CN113075659B - Self-adaptive partitioning preprocessing method and system for grid model of flat scene - Google Patents
Self-adaptive partitioning preprocessing method and system for grid model of flat scene Download PDFInfo
- Publication number
- CN113075659B CN113075659B CN202110338693.5A CN202110338693A CN113075659B CN 113075659 B CN113075659 B CN 113075659B CN 202110338693 A CN202110338693 A CN 202110338693A CN 113075659 B CN113075659 B CN 113075659B
- Authority
- CN
- China
- Prior art keywords
- target
- scene
- bounding box
- area
- circular
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
Abstract
The invention relates to a self-adaptive partitioning preprocessing method and a system for a grid model of a flat scene, wherein the method comprises the following steps: determining a target within a scene; calculating a bounding box of the target; finding a shadow area generated by the radar transmitting rays to the bounding box on the ground; forming a bounding sphere by taking the center of the bounding box as the center of the sphere and taking the farthest distance from the center of the sphere in a shadow area as the radius; and calculating a circular area formed by the intersection of the surrounding ball and the ground, and segmenting the target and the scene in the circular area. The method is suitable for a parallel accelerated large-scene self-adaptive area decomposition preprocessing algorithm, effectively reduces the memory consumption while ensuring the calculation precision, improves the algorithm parallelism, and provides theoretical and algorithm support for continuously promoting the parallelization design and development of a large-scale scene radar simulation system.
Description
Technical Field
The invention relates to the technical field of electromagnetic modeling scene simulation, in particular to a flat scene grid model self-adaptive partitioning preprocessing method and system.
Background
At present, the Synthetic Aperture Radar (SAR) technology is mature day by day, and the Synthetic Aperture Radar (SAR) has the characteristics of all-weather and all-weather, and is widely applied to the military and civil fields of remote sensing investigation, accurate strike, geographical mapping and the like. In the aspects of parameter design, signal processing algorithm, target identification application and the like of the SAR system, a large amount of radar echo data are needed, and the data should meet specific parameter requirements. The outfield radar test flight, namely the radar test of real targets and scenes, is the most direct and important means for acquiring radar echo data. However, the field radar test flight in practical engineering faces many practical difficulties, which are mainly reflected in: a large amount of financial resources, manpower and material resources are consumed in the test flight test, and the system development time is greatly increased; measurement tests are difficult to carry out for specific targets and scenarios, particularly for non-cooperative military targets and scenarios; limited experiments are difficult to ensure the completeness of echo data, such as the full attitude and the omnibearing of a target, the diversity of scenes and the target, and the like. In recent years, the computer simulation technology is rapidly developed, and the development of the radar echo modeling technology for simulating the SAR radar detection scene by adopting a computer is gradually mature. The radar echo simulation modeling technology overcomes the defects of a radar test flight test, has the characteristics of low investment, short period and flexibility, is a feasible way for acquiring radar echo data in the design of a radar system, and is an indispensable research means for deeply excavating the intrinsic mechanism and rule of radar echoes. Therefore, important theoretical significance and application value of the target and scene radar echo modeling technology are researched.
The general flow of radar echo modeling includes: the method comprises three parts of scene physical characteristic modeling, scene electromagnetic scattering characteristic modeling and sensor modeling. Modeling and describing shapes and basic physical parameters of objects, backgrounds and the like in a scene by using scene physical characteristics; the scene electromagnetic scattering modeling reveals the action mechanism and the law of electromagnetic waves and targets and environments in the scene; sensor modeling reflects the principles and processes of operation of radar systems. Wherein, the scene electromagnetic scattering modeling is the core and the foundation of radar echo modeling. Typical battlefield scenes (such as airports, ports and the like) have the characteristics of large size and scale, multiple types of backgrounds and targets, complex coupling relationship and the like, and electromagnetic scattering characteristics such as the coupling between the targets and the backgrounds and the like need to be considered respectively when radar echo modeling is carried out. High-frequency asymptotic methods represented by Physical Optics (PO), physical Diffraction Theory (PTD), ray bounce (SBR) and the like have obvious advantages in the aspects of calculation of scattering characteristics of electrically large complex targets and large-scale scenes, such as high calculation efficiency, extremely low memory overhead, low requirement on computer simulation hardware, clear explanation of Physical mechanisms, easiness in analysis of scattering mechanisms and the like. The research on the high-frequency method is mature in theory, but in the application of large scene echo modeling, the contradiction between the calculation precision and the resource overhead still exists, and the contradiction mainly appears as follows: the scene size is large (hundreds of square kilometers), the target structure is complex, the target, the background and the mutual coupling relationship between the target and the background are complex, and the bottleneck problem that how to solve the problem that the fine calculation of the complex target and the quick calculation of the large scene restrict the application engineering and the generalization of the scene echo modeling technology is solved.
The radar echo simulation technology is a technology for simulating radar echoes of a target and an environment by using a mathematical modeling method based on the action mechanism of electromagnetic waves, the target and the environment, and is commonly used for detecting the performance of a radar system, detecting various imaging algorithms and analyzing the effectiveness of the algorithms based on different models.
