A simulation method of laser pulse echo of jungle target based on KD tree acceleration

Abstract

The present invention provides a method for simulating laser pulse echoes from jungle targets based on KD tree acceleration, comprising the following steps: Step S1: constructing a physical model of a tree, comprising a trunk and a crown; Step S2: constructing a laser scattering model of the tree; Step S3: constructing a KD tree structure encompassing the jungle and background environment; Step S4: sampling the laser beam emitted by a laser radar (LiDAR), generating a photon beam based on its spatial energy distribution, and performing wavefront segmentation; Step S5: calculating the interaction between each photon in the photon beam generated in Step S4 and the KD tree structure generated in Step S3 to determine whether the photon hits the trunk or crown of the tree in the KD tree structure constructed in Step S3. Step S6: calculating the scattering events of the photons along the path; and Step S7: calculating the laser radar pulse echo based on the interaction between the photon beam and the KD tree structure obtained in Step S5. The present invention has the advantages of wide applicability and ease of use.

Description

Technical Field

[0001] The present invention belongs to the technical field of laser simulation, and in particular relates to a method for simulating laser pulse echoes of a jungle target based on a KD tree acceleration. Background Art

[0002] Rapidly simulating the scattering of laser pulses emitted by lidar in a complex jungle background environment and solving the echo time domain waveform of the laser pulse has important uses in many fields.

[0003] When solving LiDAR echoes for jungle targets, most traditional Monte Carlo ray tracing methods are unable to meet the low radiation flux requirements due to the low backscattered photon flux reaching the laser detector. This results in inefficient and inaccurate time-domain waveform calculations for the laser pulse echo. The echo's time-domain waveform depends on the spatial relationship between the target scene and the photon delay reaching the detector. Furthermore, the interaction between the laser and the scene is complex, with absorption, single / multiple reflections, and volume scattering, further complicating the echo's time-domain waveform.

[0004] Prior art describes a method for calculating the near-field echo power of a laser fuze based on beam decomposition and localized illumination. This method only considers the effects of a single reflection and ignores multiple reflections. Therefore, it can only be used for calculations involving single targets, and target structures cannot be significantly affected by multiple reflections. However, for jungle targets, multiple scattering is a major factor, making this method unsuitable for laser pulse echo simulations of jungle targets.

[0005] Therefore, there is an urgent need to propose a method that can quickly simulate the scattering effect of laser pulses emitted by lidar in a complex jungle background environment and obtain the echo time domain waveform of the laser pulse. Summary of the Invention

[0006] To solve the above problems, the present invention proposes a jungle target laser pulse echo simulation method based on KD tree acceleration, which is suitable for the rapid simulation of pulse echoes for jungle target detection by various lidars. It provides a large amount of low-cost input data for pulse echo analysis, which can be used for training various analysis algorithms, and can also be used for the demonstration and development of lidars, saving a lot of manpower and material resources.

[0007] To achieve the above objectives, the present invention provides a method for simulating laser pulse echoes from jungle targets based on KD tree acceleration, comprising the following steps: Step S1: constructing a physical model of a tree, the physical model comprising a trunk and a crown; Step S2: constructing a laser scattering model of the tree; Step S3: constructing a KD tree structure encompassing the jungle and background environment, wherein the jungle is composed of multiple trees; Step S4: sampling the laser beam emitted by a laser radar, generating a photon beam based on its spatial energy distribution, and performing wavefront segmentation; Step S5: calculating the interaction between each photon in the photon beam generated in Step S4 and the KD tree structure generated in Step S3 to determine whether the photon hits the trunk or crown of the tree in the KD tree structure constructed in Step S3; Step S6: calculating the scattering events of the photons along the path; and Step S7: calculating the laser radar pulse echo based on the interaction between the photon beam and the KD tree structure obtained in Step S5.

[0008] Preferably, the trunk portion is represented by a triangular mesh, and the crown portion is represented by a set of discrete bounding boxes.

[0009] Preferably, the triangular mesh form adopted by the trunk part includes point, surface structure index or triangular facet list form; the crown part is composed of a large number of leaves, and the leaves are converted into a continuous axis-aligned discrete bounding box structure of multiple spatial regions with similar volume scattering characteristics to form the crown part; each of the discrete bounding boxes equivalently represents the scattering characteristics of all leaves in the corresponding spatial region; wherein, the discrete bounding box contains parameters of the scattering phase function, scattering cross section, and absorption cross section volume scattering parameters.

