and tree crown classification model construction is carried out according to typical tree crown type, and the point cloud for scanning acquisition is abstracted into gabarit space structure type
Incorporate into as taper, hemispherical, semielliptical shape, and establish the surfaces externally and internally model of counter structure；Step 2, using 0.5 meter -3 meters
Sampling interval carries out image rasterizing processing to dispersion point cloud, can be converted into being available for later stage Object identifying, spatial analysis
The bidimensional image of processing；Step 3, for different tree crown transverse direction face diameters, the adjustable filtering calculation window of selection dynamic
Tree crown plane towards tree crown coverage radius less than 3m is screened and divided；Step 4, division is screened in crown surface
After out, the radius of single tree crown is calculated according to trunk center line and tree crown section window slope；Step 5, according to trunk
The radius that center line and window calculation obtain carries out two dimensional surface tree crown circular edge and delineated；Step 6, in different type tree crown
Profile is completed after delineating, the tree crown classification quantity in statistical result on different tree crown calculation windows, to count, analyzing and the later stage builds
Mould is prepared.
By above-mentioned technical proposal, in the step 4, specially local pole is found in energy restrains calculating iterative process
Value point is compared, and tree crown point is labeled as less than the maximum point of 3 meters of radiuses to the covering of tree crown center, then using tree crown point in
The heart carries out slope calculating to periphery tree crown surface point, and variant slope can calculate its angle by vertical section, according to tree
Hat surface point is calculated to the triangle relation of trunk center line horizontal range and projected according to different plant canopy models in its maximal cover
Radius on face.
For the forest land of single seeds, the overwhelming majority has similar tree crown size, but in its natural state, vegetation tool
There are different forms, significantly common size (crown canopy covering radius is about 0.5 meter -3 meters) the common tree crown of vegetation of independent characteristic
Form is cone, semielliptical shape and hemispherical.If carrying out tree crown detection using stationary window, less window can be led
Big tree crown is caused to be divided into several small tree crowns by mistake, larger window can then cause small tree crown to merge into the big tree of individual plant by mistake
Hat.So the radius under different crown projections is calculated using variable window.Because laser point cloud data has the elevation dimension number of degrees
It is believed that breath.So window can calculates its slope near local maximum.
By above-mentioned technical proposal, in the step 5, the data in the range of 3mx3m are specially calculated using detection mask
Point, tree crown locally potential extremal region is found, reduces and calculates consumption, in each mask, determined in the way of fixed slope
Search window, data point cluster is carried out from iteration mean shift algorithm based on energy convergent function using improved, set
Comb point；If calculated, cluster centre point is consistent with the extreme point in each mask, no labeled as tree crown extreme point herein
Then the cluster point is brought into cluster calculation next time as new mask point, until cluster centre point that iteration goes out and mask pole
Value point is identical.
By above-mentioned technical proposal, the energy potential-energy function between each scanning element is defined as the potential energy sum of independent each point, respectively
Potential energy P (the x individually puti) it is xjTo xiIsotropic kernel function G (xi) and bandwidth parameter hxiForm, i.e., by xiTo xjGesture
Can be P (i, j)=S (hxi)(1-g(xi-xj)/hxi), wherein S (hxi) it is potential energy constraint factor, for controlling whole function energy
Enough monotonic increases, g (xi) it is kernel function G (xi) gradient, by xiTo xjWith xjTo xiMutual potential energy summation, be the potential energy
Functional value P (xi)=Σ (P (i, j)+P (j, i)).
Average drifting is the outstanding algorithm of printenv data clusters, is widely used in image segmentation, target following, video
The fields such as analysis.In this algorithm, when carrying out recurrence contraction to data point using average drifting, bandwidth parameter needs to gradually change
To be bonded the discrete distributivity feature of data.Especially selection selection is constrained potential energy using energy potential-energy function defined above
Be worth less point quickly to be merged, such as the smaller and homogeneous distribution scanning element of change can by evade small calculating come
Reduce iterative process.Moreover, the point smaller than maximum constrained potential energy value is typically also be distributed in object exclusive.So, Ke Yitong
The point (point of distance) of excessive potential energy promotes scan data point to gather at local maximum, avoids by the point off density cluster of surrounding
Fault Distribution is formed, meanwhile, it is also preferable for the inhibitory action of noise exception scanning element.