Fig. 1 is a flow chart of radar echo simulation, and as shown in fig. 1, when radar echo simulation is performed, a radar simulation scene geometric model is first established, radar system parameters are accurately obtained, the scene geometric model (including the spatial geometric relationship between a target, an environment and a radar) is described, and a foundation is laid for electromagnetic scattering calculation later; then, performing scene electromagnetic scattering modeling to accurately describe the electromagnetic action mechanism of the electromagnetic wave and the scene, wherein the scene electromagnetic scattering modeling comprises target electromagnetic scattering modeling, environment electromagnetic scattering modeling and target-environment coupling electromagnetic scattering modeling; calculating a radar scene echo signal on the basis of establishing a radar simulation scene geometric model and describing the scene geometric model; and finally, obtaining radar echo original data by time-frequency transformation and other methods.
At present, the problems of huge memory overhead and computation time consumption exist in the simulation modeling of radar echoes in a large-scale scene, and particularly, for a flat scene, how to effectively reduce the memory overhead while ensuring the computation accuracy becomes a problem which is continuously solved in the field by improving the algorithm parallelism.
Therefore, in order to overcome the defects, a flat scene grid model adaptive partitioning preprocessing method and a flat scene grid model adaptive partitioning preprocessing system are needed, the method and the system are suitable for a large scene adaptive area decomposition preprocessing algorithm accelerated in parallel, the calculation accuracy is ensured, the memory consumption is effectively reduced, the algorithm parallelism is improved, and theoretical and algorithm support is provided for continuously promoting the parallelization design and development of a large-scale scene radar simulation system.
Disclosure of Invention
The technical problem to be solved by the invention is that the problem of huge memory overhead and computation time consumption exists during large-scale scene radar echo simulation modeling, especially for a flat scene, how to effectively reduce the memory overhead and improve the algorithm parallelism while ensuring the computation accuracy, and aiming at the defects in the prior art, the invention provides a flat scene grid model self-adaptive partition preprocessing method and system.
In order to solve the technical problem, the invention provides a self-adaptive partitioning preprocessing method for a grid model of a flat scene, which comprises the following steps: determining a target within a scene; calculating a bounding box of the target; finding a shadow area generated by the radar transmitting rays to the bounding box on the ground; forming a bounding sphere by taking the center of the bounding box as the center of a sphere and taking the farthest distance from the center of the sphere in a shadow area as a radius; and calculating a circular area formed by the intersection of the enclosing ball and the ground, and segmenting the target and the scene in the circular area.
The invention also provides a self-adaptive partitioning preprocessing method of the grid model of the flat scene, which comprises the following steps: determining all targets within a scene; calculating a bounding box of each target; finding a shadow area generated by the radar transmitting rays to any bounding box on the ground; for each target, taking the center of the corresponding bounding box as the center of a sphere and taking the farthest distance from the center of the sphere in the corresponding shadow area as a radius to form a bounding sphere; calculating a circular area formed by the intersection of each enclosing ball and the ground; regarding two or more overlapped circular areas, taking the objects corresponding to the circular areas as an integral object, calculating a bounding box, a shadow area and a bounding sphere of the integral object, calculating a circular area formed by the intersection of the bounding sphere corresponding to the integral object and the ground, and segmenting the objects and the scenes in the circular area; for circular areas that do not overlap with other circular areas, objects and scenes within the circular areas are segmented.
Preferably, the bounding box of the object is computed, comprising the steps of: for any point i in the target, its coordinate is (X) i ,Y i ,Z i ) For X i ,Y i ,Z i Has the following relationship that,
X min ≤X i ≤X max ；
Y min ≤Y i ≤Y max ；
Z min ≤Z i ≤Z max ；
determining a maximum coordinate point and a minimum coordinate point of the target; the coordinate of the maximum coordinate point is (X) max ,Y max ,Z max ) The coordinate of the minimum point of the coordinates is (X) min ,Y min ,Z min ) (ii) a And forming a bounding box by using the maximum coordinate point and the minimum coordinate point, wherein the bounding box is a cuboid, and the maximum coordinate point and the minimum coordinate point are two end points of a diagonal line of the cuboid.
Preferably, finding the shadow region on the ground generated by the radar transmitting rays into the bounding box comprises the following steps: determining coordinates (x) of a sensor of a radar 1 ,y 1 ,z 1 ) (ii) a Connecting the coordinate position of the sensor with each vertex of the bounding box and extending the coordinate position to the ground, finding out the intersection point of the extension line of each connection line and the ground, wherein the coordinate of a certain vertex k of the bounding box is (x) k ,y k ,z k ) The coordinates of the intersection point corresponding to the vertex k are (x, y, z), and satisfy the following relationship:
and can further obtain:
and the connection lines of the intersection points corresponding to all the vertexes form a shadow area.
Preferably, the area obtained by dividing the target and the scene in the circular area is a target sub-area, the remaining area is an environment area, and the environment area is divided into sub-areas by adopting a boundary size sharing mode.
