A preprocessing method for satellite observation point group mission planning

Abstract

This invention proposes a preprocessing method for satellite observation point cluster mission planning, belonging to the technical field of satellite Earth observation mission planning. This invention is used to cluster satellite observation points, dividing all satellite observation points into different regions, including the following steps: acquiring the location information of satellite observation points and setting relevant preprocessing parameters; clustering the satellite observation points according to the regional radius range of the aggregated point targets and the maximum number of aggregated point target sets, generating cluster sets; generating a convex hull set for each point set in the cluster set according to their positional relationships; and enlarging the size of each convex polygon in the convex hull set outwards according to the proportion of the outward expansion distance to obtain the final aggregated region, completing the preprocessing. This invention transforms multiple close point targets into locally clustered sets of points of interest based on the positional relationships of each point of interest, forming an efficient point cluster observation scheme, thereby improving the processing efficiency of satellite point cluster observations.

Description

Technical Field

[0001] This invention relates to a preprocessing method for satellite Earth observation point group mission planning, belonging to the technical field of satellite Earth observation mission planning. Background Technology

[0002] To improve the efficiency of satellite resource utilization, when a satellite observes a cluster of points of interest on the ground, it is necessary to preprocess these location points to form an efficient point cluster observation scheme. In this process, multiple nearby point targets can be transformed into locally clustered sets of points of interest based on their positional relationships, thereby improving the satellite's processing efficiency for point cluster observations. This is especially true when points are densely distributed, with a large number of point targets forming a group layout in a local geographical area; traditional single-point processing methods have significant room for improvement in terms of processing efficiency and resource utilization.

[0003] Furthermore, since a single point group typically contains dozens or hundreds of points of interest, using a single-point decomposition method will generate a large number of similar observation time windows. The calculation process for generating time windows is complex, and the post-merging processing is cumbersome, resulting in a serious waste of computing resources and satellite observation resources. Summary of the Invention

[0004] Based on this, this invention proposes a preprocessing method for satellite observation point cluster task planning. This method clusters observation points according to their distribution density, dividing all observation points into different regions. Furthermore, a convex hull is formed around the perimeter of each clustered region. During satellite observation, the task decomposition only needs to consider the convex hull as the entire observation object. This significantly reduces the number of observation time windows generated during satellite Earth observation, providing a simplified representation scheme for subsequent processing. It not only reduces the complexity of subsequent planning and optimization but also improves the efficiency of satellite resource utilization.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A preprocessing method for satellite observation point cluster mission planning, used to cluster satellite observation points and divide all satellite observation points into different regions, includes the following steps:

[0007] (1) Obtain the location information of satellite observation points and set relevant preprocessing parameters, including the regional radius range r of the aggregation point target, the maximum number M of the aggregation point target set, and the ratio d of the outward expansion distance, r>0, M>0, d>1;

[0008] (2) Based on the regional radius range r of the aggregation point target and the maximum number M of the aggregation point target set, the satellite observation points are clustered to generate a cluster set Z;

[0009] (3) For each set of points in the clustering set Z, generate a convex hull set C according to the positional relationship;

[0010] (4) According to the ratio d of the outward expansion distance, enlarge the size of each convex polygon in the convex hull set C outward to obtain the final aggregation area, completing the preprocessing.

[0011] Furthermore, the satellite observation point is one or more satellite observation target points with close geographical positions.

[0012] Furthermore, in step (1), let the set of all ground points to be observed be P = {P ,

[0019] }, 1 ≤ i ≤ n, where n represents the total number of points in the set. Among them, each point is represented as P i = {x i , y i}, where x i represents longitude and y i represents latitude; the clustering set is initialized to Z = {}.