By above-mentioned technical proposal, in the step 1, the mathematical modeling difference of corresponding three kinds of tree crowns gabarit spatial shape
For：
Cone：Fout=R/tan (B)
Hemisphere：Fout=[R2-(x2+y2)]1/2
Semiellipsoid：Fout=e* [R2-(x2+y2)]1/2
Wherein, fout is scan model outer surface, and B is that tree crown external profile surface projects angle with vertical plane, and R is external profile surface
The circular radius of space projection on a projection plane, x are the transverse axis coordinate of scanning element on a projection plane, and y is that scanning element exists
Ordinate of orthogonal axes on projection plane, e are the ratio of long axis to short axis of spheroid.
By above-mentioned technical proposal, according to plane trigonometry relation, complete crown mapping radius R=r/ [sin (180-2*A)], r
For current preset calculation window radius；A is any scanning element on tree crown external profile surface to the angle of trunk center line.
Tree crown cluster calculation is as shown in Figure 4.So each tree according to corresponding to can calculating above formula iteration convergence
The profile radius of hat is delineated to carry out planar range.
The present invention principle be：The characteristics of discrete type point cloud, is laser scanning beam when penetrating common non-dense object,
Surface reflection, the phenomenon of internal refraction can be formed, can be observation pair after signal collection system reclaims information and handles
Multiecho data are provided as carrying out inside and outside comprehensive scanning.Its inside and outside form of different vegetation objects is different, it is therefore necessary to
Specific aim modeling is carried out for the tree crown type of different spaces form.Tree crown type classification and extraction are carried out using LiDAR systems
It is a kind of means of active remote sensing, but due to characteristics such as the property at random of laser point cloud and big data quantities.Therefore need for big number
According to selective vegetation tree crown feature is extracted, i.e., specific aim filtering process is carried out to tree crown scan data, the operation needs
Rasterizing processing first is carried out to cloud data, tree crown dynamic window calculating then is carried out to raster data, inclined degree is obtained, borrows
Help corner space relation calculate suitable crown outline delineate parameter complete crown outline delineate.
The beneficial effect comprise that：(1) take full advantage of the configured airborne platform of scientific investigation ship and scanning senses
Equipment.Airborne platform system, airborne lidar instrument of boat-carrying etc. make use of to navigate by water sensitive measuring equipment, it is perfect that seashore is planted
The Objective extraction and recognition capability of quilt.
(2) possesses the ability for different type crown identification and extraction.Analyze the tree crown of different spaces distributional pattern
Model, Local Extremum is calculated by clustering algorithm and is used as tree crown summit, then calculates tree crown window slope according to corner relation,
Obtain the conduct of tree crown radius and delineate tree crown range parameter.
(3) it is small to calculate processing complexity, is adapted to hardware-accelerated and parallelization processing, is scientific investigation beneficial to online data processing
Islands and reefs, offshore acquisition of information and the real-time processing of modeling provide support in ship pelagic environment, improve scientific investigation ship vegetation Object identifying
Ability and efficiency.
Brief description of the drawings
Below in conjunction with drawings and Examples, the invention will be further described, in accompanying drawing：
Fig. 1 is the schematic flow sheet of plant canopy model extracting method in laser scanning environment cloud data of the embodiment of the present invention；
Crown projection spatial model schematic diagram in Fig. 2 embodiment of the present invention；
Tree crown extreme value cluster flow chart in Fig. 3 embodiment of the present invention；
Tree crown radius section model schematic diagram in Fig. 4 embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, not
For limiting the present invention.