In another aspect, the present invention provides a flat scene mesh model adaptive partitioning preprocessing system, including: the determining module is used for determining a target in the scene; the first calculation module is used for calculating the bounding box of the target; the second calculation module is used for finding a shadow area generated by the radar transmitting rays to the bounding box on the ground; the third calculation module is used for forming a bounding sphere by taking the center of the bounding box as the center of the sphere and taking the farthest distance from the center of the sphere in the shadow area as the radius; and the fourth calculation module is used for calculating a circular area formed by the intersection of the enclosing ball and the ground and segmenting the target and the scene in the circular area.
The invention also provides a self-adaptive partitioning preprocessing system for the grid model of the flat scene, which comprises the following steps: the determining module is used for determining all targets in the scene; the first calculation module is used for calculating the bounding box of each target; the second calculation module is used for finding a shadow area generated by the radar transmitting rays to any bounding box on the ground; the third calculation module is used for forming a bounding sphere by taking the center of the corresponding bounding box as the center of the sphere and the farthest distance from the center of the sphere in the corresponding shadow area as the radius for each target; the fourth calculation module is used for calculating a circular area formed by the intersection of each enclosing ball and the ground; for two or more overlapped circular areas, the determining module takes the targets corresponding to the circular areas as an integral target, the first calculating module, the second calculating module and the third calculating module respectively calculate a bounding box, a shadow area and a bounding sphere of the integral target, the fourth calculating module is also used for calculating a circular area formed by the intersection of the bounding sphere corresponding to the integral target and the ground, and dividing the targets and the scenes in the circular area; for circular areas that do not overlap with other circular areas, the fourth computing module segments the objects and scenes within these circular areas.
Preferably, the first calculation module calculates a bounding box of the object, comprising the steps of: for any of the targetsA point i having coordinates of (X) i ,Y i ,Z i ) For X i ,Y i ,Z i Has the following relationship, X min ≤X i ≤X max ；
Y min ≤Y i ≤Y max ；
Z min ≤Z i ≤Z max ；
Determining a maximum coordinate point and a minimum coordinate point of the target; the coordinate of the maximum coordinate point is (X) max ,Y max ,Z max ) The coordinate of the minimum point of the coordinates is (X) min ,Y min ,Z min ) (ii) a And forming a bounding box by using the maximum coordinate point and the minimum coordinate point, wherein the bounding box is a cuboid, and the maximum coordinate point and the minimum coordinate point are two end points of a diagonal line of the cuboid.
Preferably, the second calculation module finds the shadow area generated by the radar emitting rays to the bounding box on the ground, and comprises the following steps: determining coordinates (x) of a sensor of a radar 1 ,y 1 ,z 1 ) (ii) a Connecting the coordinate position of the sensor with each vertex of the bounding box and extending the coordinate position to the ground, finding out the intersection point of the extension line of each connection line and the ground, wherein the coordinate of a certain vertex k of the bounding box is (x) k ,y k ,z k ) The coordinates of the intersection point corresponding to the vertex k are (x, y, z), and satisfy the following relation:
and can further obtain:
and the connection lines of the intersection points corresponding to all the vertexes form a shadow area.
Preferably, the area obtained by dividing the target and the scene in the circular area is a target sub-area, and the rest area is an environment area; the system further comprises a fifth calculation module, and the fifth calculation module is used for dividing the environment area into sub-areas in a boundary size sharing mode.
The flat scene grid model self-adaptive partition preprocessing method and the system have the following beneficial effects that: the method is suitable for a parallel accelerated large-scene self-adaptive regional decomposition preprocessing algorithm, effectively reduces the memory consumption while ensuring the calculation precision, improves the algorithm parallelism, and provides theoretical and algorithm support for continuously promoting the parallel design and development of a large-scale scene radar simulation system.
Drawings
FIG. 1 is a flow chart of radar echo simulation;
FIG. 2 is a flowchart of a first embodiment of a flat scene mesh model adaptive partition preprocessing method according to the present invention;
FIG. 3 is a schematic diagram of bounding spheres generated by an embodiment of the flat scene mesh model adaptive partition preprocessing method of the present invention;
FIG. 4 is a schematic diagram of a bounding box generated by one embodiment of the flat scene mesh model adaptive partition preprocessing method of the present invention;
FIG. 5 is a schematic diagram of region segmentation according to a first embodiment of the adaptive partitioning preprocessing method for mesh models in a flat scene;
FIG. 6 is a schematic view of a scene region decomposition of an embodiment of the adaptive partitioning preprocessing method for a mesh model of a flat scene according to the present invention;
FIG. 7 is a schematic illustration of a flat scene;
FIG. 8 is a schematic diagram of the flat scene shown in FIG. 7 being subjected to area decomposition by the flat scene mesh model adaptive partitioning preprocessing method according to the present invention;
FIG. 9 (a) is a schematic diagram of a flat scene mesh model adaptive partitioning preprocessing method according to the present invention;
FIG. 9 (b) is a schematic illustration of a flat scene mesh model adaptive partition preprocessing method not employed in the present invention;
FIG. 10 is a flowchart of a second embodiment of the adaptive partitioning preprocessing method for mesh models in flat scenes according to the present invention;
fig. 11 is a schematic region segmentation diagram of a second embodiment of the adaptive partitioning preprocessing method for a mesh model of a flat scene according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
Example one
Fig. 2 is a flowchart of a first embodiment of the flat scene mesh model adaptive partitioning preprocessing method according to the present invention, and as shown in fig. 2, the flat scene mesh model adaptive partitioning preprocessing method provided in the embodiment of the present invention includes the following steps:
step S01: determining a target within a scene;
step S02: calculating a bounding box of the target;
step S03: finding a shadow area generated by the radar transmitting rays to the bounding box on the ground;
step S04: forming a bounding sphere by taking the center of the bounding box as the center of a sphere and taking the farthest distance from the center of the sphere in a shadow area as a radius;
step S05: and calculating a circular area formed by the intersection of the enclosing ball and the ground, and segmenting the target and the scene in the circular area.