[0010] Preferably, the trunk of the tree adopts the bidirectional reflectance distribution function f r Characterizes the scattering characteristics, which is related to the incident radiation brightness L i , the outgoing radiation brightness L o The relationship is: Where θ is the angle of incidence, dω i is the differential solid angle in the incident direction, Ω 2π is the hemispherical space; the bidirectional reflectance distribution function f r A bidirectional reflectance distribution function including a rough surface diffuse reflection term is selected; the crown part is divided into a set of small cubes with different volume scattering characteristics, and the volume scattering model is used to characterize the scattering characteristics. The scattered radiation brightness L s The relationship with the incident radiation brightness satisfies: where σ s (ω s ) is the scattering cross section, p(ω i ,ω s ) is the scattering phase function, ω i ,ω sare the directions of the incident vector and the outgoing vector, dω i is the differential solid angle in the incident direction.

[0011] Preferably, the KD tree structure constructed in step S3, which includes the jungle and the background environment, is defined in XML format, specifically comprising the following steps: step S301, reading the triangular mesh data of the trunk part of the tree and the discrete bounding box data of the crown part through XML to define the position and orientation of the tree, and generating the position and orientation information of the tree; step S302, reading the reflection data and the volume scattering data through XML to define the size and reflection data description of the background ground, and generating the spatial information of the background ground; step S303, organizing the position and orientation information of the tree and the spatial information of the background ground into a KD tree structure according to the information generated in step S301 and step S302, so as to accelerate the photon tracing process and the reverse ray sampling process.

[0012] Preferably, the KD tree structure includes two forms, one form is a geometric scene KD tree, which is used to store the jungle and background environment information constructed in step S3, and the other form is a photon graph structure, which is used to store discrete photon point information.

[0013] Preferably, the calculation of the scattering events of photons on the path described in step S6 specifically includes: if the photon hits the trunk part, the photon information is recorded in the photon graph structure of the KD tree structure; if the photon hits the discrete bounding box of the crown part, the photon information is recorded in the photon graph structure of the KD tree structure according to the volume scattering situation, and the next event of the photon is determined according to probability, and the event content includes being scattered or absorbed, and the scattered photon continues to step S5; wherein, the photon information recorded when the photon hits the trunk part includes the time of arrival at the focus, the direction of the photon incidence, the magnitude of the photon energy and the bidirectional reflectance distribution function information at the intersection.

[0014] Preferably, if a photon hits one of the discrete bounding boxes in the crown, the photon information is recorded in the photon graph structure of the KD tree structure according to the volume scattering situation, and the scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box to generate a uniformly distributed first random number ξ. The mean free path of the photon in the discrete bounding box is determined according to Beer's law, which is specifically determined by the following formula: Where d is the mean free path of the photon in the discrete bounding box; if the photon exceeds the discrete bounding box after traveling the distance d, the photon is considered to need to re-scatter the step, that is, continue with step S5; if the photon is still within the discrete bounding box after traveling the distance d, a roulette method is used to determine whether the photon is absorbed or scattered, and its scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box, and a uniformly distributed second random number ξ is generated. If the second random number ξ is greater than σ s / σt , the photon is absorbed and the movement of the photon ends, otherwise the photon is scattered; the importance sampling technique is used based on the scattering phase function p(ω i ,ω s ) Determine the direction of movement of the photon, and the scattered photon continues to step S5.

[0015] Preferably, the calculation of the laser radar pulse echo described in step S7 includes the following steps: step S701, traversing all photons, generating a KD tree photon graph structure, and balancing the KD tree photon graph structure; step S702, emitting sampling rays from the detector end of the laser radar to interact with the geometric scene KD tree, and if there is an intersection, searching for the illumination information at the point according to the information in the KD tree photon graph structure; step S703, multiplying the radiation brightness value, detection gain and time domain pulse signal obtained by all sampling rays, dividing the receiving time threshold into discrete time periods, and weightedly superimposing the time and energy of each effective sampling ray path to obtain the laser radar pulse echo.

[0016] Preferably, if the sampling ray intersects the tree trunk, the ray radiation brightness is calculated according to the following formula: Among them, ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the disk containing N photons nearby, f r (ω s ,ω i ) is the bidirectional reflectance distribution function. The calculated radiance is weighted by the receiving time threshold according to time and divided into discrete time periods. If the ray intersects the tree canopy, the ray radiance is calculated according to the following formula: Among them, ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the sphere containing N photons nearby, p(ω i ,ω s ) is the scattering phase function, and the calculated radiation brightness is weighted to the receiving time threshold according to time and divided into discrete time periods.