[0013] Furthermore, step (2) is specifically as follows:

[0014] (2-1) Select an arbitrary point P i in the set P as the center of the circle;

[0015] (2-2) Draw a circle with radius r, and call this circle the neighborhood circle of point P i . Denote the total number of points in the neighborhood circle of P i as m;

[0016] (2-3) If m ≥ M, then call P i a core point; add all points in the neighborhood circle of the core point P i [[ID=F43]] to the clustering set Z i ; if m < M, then return to step (2-1) to select another point;

[0017] (F2-4) Select an arbitrary point in Z i , and determine whether this point is a core point. If so, add all points in the neighborhood circle of this point to the set Z<00 Fx i ; otherwise, select another point in Z i ; repeat this step until the set Z i no longer grows. At this time, Z i is the newly generated cluster; add this cluster to the clustering set Z, and update the set P = P - Z i ;

[0018] (2-5) Repeat steps (2-1) to (2-4) until no new core points can be found in the set P;

[0019] (2-6) Output cluster set Z as the clustering result, and the remaining points in set P are noise points.

[0020] Furthermore, step (3) specifically involves:

[0021] (3-1) Initialize the convex hull set of all clusters to C = {};

[0022] (3-2) Select a cluster Z from the cluster set Z. i ={P j}, Representing set Z i The total number of midpoints; cluster Z i The set of convex hull vertices is initialized to

[0023] (3-3) Z i The point with the lowest mid-latitude is denoted as P. k If points with the same latitude exist, select the point with the smallest longitude among those points. k Point P is the initial point for constructing the convex hull. k Add to collection middle;

[0024] (3-4) Change the initial point from set Z i Delete;

[0025] (3-5) Traverse set Z i The initial point and Z i Any point in the vector is formed by the initial point and Z. i When the vector formed by a certain point in the equation makes the smallest counterclockwise angle with the equator, Z is... i Add this point to the set In, and Z i Set this point as the new initial point;

[0026] (3-6) Repeat steps (3-4) to (3-5) until the convex hull is closed. For cluster Z i The convex hull;

[0027] (3-7) will Add it to the convex hull set C, and set Z i Remove from cluster set Z;

[0028] (3-8) Repeat steps (3-2) to (3-7) until the cluster set Z is empty, and output the convex hull set C.

[0029] Furthermore, step (4) specifically involves:

[0030] (4-1) Select any set from the convex hull set C. Representative set The total number of midpoints, where each point is denoted as P. s ={x s ,y s}, x s Represents longitude, y s Represents the latitude; let the enlarged set of the convex hull be C′={}, and initialize the set.

[0031] (4-2) Finding the set The center point of all points is

[0032] (4-3) Select Any point P in s As the endpoint, with P o Starting from the vector, construct a vector vector The magnitude is multiplied by d to form a new vector. Point P s 'Add to set' middle;

[0033] (4-4) Point P s From the set Delete;

[0034] (4-5) Repeat steps (4-3) to (4-4) until the set is complete. If empty, the set will be empty. Add to set C′;

[0035] (4-6) Traverse the convex hull set C until all convex hulls in the convex hull set C have been enlarged;

[0036] (4-7) Traverse set C′, for each element Connecting the vertices counterclockwise to form a closed figure, this closed figure is the corresponding cluster set Z. i The final convex hull, each of which is a final aggregation region.

[0037] Compared with the prior art, the advantages of this invention are:

[0038] 1. After processing by this method, the regional distribution of satellite observation points is more distinct, and the number of regions after division is much smaller than the number of observation points, making the optimization efficiency of satellite observation mission planning higher.

[0039] 2. This method requires fewer parameters to be set, does not require pre-determining the number of clusters or the shape of the cluster regions, has high processing efficiency, and can quickly converge even for a large number of discrete ground observation points.

[0040] In summary, this invention flexibly divides a point group into multiple convex polygonal regions based on the geographical distribution of the points, and flexibly transforms it into a unified plan for regional and scattered point observation tasks. Detailed Implementation

[0041] To make the above features and advantages of the present invention more apparent and understandable, the technical solution of the present invention will be further described below through specific embodiments.