In the embodiment of the present invention, there is provided plant canopy model extracting method in a kind of laser scanning environment cloud data, including with
Lower step, step 1, the scanning of discretization point cloud is carried out to object of observation using airborne LiDAR systems, according to typical tree crown type
Tree crown classification model construction is carried out, the point cloud for scanning acquisition is abstracted into gabarit space structure type incorporates into as taper, hemispherical, semielliptical
Shape, and establish the surfaces externally and internally model of counter structure；Step 2, shadow is carried out to dispersion point cloud using 0.5 meter of -3 meters of sampling interval
As rasterizing processing, can be converted into being available for later stage Object identifying, the bidimensional image of spatial analysis processing；Step 3, pin
To different tree crown transverse direction face diameters, the adjustable filtering calculation window of selection dynamic is less than towards tree crown coverage radius
3m tree crown plane is screened and divided；Step 4, after crown surface is screened and marks off, according to trunk center line and tree
The radius of single tree crown is calculated in hat section window slope；Step 5, half obtained according to trunk center line and window calculation
Footpath carries out two dimensional surface tree crown circular edge and delineated；Step 6, after different type crown outline is completed to delineate, in statistical result
On different tree crown calculation windows tree crown classification quantity, for count, analyze and the later stage modeling prepare.
Further, in the step 4, specially searching local extremum clicks through in energy is restrained and calculates iterative process
Row compares, and tree crown point is labeled as less than the maximum point of 3 meters of radiuses to the covering of tree crown center, then to week centered on tree crown point
Side tree crown surface point carries out slope calculating, and variant slope can calculate its angle by vertical section, according to tree crown surface
Point is calculated according to different plant canopy models on its maximal cover perspective plane to the triangle relation of trunk center line horizontal range
Radius.
For the forest land of single seeds, the overwhelming majority has similar tree crown size, but in its natural state, vegetation tool
There are different forms, significantly common size (crown canopy covering radius is about 0.5 meter -3 meters) the common tree crown of vegetation of independent characteristic
Form is cone, semielliptical shape and hemispherical.If carrying out tree crown detection using stationary window, less window can be led
Big tree crown is caused to be divided into several small tree crowns by mistake, larger window can then cause small tree crown to merge into the big tree of individual plant by mistake
Hat.So the radius under different crown projections is calculated using variable window.Because laser point cloud data has the elevation dimension number of degrees
It is believed that breath.So window can calculates its slope near local maximum.
Further, in the step 5, specially in a large amount of point clouds of processing, how data point is quickly scanned
Feature point extraction is the committed step that whole Objective extraction calculates.This also contributes to the cost entirely calculated and efficiency.In order to keep away
Exempt from the calculating pixel-by-pixel to rasterizing point cloud image, the data point in the range of 3mx3m is calculated using detection mask, finds tree crown office
The potential extremal region in portion, reduce and calculate consumption, in each mask, search window is determined in the way of fixed slope, use
It is improved that data point cluster is carried out from iteration mean shift algorithm based on energy convergent function, obtain tree crown point；If calculate
Go out that cluster centre point is consistent with the extreme point in each mask, then labeled as tree crown extreme point herein, otherwise by the cluster point
Cluster calculation next time is brought into as new mask point, until the cluster centre point that iteration goes out is identical with mask extreme point.
Further, the energy potential-energy function between each scanning element is defined as the potential energy sum of independent each point, each independent point
Potential energy P (xi) it is xjTo xiIsotropic kernel function G (xi) and bandwidth parameter hxiForm, i.e., by xiTo xjPotential energy be P
(i, j)=S (hxi)(1-g(xi-xj)/hxi), wherein S (hxi) it is potential energy constraint factor, for controlling whole function dull
It is incremented by, g (xi) it is kernel function G (xi) gradient, by xiTo xjWith xjTo xiMutual potential energy summation, be this potential-energy function value
P(xi)=Σ (P (i, j)+P (j, i)).