The self-adaptive partitioning preprocessing method for the flat scene grid model divides the target area into the target sub-areas independently and decomposes the target area and the background environment area so as to consider the area decomposition of the environment area independently later. The correctness of the simulation result, namely the correctness of the target simulation result is ensured during the regional decomposition. In the simulation calculation of the radar echo in the large scene, a target can form a shadow region of the target, the shadow region cannot be separated from the target during decomposition, otherwise the accuracy of a simulation result can be influenced, and the coupling scattering contribution of the environment in a region far away from the target to the target is very small and can be ignored. The invention provides a bounding box method, which integrally divides a target, a shadow and a coupling area into the same area.
FIG. 3 is a schematic diagram of bounding spheres generated by an embodiment of the flat scene mesh model adaptive partition preprocessing method of the present invention; as shown in fig. 3, the bounding volume generation method includes the steps of (1) finding a bounding volume of a target (a small sphere in the figure) as shown by a black dashed cube in the figure; (2) Finding the shadow boundary (the oblique line part at the position of the lower right corner of the black dotted cube in the figure) generated by the emission ray and the bounding box on the ground; (3) Finding the farthest distance from the center of the bounding box in the shadow area as a radius, and taking the center of the bounding box as the center of the sphere to make a sphere (shown by a dotted circle in the figure) to calculate the bounding sphere; (4) The target area can be segmented from the scene by taking the target and the scene in a circular area (shown by a curve formed by dot-dash lines in the figure) surrounding the intersection of the ball and the ground, the shadow area of the target is included, and meanwhile, enough background environment area is arranged around the target area to calculate the coupling electromagnetic scattering contribution of the environment to the target, so that the accuracy of the final simulation result is ensured.
Fig. 4 is a schematic diagram of a bounding box generated by an embodiment of the flat scene mesh model adaptive partition preprocessing method of the present invention, as shown in fig. 4, in this embodiment, a bounding box of a calculation target includes the following steps: for any point i in the target, its coordinate is (X) i ,Y i ,Z i ) For X i ,Y i ,Z i Has the following relationship that,
X min ≤X i ≤X max ；
Y min ≤Y i ≤Y max ；
Z min ≤Z i ≤Z max ；
determining a maximum coordinate point and a minimum coordinate point of the target; the coordinate of the maximum coordinate point is (X) max ,Y max ,Z max ) The coordinate of the minimum point of the coordinates is (X) min ,Y min ,Z min ) (ii) a And forming a bounding box by using the maximum coordinate point and the minimum coordinate point, wherein the bounding box is a cuboid, and the maximum coordinate point and the minimum coordinate point are two end points of a diagonal line of the cuboid.
In this embodiment, can simulate carrying the sensor on the aircraft during radar echo simulation, the position coordinate of radar sensor during simulation can obtain in the procedure, does the line and projects subaerial to each summit of every sensor position coordinate and bounding box, finds the nodical with ground, finds the produced shadow region of radar transmission ray to bounding box subaerial, includes following step: determining coordinates (x) of a sensor of a radar 1 ,y 1 ,z 1 ) (ii) a Connecting the coordinate position of the sensor with each vertex of the bounding box and extending the coordinate position to the ground, finding out the intersection point of the extending line of each connecting line and the ground, wherein the coordinate of a certain vertex k of the bounding box is (x) k ,y k ,z k ) The coordinates of the intersection point corresponding to the vertex k are (x, y, z), and the following relationship is satisfied by using a space straight line formula:
and knowing that the coordinate of the point z projected by the connecting line to the ground is 0, the solving formula of the projection points x and y can be obtained, and further the following can be obtained:
and the connection lines of the intersection points corresponding to all the vertexes form a shadow area.
FIG. 5 is a schematic diagram of region segmentation according to a first embodiment of the adaptive partitioning preprocessing method for mesh models in a flat scene; as shown in fig. 5, in this embodiment, a region obtained by dividing the target and the scene in the circular region is a target sub-region, the remaining region is an environment region, and the environment region is divided into sub-regions by a boundary size sharing method. Targets in a scene are extracted independently after the method of surrounding the ball is used, the remaining environment part still consists of a large number of surface elements, and memory and time consumption are still huge when echo simulation is directly carried out, so that regional decomposition pretreatment needs to be carried out on a background environment region.