[0042] A preprocessing method for satellite observation point group mission planning includes the following steps:

[0043] (1) Input the location information and related parameters of discrete points

[0044] Assume the set of all ground observation points is P = {P i}, 1≤i≤n, where n represents the total number of points in the set. Each point is represented as P. i ={x i ,y i}, x i Represents longitude, y i Represents latitude. Parameter r is set to any positive number, and parameter M is set to any positive integer. The set of clustering results is initialized as Z = {}.

[0045] (2) Clustering the point group

[0046] (2-1) Select any point P in set P. i The center is [the point of the circle].

[0047] (2-2) Draw a circle with radius r, and call this circle P. i The neighborhood circle of point P i The total number of points in the neighborhood circle is denoted as m.

[0048] (2-3) If m ≥ M, then P is called P i The core point is P. i All points in the neighborhood circle are added to the set Z. i .

[0049] (2-4) Select Z i Using any point in the circle as the center, repeat steps (2-2) to (2-3) until Z is reached. i Unable to find a new core point, Z at this point i For the newly generated clusters, update the clustering result set Z = Z∪Z. i Update the set of observation points P = PZ i .

[0050] (2-5) If m < M, select any remaining point in P as the center, and repeat steps (2-1) to (2-4) until no new core points can be found in set P.

[0051] (2-6) Output the clustering set Z as the clustering result, and call the remaining points in set P as noise points.

[0052] (3) Generate a convex polygon

[0053] (3-1) Select the cluster Z that needs to form a convex hull from the clustering set Z i ={P j} h Zi represents the total number of points in set Z i . Initialize the convex hull vertex set of cluster Z i to Initialize the convex hull set of all clusters to C = {}.

[0054] (3-2) Denote the point with the smallest latitude in Z i as P k . If there are the same latitude values, select the point with the smallest longitude among the same latitudes. The point P k is the initial point for constructing the convex hull. Add P k to the set

[0055] (3-3) Update the set Z i = Z i - P k .

[0056] (3-4) Traverse the set Z i , and denote the vector formed by P k and any point P i in Z j as When the counterclockwise angle between the vector and the equator is the smallest, add P j to the set C Zi . And set P j as the new initial point, and add P j to the set

[0057] (3-5) Repeat steps (3-3) to (3-4) until the convex hull is closed. At this time is the convex hull of cluster Z i .

[0058] (3-6) Update the convex hull result set Update the clustering result set Z = Z - Zi .

[0059] (3-7) Repeat steps (3-1) to (3-6) until the clustering result set Z is empty. Output the convex hull set C.

[0060] (4) Enlarge the convex polygon by a certain size

[0061] (4-1) Select any set from the convex hull set C. Representative set The total number of midpoints, where each point is denoted as P. s ={x s ,y s}, x s Represents longitude, y s Represents the latitude. Let the enlarged set of the convex hull be C′={}. Set the parameter d to any rational number greater than 1.

[0062] (4-2) Finding the set The center point of all points is

[0063] (4-3) Select Any point P in s As the endpoint, with P o Starting from the vector, construct a vector vector The magnitude is increased by a factor of d to form a new vector. P s 'Add to set'

[0064] (4-4) Update the set

[0065] (4-5) Repeat steps (4-3) to (4-4) until the set is complete. Empty. Update the set.

[0066] (4-6) Traverse the convex hull set C until the convex hulls of all clusters in C are enlarged.

[0067] (4-7) Traverse set C′ and process each of its subsets. Connecting the vertices in the diagram counterclockwise forms a closed figure, which is the corresponding cluster set Z. i The convex hull.

[0068] In existing technologies, when point distribution is relatively dense, a large number of point targets form a cluster layout in a local geographical area. Traditional single-point processing methods have significant room for improvement in terms of processing efficiency and resource utilization. Since a single point group typically contains dozens or hundreds of points of interest, using single-point decomposition will generate a large number of similar observation time windows. The calculation process for generating time windows is complex, and the post-merging processing is cumbersome, resulting in a serious waste of computing resources and satellite observation resources. To address this, this invention flexibly divides the point group into multiple convex polygon regions based on the geographical distribution relationship of multiple points, flexibly transforming it into a unified plan for regional and scattered point observation tasks.