Average drifting is the outstanding algorithm of printenv data clusters, is widely used in image segmentation, target following, video
The fields such as analysis.In this algorithm, when carrying out recurrence contraction to data point using average drifting, bandwidth parameter needs to gradually change
To be bonded the discrete distributivity feature of data.Especially selection selection is constrained potential energy using energy potential-energy function defined above
Be worth less point quickly to be merged, such as the smaller and homogeneous distribution scanning element of change can by evade small calculating come
Reduce iterative process.Moreover, the point smaller than maximum constrained potential energy value is typically also be distributed in object exclusive.So, Ke Yitong
The point (point of distance) of excessive potential energy promotes scan data point to gather at local maximum, avoids by the point off density cluster of surrounding
Fault Distribution is formed, meanwhile, it is also preferable for the inhibitory action of noise exception scanning element.
Further, in the step 1, the mathematical modeling of corresponding three kinds of tree crowns gabarit spatial shape is respectively：
Cone：Fout=R/tan (B)
Hemisphere：Fout=[R2-(x2+y2)]1/2
Semiellipsoid：Fout=e* [R2-(x2+y2)]1/2
Wherein, fout is scan model outer surface, and B is that tree crown external profile surface projects angle with vertical plane, and R is external profile surface
The circular radius of space projection on a projection plane, x are the transverse axis coordinate of scanning element on a projection plane, and y is that scanning element exists
Ordinate of orthogonal axes on projection plane, e are the ratio of long axis to short axis of spheroid.
Further, according to plane trigonometry relation, complete crown mapping radius R=r/ [sin (180-2*A)], r is current
It is default to calculate windows radius；A is any scanning element on tree crown external profile surface to the angle of trunk center line.
Tree crown cluster calculation is as shown in Figure 4.So each tree according to corresponding to can calculating above formula iteration convergence
The profile radius of hat is delineated to carry out planar range.
In presently preferred embodiments of the present invention, there is provided plant canopy model extracting method in a kind of laser scanning environment cloud data,
As shown in figure 1, calculation process delineated by tree crown classification model construction, point cloud rasterizing, grid filtering, tree crown window calculation, profile and
As a result six link compositions of statistic of classification.
Tree crown classification model construction be according to typical tree crown type, by scan obtain point cloud abstract space structure type incorporate into for
Taper, hemispherical, semielliptical shape, its ideal distribution form and actual point cloud example are as shown in Fig. 2 and establish the interior of counter structure
Appearance surface model, different type crown projection spatial model.Afterwards, rasterizing image processing is carried out for a cloud, using 0.5
The sampling interval of -3 meters of rice carries out rasterizing to dispersion point cloud, makes to be converted into being available for later stage Object identifying, spatial analysis etc. to handle
Bidimensional image.
Then, the adjustable filtering calculation window of grid filtering selection dynamic carries out extreme value towards the tree crown plane of suitable size
The screening of point cluster and division, schematic diagram are as shown in Figure 3.
Tree crown window calculation be crown surface be screened mark off come after, it is oblique according to trunk center line and tree crown section window
Rate, schematic diagram is as shown in figure 4, be calculated the radius of single tree crown.
The tree crown radius that upper step is calculated, progress profile is delineated at each tree crown extreme point being calculated, according to
The radius that trunk center line and window calculation obtain carries out two dimensional surface tree crown circular edge and delineated.
Finally, after different type crown outline is completed to delineate, the tree crown in statistical result on different tree crown calculation windows
Classify quantity, for count, analyze and the later stage modeling prepare.
The present invention utilizes scientific investigation ship ship-borne helicopter or airborne laser spatial digitizer (LiDAR) system of UAV flight
Carry out active target identification, it is possible to achieve the modeling to the vegetation scan data in outboard space such as island/reef, Near shore environment
With extraction；By the collection to related vegetation information, the perception to Near shore environment is realized, lifts the space exploration of scientific investigation ship with determining
Plan supporting capacity；A set of different types vegetation plant canopy model is proposed, optimizes the process and result of filtering.