FIG. 6 is a schematic view of a scene region decomposition of an embodiment of the adaptive partitioning preprocessing method for a mesh model of a flat scene according to the present invention; as shown in fig. 6, after the bounding sphere method is used, the target region in the scene is extracted separately, the remaining background environment region still consists of a large number of surface elements, and the memory and time consumption are still huge when echo simulation is directly performed, so that region decomposition preprocessing needs to be performed on the background environment region; considering that the background environment area is huge under the condition of a large scene, if the average grid number is used for carrying out area decomposition, the problems of difficult adaptation, complex operation and the like exist. For a flat scene, mesh subdivision is uniform under most conditions, obvious shielding and coupling relations do not exist among meshes, sub-regions of a background environment are divided in a boundary size sharing mode, and the size of each region can be approximately equal to the number of the meshes. The scene is divided into a target sub-area and a plurality of background environment sub-areas, so that the number of surface elements of each background environment sub-area is approximately close, the computing time and the memory overhead of each background environment sub-area are balanced, and the parallel computing efficiency is improved. The final decomposition effect is shown in fig. 6. Generally, after dividing the background environment area into M equal parts, the memory overhead can be approximately changed to 1/M. And performing electromagnetic calculation on the ground environment background area according to an environment area calculation strategy to obtain the background scattering characteristic. And then carrying out correlation synthesis with the scattering characteristics of the target area to obtain the scattering characteristics of the scene.
FIG. 7 is a schematic illustration of a flat scene; FIG. 8 is a schematic diagram of the flat scene shown in FIG. 7 being subjected to area decomposition by the flat scene mesh model adaptive partitioning preprocessing method according to the present invention; as shown in FIG. 7, a simple flat scene contains a complex target, the scene is square, the side length is 300m, a tank target is arranged in the center of the scene, and the maximum side length of a bounding box of the tank target is 8m. The invention can be used for carrying out regional decomposition on the flat scene shown in the figure 7, as shown in figure 8, and can be decomposed into a target sub-region and a plurality of environment sub-regions. And performing simulation calculation on the decomposed sub-regions respectively, so that the memory and time consumption of the simulation calculation can be reduced.
And (3) setting SAR echo simulation parameters according to the table 1, and uniformly dividing an environment part into a left part and a right part when a scene is processed by using a region decomposition method, thereby verifying the accuracy and the effectiveness of the method.
TABLE 1 SAR echo simulation parameter set
| Imaging mode | Bunching | Center frequency | 10GHz | Flying height | 2500m |
| Number of flight path sampling points | 300 | Number of frequency sampling points | 300 | Flying speed | 200m/s |
| Bandwidth of signal | 750MHz | Polarization mode | HH | PRF | 550Hz |
FIG. 9 (a) is a schematic diagram of a flat scene mesh model adaptive partitioning preprocessing method according to the present invention; FIG. 9 (b) is a schematic diagram of the adaptive partitioning preprocessing method for the flat scene mesh model without the invention. The original overall simulation calculation method and the method provided by the invention are used for respectively carrying out simulation calculation on the flat scene, and two simulation results are obtained and are shown in fig. 9. Simulation results show that the simulation results obtained by calculation of the two methods are completely the same, and the feasibility and the accuracy of the method are verified.
And then verifying the effectiveness of the region decomposition method, calculating by respectively using 6 different subdivision accuracies, counting the memory and time consumption under 6 conditions, and counting data to obtain a table 2.
TABLE 2 memory and time consumption under different subdivision accuracy conditions
The memory and time consumption of the simulation calculation after the direct calculation and the area decomposition are listed in the table, and the analysis shows that:
(1) The scene is divided into 1 target sub-area and 2 environment sub-areas, because the grid number of the environment sub-areas is far larger than that of the target sub-areas, the memory cost only depends on the environment sub-areas with the most grid units; the grid number of the two environment sub-areas is approximately equal, and the maximum memory in partition calculation is about half of that in direct calculation. Therefore, the number of partitions should be increased as much as possible to reduce the number of meshes in each partition.
(2) The total time of the partition calculation is equivalent to the direct calculation time, the calculation time consumption of the environment area is far larger than that of the target area, the calculation time of the environment area is closely related to the grid number, the grid spatial relationship, the calculation resolution and the like, and the total calculation time is approximately unchanged. Therefore, the respective partitions are combined after parallel computation, and ideal parallel acceleration can be achieved.
(3) Compared with the calculation results under different subdivision accuracies, the correlation between the memory overhead and the grid number is realized, and the calculation time is the minimum when the calculation time is 1000mm, because the clutter calculation unit subdivision in the clutter calculation strategy is closely related to the resolution unit, when the grid unit size and the resolution unit meet the specific multiple relation, the unit subdivision has the highest efficiency, and the grid number and the calculation time of the unit subdivision can be greatly reduced.