[0017] In summary, compared with the existing technology, the jungle target laser pulse echo simulation method based on KD tree acceleration provided by the present invention has the following beneficial effects: 1. It has a wide range of applications. The present invention is suitable for rapid simulation of pulse echoes for jungle target detection by various laser radars; 2. It can provide a large amount of low-cost input data for pulse echo analysis, which can be used for training various analysis algorithms and demonstration and research and development of laser radars; 3. The present invention is easy to use and can save a lot of manpower and material costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1This is a flow chart of the jungle target laser pulse echo simulation method based on KD tree acceleration of the present invention;

[0019] Figure 2 Schematic diagram of photon energy estimation in the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the appended Figure 1 ~Attached Figure 2 , the technical solutions, structural features, objectives achieved and effects in the embodiments of the present invention are described in detail.

[0021] It should be noted that, in the present invention, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus comprising a series of elements includes not only the elements explicitly listed, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus.

[0022] The present invention provides a method for simulating laser pulse echoes of jungle targets based on KD tree acceleration, wherein the KD tree (short for k-dimensional tree) is a data structure for partitioning k-dimensional data space and is currently an effective means of accelerating the space traversal process. The simulation steps are as follows: Figure 1 As shown, it mainly includes the following steps:

[0023] Step S1: construct a physical model of the tree.

[0024] The physical model of the tree includes a trunk and a crown; the trunk is represented by a triangular mesh, and the crown is represented by a set of discrete bounding boxes; the triangular mesh used for the trunk includes a point, a surface structure index, or a triangular facet list; the crown is obtained by converting the physical properties of the tree (such as spatial geometric distribution, leaf angle distribution, leaf area density, etc.); further, the crown is composed of a large number of leaves, which are converted into a continuous axis-aligned discrete bounding box structure of multiple spatial regions with similar volume scattering characteristics to form the crown; each discrete bounding box equivalently represents the scattering characteristics of all leaves in the corresponding spatial region, and the size of the discrete bounding box can be a fixed value or adaptively determined according to the leaf area density in the spatial region. The physical properties of the tree are obtained by lidar measurement or simulation. The discrete bounding boxes contain volume scattering parameters such as the parameters of the scattering phase function, the scattering cross section, and the absorption cross section.

[0025] Step S2: constructing a laser scattering model of trees.

[0026] The scattering characteristics of the trunk of the tree are characterized by the bidirectional reflectance distribution function, and the scattering characteristics of the crown of the tree are characterized by the volume scattering model.

[0027] Specifically, the trunk of the tree adopts the bidirectional reflectance distribution function f r Characterizes the scattering characteristics, which is related to the incident radiation brightness L i , the outgoing radiation brightness L o The relationship is:

[0028]

[0029] Where θ is the angle of incidence, dω i is the differential solid angle in the incident direction, Ω 2π is the hemispherical space (i.e., the integration range); the bidirectional reflectance distribution function f r There are a variety of models that can be used. For tree trunk scattering, a bidirectional reflectance distribution function including rough surface diffuse reflection terms can be used.

[0030] The crown is divided into a set of small cubes with different volume scattering properties according to the crown composition rules described in step S1 (equivalent to a discrete bounding box structure with multiple continuous axis alignments in a spatial region). The volume scattering model is used to characterize the scattering properties. When the laser is transmitted in this volume scattering medium, Beer's law is satisfied, that is:

[0031]

[0032] Where T r is the transmittance of photons passing through distance d, σ t is the extinction cross section. Scattered radiation brightness L s The relationship with the incident radiation brightness satisfies:

[0033]

[0034] where σ s (ω s ) is the scattering cross section, p(ω i ,ω s ) is the scattering phase function, ω i ,ω s are the directions of the incident vector and the outgoing vector, dω i is the differential solid angle of the incident direction, the scattering cross section σ in the function s (ω s ) and extinction cross section σ t The parameters are determined by the discrete bounding box structure of step S1.

[0035] Step S3: construct a KD tree structure including a jungle and a background environment, wherein the jungle is composed of multiple trees.

[0036] Among them, the KD tree structure constructed in step S3, which includes the jungle and background environment, is defined in the form of XML (Extensible Markup Language), and specifically includes the following steps: Step S301, respectively reading the triangular mesh data of the trunk part of the tree and the discrete bounding box data of the crown part through XML to define the position and orientation of the tree, and generate the position and orientation information of the tree; Step S302, respectively reading the reflection data and volume scattering data through XML to define the size and reflection data description of the background ground, and generate the spatial information of the background ground; Step S303, according to the information generated in step S301 and step S302, the position and orientation information of the tree and the spatial information of the background ground are organized into a KD tree structure to accelerate the photon tracing process and the reverse ray sampling process.