It should be appreciated that for those of ordinary skills, can according to the above description be improved or converted,
And all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (6)
1. plant canopy model extracting method in a kind of laser scanning environment cloud data, it is characterised in that comprise the following steps, step
One, the scanning of discretization point cloud is carried out to object of observation using airborne LiDAR systems, tree crown classification is carried out according to typical tree crown type
Modeling, the point cloud that obtains will be scanned it is abstracted gabarit space structure type and incorporate into as taper, hemispherical, semielliptical shape, and establishes pair
Answer the surfaces externally and internally model of structure；Step 2, dispersion point cloud is carried out at image rasterizing using 0.5 meter of -3 meters of sampling interval
Reason, it can be converted into being available for later stage Object identifying, the bidimensional image of spatial analysis processing；Step 3, for different trees
Horizontal face diameter is preced with, tree crown of the adjustable filtering calculation window of selection dynamic towards tree crown coverage radius less than 3m is put down
Screened and divided in face；Step 4, after crown surface is screened and marks off, according to trunk center line and tree crown section window
The radius of single tree crown is calculated in slope；Step 5, the radius obtained according to trunk center line and window calculation carry out two dimension
Plane tree crown circular edge is delineated；Step 6, after different type crown outline is completed to delineate, different tree crown meters in statistical result
Calculate window on tree crown classification quantity, for count, analyze and the later stage modeling prepare.
2. plant canopy model extracting method in laser scanning environment cloud data according to claim 1, it is characterised in that institute
State in step 4, specially searching Local Extremum is compared in energy is restrained and calculates iterative process, and tree crown center is covered
Lid is labeled as tree crown point less than the maximum point of 3 meters of radiuses, then periphery tree crown surface point is carried out centered on tree crown point oblique
Rate calculates, and variant slope can calculate its angle by vertical section, according to tree crown surface point to trunk center line level away from
From triangle relation radius according to different plant canopy models on its maximal cover perspective plane is calculated.
3. plant canopy model extracting method in laser scanning environment cloud data according to claim 1 or 2, its feature exist
In in the step 5, specially use detection mask calculates the data point in the range of 3mx3m, finds tree crown locally potential pole
It is worth region, reduces and calculate consumption, in each mask, search window is determined in the way of fixed slope, using improved base
Data point cluster is carried out from iteration mean shift algorithm in energy convergent function, obtains tree crown point；If calculate in cluster
Heart point is consistent with the extreme point in each mask, then labeled as tree crown extreme point herein, otherwise using the cluster point as newly
Mask point brings cluster calculation next time into, until the cluster centre point that iteration goes out is identical with mask extreme point.
4. plant canopy model extracting method in laser scanning environment cloud data according to claim 3, it is characterised in that every
Sweep the potential energy sum that the energy potential-energy function between described point is defined as independent each point, each potential energy P (x individually puti) it is xjTo xiRespectively
To the kernel function G (x of the same sexi) and bandwidth parameter hxiForm, i.e., by xiTo xjPotential energy be P (i, j)=S (hxi)(1-g(xi-
xj)/hxi), wherein S (hxi) be potential energy constraint factor, for control whole function can monotonic increase, g (xi) it is kernel function G
(xi) gradient, by xiTo xjWith xjTo xiMutual potential energy summation, be this potential-energy function value P (xi)=Σ (P (i, j)+P
(j,i))。
5. plant canopy model extracting method in laser scanning environment cloud data according to claim 1 or 2, its feature exist
In in the step 1, the mathematical modeling of corresponding three kinds of tree crowns gabarit spatial shape is respectively：
Cone：Fout=R/tan (B)
Hemisphere：Fout=[R2-(x2+y2)]1/2
Semiellipsoid：Fout=e* [R2-(x2+y2)]1/2
Wherein, fout is scan model outer surface, and B is that tree crown external profile surface projects angle with vertical plane, and R is that external profile surface is being thrown
The radius of space projection circle in shadow plane, x are the transverse axis coordinate of scanning element on a projection plane, and y is that scanning element is projecting
Ordinate of orthogonal axes in plane, e are the ratio of long axis to short axis of spheroid.