Example two
FIG. 10 is a flowchart of a second embodiment of the adaptive partitioning preprocessing method for mesh models in flat scenes according to the present invention; as shown in fig. 10, the method for preprocessing mesh model adaptive partitions in a flat scene provided in the embodiment of the present invention includes the following steps:
step Q01: determining all targets within a scene;
step Q02: calculating a bounding box of each target;
step Q03: finding a shadow area generated by the radar transmitting rays to any bounding box on the ground;
step Q04: for each target, taking the center of the corresponding bounding box as the center of a sphere and taking the farthest distance from the center of the sphere in the corresponding shadow area as a radius to form a bounding sphere;
step Q05: calculating a circular area formed by the intersection of each enclosing ball and the ground; regarding two or more overlapped circular areas, taking the objects corresponding to the circular areas as an integral object, calculating a bounding box, a shadow area and a bounding sphere of the integral object, calculating a circular area formed by the intersection of the bounding sphere corresponding to the integral object and the ground, and segmenting the objects and the scenes in the circular area; for circular areas that do not overlap with other circular areas, objects and scenes within the circular areas are segmented.
The second embodiment is basically the same as the first embodiment, and the same parts are not described again, except that: in a large scene, a plurality of targets often exist, when the regional decomposition preprocessing is performed, if the distance between the targets is long, the targets can be correctly segmented from the scene by using the method of the first embodiment, but when the distance between the targets is short, the problem that the shadow regions of the targets are overlapped may exist, and the targets may be mistakenly segmented by using the method of the first embodiment, so that the second embodiment of the invention provides an improved ball surrounding method for the large scene under the condition of multiple targets.
FIG. 11 is a schematic diagram of region segmentation according to a second embodiment of the adaptive partitioning preprocessing method for mesh models in a flat scene; which shows a region segmentation schematic obtained by using the improved bounding sphere method in a large scene multi-target situation. The specific method comprises the following steps: (1) Performing a bounding sphere on each target in the scene, finding a target area (namely a circular area) of each target, but not dividing; (2) After each divisible area is found, judging whether each target area is overlapped or not, if the overlapped area exists, taking the overlapped targets as a whole, carrying out bounding on the whole again, and finding a new divisible area, wherein if a plurality of targets are overlapped, the targets are classified as the whole, and the target areas which are not overlapped are not processed; (3) And (4) performing regional decomposition on each newly divided region, so as to complete regional decomposition target pretreatment of the scene with multiple targets.
According to the process, all targets are subjected to bounding sphere calculation, the targets and the peripheral area in the scene can be separated into a plurality of target areas, and independent calculation is carried out according to the target electromagnetic calculation strategy to obtain the scattering characteristics of the target areas.
In conclusion, it can be seen from the simulation result that the problem of large-scene radar echo simulation memory and large time consumption can be effectively solved by adopting the area decomposition technology, and the algorithm calculation efficiency and the calculation capability can be improved to the greatest extent by combining the unit grid optimization technology.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present 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 solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. A self-adaptive partitioning preprocessing method for a flat scene grid model is characterized by comprising the following steps:
determining a target within a scene;
calculating a bounding box of the target;
finding a shadow area on the ground generated by the radar transmitting rays to the bounding box;
taking the center of the bounding box as a sphere center, and taking the farthest distance from the sphere center in the shadow area as a radius to form a bounding sphere;
calculating a circular area formed by the intersection of the surrounding ball and the ground, and segmenting the target and the scene in the circular area;
and dividing the target and the scene in the circular area into areas which are target sub-areas, and dividing the remaining areas into environment areas by adopting a boundary size sharing mode.
2. A self-adaptive partitioning preprocessing method for a flat scene grid model is characterized by comprising the following steps:
determining all targets within a scene;
calculating a bounding box for each of the objects;
finding a shadow area on the ground generated by the radar transmitting rays to any one of the bounding boxes;
for each target, taking the center of the corresponding bounding box as a sphere center, and taking the farthest distance from the sphere center in the corresponding shadow area as a radius to form a bounding sphere;
calculating a circular area formed by the intersection of each enclosing ball and the ground; regarding two or more overlapped circular areas, taking the targets corresponding to the circular areas as a whole target, calculating a bounding box, a shadow area and a bounding sphere of the whole target, calculating a circular area formed by the intersection of the bounding sphere corresponding to the whole target and the ground, and dividing the target and the scene in the circular area; for circular areas which are not overlapped with other circular areas, dividing the targets and the scenes in the circular areas;
and dividing the target and the scene in the circular area into areas which are target sub-areas, and dividing the remaining areas into environment areas by adopting a boundary size sharing mode.
3. The adaptive partitioning preprocessing method for mesh models of flat scenes according to claim 1 or 2, characterized in that the calculating the bounding box of the target comprises the following steps:
for any point i in the target, its coordinate is
To a
Has the following relationship that,
determining a maximum coordinate point and a minimum coordinate point of the target; the coordinate of the maximum coordinate point is
The coordinate of the minimum point of the coordinates is
；
And forming the bounding box by using the maximum coordinate point and the minimum coordinate point, wherein the bounding box is a cuboid, and the maximum coordinate point and the minimum coordinate point are two end points of a diagonal line of the cuboid.