[0037] The KD tree structure consists of two forms: a geometric scene KD tree, which is used to store the jungle and background environment information constructed in step S3; and a photon graph structure, which is used to store discrete photon point information. The spatial positional structure of the tree targets and the background ground in the jungle background environment constructed in step S3 is organized using a KD tree structure, which greatly accelerates calculation speed.

[0038] Step S4: sample the laser beam emitted by the laser radar, generate a photon beam according to the spatial energy distribution, and perform wavefront segmentation.

[0039] Specifically, since the emission energy of the laser beam is a spatial Gaussian distribution, different regions correspond to different energies. Therefore, it is necessary to discretize the laser beam in the region with similar energy into sub-beams, and perform a wavefront segmentation operation on the laser beam emitted by the lidar according to the distribution characteristics of the spatial energy before the photon traversal. The spatial energy distribution of the laser beam can be defined in various forms, such as discrete spatial sampling points, or expressed in analytical form, for example:

[0040] I=I0exp(-2r 2 / ω 2 )

[0041] Where I0 is the peak power, ω is the beam divergence angle, and r is the distance to the maximum power.

[0042] Sampling of the laser beam emitted by the lidar can be performed in a variety of ways, such as uniform sampling, which adds weight information to the photons themselves; or importance sampling, which gives the same weight to each photon, but the number of samples is determined according to the spatial distribution of energy. Furthermore, the direction of photon emission is determined according to the direction of laser beam emission, such as Figure 2 shown.

[0043] Step S5: Calculate the result of the interaction between each photon in the photon beam generated in step S4 and the KD tree structure generated in step S3 to determine whether the photon hits the trunk or crown of the KD tree structure constructed in step S3.

[0044] Among them, the KD tree structure is used to accelerate the photon intersection process, and different labels are used for the trunk part and the crown part to distinguish their different types, which can greatly accelerate the calculation of the KD tree intersection of photons and geometric scenes; further, since the scattering models of the trunk part and the crown part are different, the calculation methods used in the KD tree intersection process of photons and geometric scenes are also different. Specifically, the intersection process of photons and trunk parts is surface scattering calculation, and the intersection process of photons and crown parts is volume scattering calculation.

[0045] Step S6: Calculate the scattering events of photons along the path.

[0046] Specifically, if a photon hits the trunk, the photon information is recorded in the photon graph structure of the KD tree structure. If a photon hits the discrete bounding box of the crown, the photon information is recorded in the photon graph structure of the KD tree structure based on the volume scattering situation, and the next event of the photon is determined by probability. The event content includes scattering or absorption. The scattered photon proceeds to step S5.

[0047] Among them, the photon information recorded when the photon hits the tree trunk includes the time of arrival at the focus, the direction of the photon incidence, the magnitude of the photon energy, and the bidirectional reflection distribution function information at the intersection.

[0048] If a photon hits one of the discrete bounding boxes in the crown, the photon information is recorded in the photon graph structure of the KD tree structure according to the volume scattering situation. The scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box, and a uniformly distributed first random number ξ is generated. The mean free path of the photon in the discrete bounding box is determined according to Beer's law, which is determined by the following formula:

[0049]

[0050] Where d is the mean free path of the photon in the discrete bounding box; if the photon exceeds the discrete bounding box after traveling the distance d, that is, the mean free path d of the photon is greater than the length of the discrete bounding box, then the photon is considered to need to re-perform the scattering step, that is, continue with step S5.

[0051] If the photon is still within the discrete bounding box after traveling the distance d, that is, the mean free path d of the photon is less than the length of the discrete bounding box, a roulette wheel method is used to determine whether the photon is absorbed or scattered. The scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box, and a uniformly distributed second random number ξ is generated. If the second random number ξ is greater than σ s / σ t , the photon is absorbed and the movement of the photon ends, otherwise the photon is scattered; the importance sampling technique is used based on the scattering phase function p(ω i ,ω s ) Determine the direction of movement of the photon, and the scattered photon continues to step S5.

[0052] Step S7: Calculate the laser radar pulse echo based on the result of the interaction between the photon beam and the KD tree obtained in step S5.