[0069] In summary, this invention transforms multiple similar point targets into a locally clustered set of points of interest based on the positional relationship of each point of interest, forming an efficient point group observation scheme, thereby improving the satellite's processing efficiency for point group observations.

Claims

1. A preprocessing method for satellite observation point group mission planning, characterized in that, The method for clustering satellite observation points, dividing all satellite observation points into different regions, includes the following steps: (1) Obtain the location information of satellite observation points and set relevant preprocessing parameters, including the regional radius range r of the aggregation point target, the maximum number M of the aggregation point target set, and the ratio d of the outward expansion distance, where r > 0, M > 0, and d > 1; where the set of all ground observation points is assumed to be , , This represents the total number of points in the set, where each point is represented as... , Represents longitude. Represents dimension; cluster set initialized to ; (2) Based on the regional radius range r of the aggregation point target and the maximum number M of the aggregation point target set, the satellite observation points are clustered to generate cluster sets. Specifically: (2-1) Selecting a set any point in Centered on; (2-2) with A circle drawn with a radius of 1 is called a circle. The neighborhood circle of a point will The total number of the midpoints of the neighborhood circle is denoted as ; (2-3) If Then it is called As the core point; to make the core point All points in the neighborhood circle are added to the cluster set. ;if If so, return to step (2-1) and select another point; (2-4) Select For any point in the set, determine whether it is a core point. If it is, add all points in the neighborhood circle of that point to the set. In, otherwise in Choose another point; repeat this step until the set is complete. No longer growing, at this point This is the newly generated cluster; add this cluster to the cluster set Z and update the set. ; (2-5) Repeat steps (2-1) to (2-4) until the set is complete. No new core points can be found in it; (2-6) Output cluster set For clustering results, set The remaining points are noise points; (3) For cluster sets For each set of points in the array, generate a convex hull set according to their positional relationships. Specifically: (3-1) Initialize the convex hull set of all clusters to ; (3-2) From cluster sets Select a cluster , , Representative set The total number of midpoints; the cluster The set of convex hull vertices is initialized to ; (3-3) will The point with the lowest mid-latitude is denoted as If points with the same latitude exist, the point with the smallest longitude among those points at that latitude is selected. The point is the initial point for constructing the convex hull. Add to collection middle; (3-4) Change the initial point from the set Delete; (3-5) Traversing the set initial point and Any point in the vector is a vector, and when the initial point and... When the vector formed by a point in the equator makes the smallest counterclockwise angle with the equator, then... Add this point to the set In, and will Set this point as the new initial point; (3-6) Repeat steps (3-4) to (3-5) until the convex hull is closed. For clusters The convex hull; (3-7) Add it to the convex hull set C, and Remove from cluster set Z; (3-8) Repeat steps (3-2) to (3-7) until the cluster set is obtained. If empty, output the convex hull set. ; (4) Based on the ratio d of the outer expansion distance, for the convex hull set Each convex polygon in the image is enlarged outwards to obtain the final aggregated region, completing the preprocessing; specifically: (4-1) Select the convex hull set any set , , Representative set The total number of midpoints, where each point is represented as , Represents longitude. Represents latitude; set the enlarged set of the convex hull as... Initialize the collection ={}; (4-2) Find the set The center point of all points is ; (4-3) Select any point in As the endpoint, with Starting from, construct a vector ; to vector The modulus is expanded to Multiply, forming a new vector , will point Add to collection middle; (4-4) Point From the set Delete; (4-5) Repeat steps (4-3) to (4-4) until the set is complete. If empty, the set will be empty. Add to collection middle; (4-6) Traversing the convex hull set This continues until all convex hulls in the convex hull set C are magnified; (4-7) Traversing the set For each element Connect the vertices counterclockwise to form a closed figure; this closed figure is the corresponding cluster set. The final convex hull, each of which is a final aggregation region.

2. The satellite observation point group mission planning preprocessing method according to claim 1, characterized in that, The satellite observation point refers to one or more satellite observation target points that are geographically close to each other.

Images