6. plant canopy model extracting method in laser scanning environment cloud data according to claim 5, it is characterised in that according to
According to plane trigonometry relation, complete crown mapping radius R=r/ [sin (180-2*A)], r is current preset calculation window radius；A
For any scanning element on tree crown external profile surface to the angle of trunk center line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710891872.5A CN107705344A (en) | 2017-09-27 | 2017-09-27 | Plant canopy model extracting method in laser scanning environment cloud data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710891872.5A CN107705344A (en) | 2017-09-27 | 2017-09-27 | Plant canopy model extracting method in laser scanning environment cloud data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107705344A true CN107705344A (en) | 2018-02-16 |
Family
ID=61176205
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710891872.5A Pending CN107705344A (en) | 2017-09-27 | 2017-09-27 | Plant canopy model extracting method in laser scanning environment cloud data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107705344A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108957476A (en) * | 2018-09-26 | 2018-12-07 | 中国计量大学 | A kind of Diameter measure device and tree-walk measuring method based on unmanned plane |
| CN110070550A (en) * | 2019-04-26 | 2019-07-30 | 中国农业大学 | Finishing strategy acquisition methods, device and the electronic equipment of forest |
| CN111160236A (en) * | 2019-12-27 | 2020-05-15 | 北京林业大学 | Automatic dividing method for watershed canopy by combining forest region three-dimensional morphological filtering |
| CN116862985A (en) * | 2023-09-01 | 2023-10-10 | 江苏中安建设集团有限公司 | Land high-precision positioning electronic informatization monitoring method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761744A (en) * | 2014-02-21 | 2014-04-30 | 哈尔滨工业大学 | General type two-dimensional point cloud grouping method |
| CN104849722A (en) * | 2015-05-08 | 2015-08-19 | 南京林业大学 | Light detection and ranging (LiDAR) waveform comprehensive feature-based individual tree identification method |
| CN105046688A (en) * | 2015-06-23 | 2015-11-11 | 北京工业大学 | Method for automatically identifying multiple planes in three-dimensional point cloud |
| US20160063754A1 (en) * | 2014-08-26 | 2016-03-03 | The Boeing Company | System and Method for Detecting a Structural Opening in a Three Dimensional Point Cloud |
| CN105469098A (en) * | 2015-11-20 | 2016-04-06 | 中北大学 | Precise LINDAR data ground object classification method based on adaptive characteristic weight synthesis |
| CN106077251A (en) * | 2016-06-27 | 2016-11-09 | 哈尔滨工业大学 | A kind of induced with laser free forming method and device of cross bar integral panel |
| CN106815847A (en) * | 2017-01-12 | 2017-06-09 | 非凡智慧（宁夏）科技有限公司 | Trees dividing method and single tree extracting method based on laser radar point cloud |
-
2017
- 2017-09-27 CN CN201710891872.5A patent/CN107705344A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761744A (en) * | 2014-02-21 | 2014-04-30 | 哈尔滨工业大学 | General type two-dimensional point cloud grouping method |
| US20160063754A1 (en) * | 2014-08-26 | 2016-03-03 | The Boeing Company | System and Method for Detecting a Structural Opening in a Three Dimensional Point Cloud |
| CN104849722A (en) * | 2015-05-08 | 2015-08-19 | 南京林业大学 | Light detection and ranging (LiDAR) waveform comprehensive feature-based individual tree identification method |
| CN105046688A (en) * | 2015-06-23 | 2015-11-11 | 北京工业大学 | Method for automatically identifying multiple planes in three-dimensional point cloud |
| CN105469098A (en) * | 2015-11-20 | 2016-04-06 | 中北大学 | Precise LINDAR data ground object classification method based on adaptive characteristic weight synthesis |
| CN106077251A (en) * | 2016-06-27 | 2016-11-09 | 哈尔滨工业大学 | A kind of induced with laser free forming method and device of cross bar integral panel |
| CN106815847A (en) * | 2017-01-12 | 2017-06-09 | 非凡智慧（宁夏）科技有限公司 | Trees dividing method and single tree extracting method based on laser radar point cloud |
Non-Patent Citations (1)
| Title |