[0053] Specifically, step S7 calculates the pulse echo of the laser radar and includes the following steps: step S701, traverses all photons, generates a KD tree photon graph structure, and balances the KD tree photon graph structure; step S702, emits a sampling ray from the detector end of the laser radar to interact with the geometric scene KD tree, and if there is an intersection, searches for the illumination information at the point according to the information in the KD tree photon graph structure; step S703, multiplies the radiation brightness value, detection gain and time domain pulse signal obtained by all sampling rays, divides the receiving time threshold into discrete time periods, and weightedly superimposes the time and energy of each effective sampling ray path to obtain the pulse echo of the laser radar.

[0054] If the sampling ray intersects the tree trunk, the ray radiation brightness is calculated according to the following formula:

[0055]

[0056] Where ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the disk containing N photons nearby, f r (ω s ,ω i ) is the bidirectional reflectance distribution function. The calculated radiance is weighted to the receiving time threshold according to time and divided into discrete time periods.

[0057] If the sampling ray intersects the tree crown, the ray radiation brightness is calculated according to the following formula:

[0058]

[0059] Where ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the sphere containing N photons nearby, p(ω i ,ωs ) is the scattering phase function. The calculated radiance is weighted to the receiving time threshold according to time and divided into discrete time periods.

[0060] The radiation brightness values ​​obtained from all sampling rays are multiplied by the detection gain and the time domain pulse signal, and the time and energy of each effective ray path are weighted and superimposed to obtain the pulse echo of the laser radar; if the emission pulse energy of the sampling ray is a time series The corresponding emission times are t0, t1, t2...t n , the time interval is Δt. The number of simulated photons emitted at time t0 is N0, t n The number of simulated photons emitted at the moment is N0 I tn / I t0 Assume that the echo sampling time interval is Δt', the first echo sampling time is t0', and the number of photons in the interval [t0'-Δt' / 2, t0'+Δt' / 2] is superimposed to obtain the echo intensity at t0'. Similarly, [t n '-Δt' / 2,t n The number of photons in the interval '+Δt' / 2] is superimposed to obtain t n 'The echo strength at that moment.

[0061] In summary, compared with the existing technology, the jungle target laser pulse echo simulation method based on KD tree acceleration provided by the present invention has the advantages of wide application range and ease of use.

[0062] Although the present invention has been described in detail through the above preferred embodiments, it should be understood that the above description is not intended to limit the present invention. After reading the above description, various modifications and substitutions of the present invention will become apparent to those skilled in the art. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims

1. A jungle target laser pulse echo simulation method based on KD tree acceleration, characterized in that: The steps include: Step S1: constructing a physical model of a tree, wherein the physical model of the tree includes a trunk part and a crown part; Step S2: constructing a laser scattering model of trees; Step S3: constructing a KD tree structure including a jungle and a background environment, wherein the jungle is composed of multiple trees; Step S4: sampling the laser beam emitted by the laser radar, generating a photon beam according to the spatial energy distribution, and performing wavefront segmentation; Step S5: Calculate the result of each photon in the photon beam generated in step S4 interacting with the KD tree structure generated in step S3 to determine whether the photon hits the trunk or crown of the KD tree structure constructed in step S3; Step S6: calculating the scattering events of the photons along the path; wherein the calculating the scattering events of the photons along the path in step S6 specifically includes: If a photon hits the trunk, the photon information is recorded in the photon graph structure of the KD tree structure; if a photon hits the discrete bounding box of the crown, the photon information is recorded in the photon graph structure of the KD tree structure according to the volume scattering situation, and the next event of the photon is determined by probability. The event content includes scattering or absorption. The importance sampling technology is used based on the scattering phase function p(ω i ,ω s ) determines the direction of motion of the photon, and the scattered photon proceeds to step S5, where ω i ,ω s are the directions of the incident vector and the outgoing vector respectively; The photon information recorded when the photon hits the tree trunk includes the time of arrival at the focus, the direction of the photon incident, the magnitude of the photon energy, and the bidirectional reflectance distribution function information at the intersection; Step S7: Calculate the laser radar pulse echo based on the result of the interaction between the photon beam obtained in step S5 and the KD tree.

2. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 1, characterized in that: The trunk is represented by a triangle mesh, and the crown is represented by a set of discrete bounding boxes.

3. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 2, characterized in that: The triangular mesh form used in the trunk part includes point, surface structure index or triangular facet list form; The crown portion is composed of a large number of leaves, and the leaves are converted into a plurality of continuous axis-aligned discrete bounding box structures of spatial regions with similar volume scattering characteristics to form the crown portion; Each of the discrete bounding boxes equivalently represents the scattering characteristics of all leaves in the corresponding spatial region; The discrete bounding box contains parameters of the scattering phase function, the scattering cross section, and the absorption cross section volume scattering parameters.

4. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 3, characterized in that: The trunk of the tree adopts the bidirectional reflectance distribution function f r Characterizes the scattering characteristics, which is related to the incident radiation brightness L i , the outgoing radiation brightness L o The relationship is: Where θ is the angle of incidence, dω i is the differential solid angle in the incident direction, Ω 2π is the hemispherical space; the bidirectional reflectance distribution function f r Select the bidirectional reflectance distribution function that includes the rough surface diffuse reflection term; The crown is divided into a set of small cubes with different volume scattering characteristics. The volume scattering model is used to characterize the scattering characteristics. The scattered radiation brightness L s The relationship with the incident radiation brightness satisfies: where σ s (ω s ) is the scattering cross section, p(ω i ,ω s ) is the scattering phase function, ω i ,ω s are the directions of the incident vector and the outgoing vector, dω i is the differential solid angle in the incident direction.

5. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 4, characterized in that: The construction of the KD tree structure including the jungle and the background environment described in step S3 is defined in XML format and specifically includes the following steps: Step S301: Read the triangular mesh data of the trunk and the discrete bounding box data of the crown of the tree through XML to define the position and orientation of the tree, and generate the position and orientation information of the tree; Step S302: Read the reflection data and volume scattering data through XML to define the size of the background ground and the description of the reflection data, and generate spatial information of the background ground; Step S303: Based on the information generated in steps S301 and S302, the position orientation information of the trees and the spatial information of the background ground are organized into a KD tree structure to accelerate the photon tracing process and the reverse ray sampling process.

6. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 5, characterized in that: The KD tree structure includes two forms, one form is a geometric scene KD tree, which is used to store the jungle and background environment information constructed in step S3, and the other form is a photon graph structure, which is used to store discrete photon point information.

7. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 6, characterized in that: If a photon hits one of the discrete bounding boxes in the crown, the photon information is recorded in the photon graph structure of the KD tree structure according to the volume scattering situation. The scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box, and a uniformly distributed first random number ξ is generated. The mean free path of the photon in the discrete bounding box is determined according to Beer's law, which is determined by the following formula: Where d is the mean free path of the photon in the discrete bounding box; if the photon exceeds the discrete bounding box after traveling the mean free path d, the photon is considered to need to undergo the scattering step again, i.e., continue to step S5; If the photon is still within the discrete bounding box after the mean free path d, the roulette wheel method is used to determine whether the photon is absorbed or scattered. The scattering characteristics are calculated according to the volume scattering parameters in the discrete bounding box, and a uniformly distributed second random number ξ is generated. If the second random number ξ is greater than σ s / σ t , the photon is absorbed and ends its motion, otherwise the photon is scattered; where σ s is the scattering cross section, σ t is the extinction cross section.

8. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 7, characterized in that: The step S7 of calculating the laser radar pulse echo comprises the following steps: Step S701: traverse all photons to generate a KD tree photon graph structure, and balance the KD tree photon graph structure; Step S702: A sampling ray is emitted from the detector end of the laser radar to interact with the KD tree of the geometric scene. If there is an intersection, the illumination information at the point is searched based on the information in the KD tree photon graph structure. Step S703: Multiply the radiation brightness values, detection gains, and time domain pulse signals obtained by all sampling rays, divide the receiving time threshold into discrete time periods, perform weighted superposition of the time and energy of each effective sampling ray path, and obtain the laser radar pulse echo.

9. The jungle target laser pulse echo simulation method based on KD tree acceleration according to claim 8, characterized in that: If the sampling ray intersects the tree trunk, the ray radiation brightness is calculated according to the following formula: Among them, ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the disk containing N photons nearby, f r (ω s ,ω i ) is the bidirectional reflectance distribution function, and the calculated radiance is weighted to the receiving time threshold according to time and divided into discrete time periods; If the ray intersects the tree crown, the ray radiation brightness is calculated according to the following formula: Among them, ΔΦ(ω i ) is the radiant illumination of photons in the photon diagram, r is the radius of the sphere containing N photons nearby, p(ω i ,ω s ) is the scattering phase function, and the calculated radiation brightness is weighted according to time to the receiving time threshold and divided into discrete time periods.

Images