4. The flat scene mesh model adaptive partition preprocessing method according to claim 1 or 2, wherein the finding of the shadow area generated by the radar emitting ray to the bounding box on the ground comprises the following steps:
determining coordinates of a sensor of the radar
；
Connecting the coordinate position of the sensor with each vertex of the bounding box and extending the coordinate position to the ground, finding out the intersection point of the extending line of each connecting line and the ground, wherein the coordinate of a certain vertex k of the bounding box is
The coordinate of the intersection point corresponding to the vertex k is
The following relationship is satisfied:
and can further obtain:
and the connection lines of the intersection points corresponding to all the vertexes form the shadow area.
5. An adaptive partition preprocessing system for a mesh model of a flat scene, comprising:
a determination module to determine a target within a scene;
a first computation module to compute a bounding box of the target;
a second calculation module for finding on the ground a shadow region created by the radar transmitting rays to the bounding box;
a third calculation module for forming a bounding sphere with a center of the bounding box as a center of sphere and a farthest distance from the center of sphere in the shadow region as a radius;
the fourth calculation module is used for calculating a circular area formed by the intersection of the surrounding ball and the ground and segmenting the target and the scene in the circular area;
the area obtained by dividing the target and the scene in the circular area is a target sub-area, and the rest area is an environment area; the system also comprises a fifth calculation module, wherein the fifth calculation module is used for dividing the environment area into sub-areas in a boundary size sharing mode.
6. An adaptive partitioning preprocessing system for a mesh model of a flat scene, comprising:
a determination module to determine all targets within a scene;
a first computation module for computing a bounding box for each of the objects;
a second calculation module for finding on the ground a shadow region created by the radar transmitting a ray to any one of the bounding boxes;
a third calculation module, for each target, the third calculation module is configured to form a bounding sphere with a center of the bounding box corresponding to the target as a sphere center and a farthest distance from the sphere center in the shadow area corresponding to the target as a radius;
the fourth calculation module is used for calculating a circular area formed by the intersection of each enclosing ball and the ground; for two or more overlapped circular regions, the determining module takes the targets corresponding to the circular regions as a whole target, the first calculating module, the second calculating module and the third calculating module respectively calculate a bounding box, a shadow region and a bounding sphere of the whole target, and the fourth calculating module is further used for calculating a circular region formed by the intersection of the bounding sphere corresponding to the whole target and the ground and segmenting the targets and the scene in the circular region; for circular areas which are not overlapped with other circular areas, the fourth calculation module divides the targets and the scenes in the circular areas;
the region obtained by dividing the target and the scene in the circular region is a target sub-region, and the rest region is an environment region; the system also comprises a fifth calculation module, wherein the fifth calculation module is used for dividing the environment area into sub-areas in a boundary size sharing mode.
7. The flat scene mesh model adaptive partition preprocessing system according to claim 5 or 6, wherein the first computing module calculates the bounding box of the object, comprising the steps of:
for any point i in the target, its coordinates are
To a
Has the following relationship that,
determining a maximum coordinate point and a minimum coordinate point of the target; the coordinate of the maximum coordinate point is
The coordinate of the minimum point of the coordinates is
；
And forming the bounding box by using the maximum coordinate point and the minimum coordinate point, wherein the bounding box is a cuboid, and the maximum coordinate point and the minimum coordinate point are two end points of a diagonal line of the cuboid.