|---|
| LIU, LIKUN: "Individual Tree Shape Modeling for Canopy Delineation from Airborne LiDAR Data", 《PROCEEDINGS OF SPIE》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108957476A (en) * | 2018-09-26 | 2018-12-07 | 中国计量大学 | A kind of Diameter measure device and tree-walk measuring method based on unmanned plane |
| CN110070550A (en) * | 2019-04-26 | 2019-07-30 | 中国农业大学 | Finishing strategy acquisition methods, device and the electronic equipment of forest |
| CN110070550B (en) * | 2019-04-26 | 2021-09-28 | 中国农业大学 | Forest tree trimming strategy obtaining method and device and electronic equipment |
| CN111160236A (en) * | 2019-12-27 | 2020-05-15 | 北京林业大学 | Automatic dividing method for watershed canopy by combining forest region three-dimensional morphological filtering |
| CN116862985A (en) * | 2023-09-01 | 2023-10-10 | 江苏中安建设集团有限公司 | Land high-precision positioning electronic informatization monitoring method |
| CN116862985B (en) * | 2023-09-01 | 2023-11-10 | 江苏中安建设集团有限公司 | Land high-precision positioning electronic informatization monitoring method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Individual tree crown delineation using localized contour tree method and airborne LiDAR data in coniferous forests | |
| Yang et al. | An individual tree segmentation method based on watershed algorithm and three-dimensional spatial distribution analysis from airborne LiDAR point clouds | |
| CN112381861B (en) | Forest land point cloud data registration and segmentation method based on foundation laser radar | |
| CN111898688B (en) | Airborne LiDAR data tree classification method based on three-dimensional deep learning | |
| CN106199557B (en) | A kind of airborne laser radar data vegetation extracting method | |
| CN105389538B (en) | A method of based on a cloud hemisphere slice estimation Forest Leaf Area Index | |
| Allouis et al. | Stem volume and above-ground biomass estimation of individual pine trees from LiDAR data: Contribution of full-waveform signals | |
| CN113591766B (en) | Multi-source remote sensing tree species identification method for unmanned aerial vehicle | |
| CN107705309A (en) | Forest parameter evaluation method in laser point cloud | |
| CN107273902B (en) | A method of pylon point cloud is automatically extracted from on-board LiDAR data | |
| CN109146889A (en) | A kind of field boundary extracting method based on high-resolution remote sensing image | |
| CN107705344A (en) | Plant canopy model extracting method in laser scanning environment cloud data | |
| CN112819830A (en) | Individual tree crown segmentation method based on deep learning and airborne laser point cloud | |
| CN111612896B (en) | Method for reconstructing three-dimensional tree model based on airborne laser radar tree point cloud | |
| CN107479065B (en) | Forest gap three-dimensional structure measuring method based on laser radar | |
| Bienert et al. | Voxel space analysis of terrestrial laser scans in forests for wind field modelling | |
| CN107330422A (en) | A kind of method for carrying out mima type microrelief classification to semiarid zone based on high accuracy number elevation model | |
| CN109766824B (en) | Active and passive remote sensing data fusion classification method based on fuzzy evidence theory | |
| Yan | Pixel based and object oriented image analysis for coal fire research | |
| Dutta et al. | Characterizing vegetation canopy structure using airborne remote sensing data | |
| CN111091079A (en) | TLS-based method for measuring dominant single plant structural parameters of vegetation in alpine and fragile regions | |
| Li et al. | An iterative-mode scan design of terrestrial laser scanning in forests for minimizing occlusion effects | |
| Ferraz et al. | Canopy density model: A new ALS-derived product to generate multilayer crown cover maps | |
| CN115880487A (en) | Forest laser point cloud branch and leaf separation method based on deep learning method | |
| Liu et al. | Individual tree identification using a new cluster-based approach with discrete-return airborne LiDAR data |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180216 |