8. The adaptive partitioning preprocessing system for mesh models in flat scenes according to claim 5 or 6, wherein the second computing module finds the shadow area generated by the rays emitted by the radar to the bounding box on the ground, comprising the following steps:
determining coordinates of a sensor of the radar
；
Connecting the coordinate position of the sensor with each vertex of the bounding box and extending the coordinate position to the ground, finding out the intersection point of the extending line of each connecting line and the ground, wherein the coordinate of a certain vertex k of the bounding box is
The coordinate of the intersection point corresponding to the vertex k is
The following relationship is satisfied:
and can further obtain:
and the connection lines of the intersection points corresponding to all the vertexes form the shadow area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110338693.5A CN113075659B (en) | 2021-03-30 | 2021-03-30 | Self-adaptive partitioning preprocessing method and system for grid model of flat scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110338693.5A CN113075659B (en) | 2021-03-30 | 2021-03-30 | Self-adaptive partitioning preprocessing method and system for grid model of flat scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113075659A CN113075659A (en) | 2021-07-06 |
| CN113075659B true CN113075659B (en) | 2022-10-04 |
Family
ID=76611389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110338693.5A Active CN113075659B (en) | 2021-03-30 | 2021-03-30 | Self-adaptive partitioning preprocessing method and system for grid model of flat scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113075659B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105842698A (en) * | 2016-05-18 | 2016-08-10 | 西安电子科技大学 | Fast simulation method of high resolution SAR image |
| CN107315881A (en) * | 2017-06-30 | 2017-11-03 | 电子科技大学 | Half space Green's function and ray-tracing procedure for electromagnetic scattering simulation model |
| CN107831475A (en) * | 2017-11-06 | 2018-03-23 | 电子科技大学 | A kind of synthetic aperture radar Deceiving interference method based on shadow character |
| CN112083415A (en) * | 2020-10-12 | 2020-12-15 | 吉林大学 | Millimeter wave radar model target visibility judgment method based on 3D information |
| CN112199865A (en) * | 2020-11-11 | 2021-01-08 | 西南科技大学 | SBR-based ray tracking method for multilayer closed cabin, storage medium and terminal |
| CN112204624A (en) * | 2019-09-16 | 2021-01-08 | 深圳市大疆创新科技有限公司 | Method and device for automatically shearing model and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10948300B2 (en) * | 2018-12-27 | 2021-03-16 | Beijing Voyager Technology Co., Ltd. | Systems and methods for path determination |
-
2021
- 2021-03-30 CN CN202110338693.5A patent/CN113075659B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105842698A (en) * | 2016-05-18 | 2016-08-10 | 西安电子科技大学 | Fast simulation method of high resolution SAR image |
| CN107315881A (en) * | 2017-06-30 | 2017-11-03 | 电子科技大学 | Half space Green's function and ray-tracing procedure for electromagnetic scattering simulation model |
| CN107831475A (en) * | 2017-11-06 | 2018-03-23 | 电子科技大学 | A kind of synthetic aperture radar Deceiving interference method based on shadow character |
| CN112204624A (en) * | 2019-09-16 | 2021-01-08 | 深圳市大疆创新科技有限公司 | Method and device for automatically shearing model and storage medium |
| CN112083415A (en) * | 2020-10-12 | 2020-12-15 | 吉林大学 | Millimeter wave radar model target visibility judgment method based on 3D information |
| CN112199865A (en) * | 2020-11-11 | 2021-01-08 | 西南科技大学 | SBR-based ray tracking method for multilayer closed cabin, storage medium and terminal |
Non-Patent Citations (2)
| Title |
|---|
| 基于OSG的三维实时场景模拟系统的研究;张桂花等;《大众科技》;20111020(第10期);正文第80-82页 * |
| 基于多核CPU+GPU运算的电磁场高效体绘制算法研究;陈宇峰等;《计算机工程与应用》;20180323(第18期);正文第218-221页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113075659A (en) | 2021-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103760532B (en) | The SAR cheating interference method of jammer and receiver network interworking | |
| CN104865562B (en) | Identification method for radar disoperative target based on mixed model | |
| CN105911533B (en) | A kind of down-sampled fast scanning method of three-dimensional imaging based on flat scanning structure | |
| CN107783092A (en) | To rcs measurement system and method behind near field based on Chain relation | |
| CN110580742A (en) | method for achieving modeling and analysis of target electromagnetic scattering characteristics based on GPU parallel SBR | |
| CN113569958A (en) | Laser point cloud data clustering method, device, equipment and medium | |
| CN113376597A (en) | Complex terrain electromagnetic scattering rapid simulation method based on digital elevation map and GPU | |
| CN110705058A (en) | Near-field electromagnetic scattering simulation method for ultra-electric large-scale target | |
| CN113075659B (en) | Self-adaptive partitioning preprocessing method and system for grid model of flat scene | |
| Willis et al. | Hardware-accelerated SAR simulation with NVIDIA-RTX technology | |
| CN111123225B (en) | Sea background foil strip cloud scattering method based on vector transport theory | |
| CN113064160A (en) | Target positioning method under U-shaped building layout | |
| CN111830500A (en) | Radar image simulation method of sea surface ship target based on improved SBR (sequencing batch reactor) rapid imaging technology | |
| CN113805179B (en) | Three-dimensional modeling method for airborne weather radar target | |
| CN113447896B (en) | Undulating terrain SAR echo simulation method based on dynamic occlusion judgment | |
| CN111413679B (en) | Method for calculating RCS of unmanned aerial vehicle swarm | |
| Douchin et al. | SE-Workbench-RF: Performant and high-fidelity raw data generation for various radar applications | |
| CN113066161A (en) | Modeling method of urban radio wave propagation model | |
| CN111693958A (en) | Passive radar three-dimensional space direction finding method and system | |
| CN114089336A (en) | Adaptive partitioning method for complex fluctuating scene and radar echo simulation method | |
| Minjie et al. | An efficient adaptive space partitioning algorithm for electromagnetic scattering calculation of complex 3D models | |
| Tang et al. | A moving target detection method based on YOLO for dual-beam SAR | |
| CN113283526B (en) | Early warning method based on radiation source positioning, computer equipment and readable storage medium | |
| Artyushenko et al. | Modeling of reflections from the Earth's surface using the two-dimensional geometric models | |
| Cheng et al. | Research on Visualization Method for Detection Power of Guidance and Warning Radar |
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 | ||
| GR01 | Patent grant |