CN111666661B

CN111666661B - Method and system for planning imaging multi-strip splicing task in single track of agile satellite

Info

Publication number: CN111666661B
Application number: CN202010437180.5A
Authority: CN
Inventors: 沈欣; 李仕学; 许俊飞; 蒋永华; 张过
Original assignee: Wuhan University WHU
Current assignee: Wuhan University WHU
Priority date: 2020-05-21
Filing date: 2020-05-21
Publication date: 2022-04-26
Anticipated expiration: 2040-05-21
Also published as: CN111666661A

Abstract

A method and a system for planning imaging multi-strip splicing tasks in a single track of an agile satellite comprise the steps of establishing a circumscribed rectangle of an area based on the principle of rotating a clamping shell aiming at an imaging task area, and obtaining a plurality of strips through segmentation; calculating the corresponding coverage rate of each strip, the midpoint coordinates of the starting edge and the ending edge of each strip, and calculating the imaging time windows of the starting end point and the ending end point of each strip; the satellite attitude motion is reduced into the plane motion of a camera pointing point, the plane motion constraint of the satellite camera pointing point in the multi-strip splicing imaging process is constructed, and the constraint condition is determined; cutting an imaging time window and normalizing the imaging time to determine a decision variable; constructing an imaging multi-strip splicing task planning mathematical model in the single track of the agile satellite, and determining the quantitative relation between a model decision variable and a target function and a constraint condition; and solving by adopting a PSO optimization algorithm to obtain an imaging multi-strip splicing task planning scheme, thereby realizing the maximum coverage of an imaging task area.

Description

Method and system for planning imaging multi-strip splicing task in single track of agile satellite

Technical Field

The invention relates to the technical field of satellite remote sensing, in particular to an agile satellite single-track in-motion imaging multi-strip splicing task planning technology, and provides a satellite imaging optimization technical scheme.

Background

Because the agile satellite has the flexibility of fast and stable imaging around any axis (pitching, rolling and yawing), the single-track multi-strip splicing imaging task can be realized. How to make a scientific and reasonable planning strategy according to task requirements and rapidly solve the problem of single-track multi-strip imaging mode scheduling in the process of high-speed operation of an in-orbit satellite to acquire images with high spatial resolution and high temporal resolution is a hot problem in the current domestic and foreign agile satellite scheduling technical research, and a single-track multi-strip splicing imaging mode is shown in fig. 4.

Different from the traditional agile satellite which implements attitude maneuver in the interval of the imaging task, the agile satellite which is imaged in the process needs to implement continuous attitude maneuver in the rolling, pitching and yawing directions continuously, so that the range and the execution efficiency of the executable task are greatly widened, and great challenges are provided for planning the satellite imaging task while diversified services are brought to users.

The existing multi-strip splicing task planning research mainly focuses on two aspects: the method comprises the steps of firstly, regional decomposition and secondly, planning model construction and solution. Typical methods of the existing region decomposition include a scene division method based on a monoscopic decomposition method, a scene division method based on a predefined reference system, and the like. The existing modeling solving aspect mainly comprises regional target dynamic decomposition and the like. The method mainly comprises the steps of decomposing a target area into independent strips with a certain overlapping degree according to the input of the position of the target area, the satellite orbit, sensor parameters and the like, considering the satellite attitude mobility as a constraint condition, and solving a model by adopting a traditional optimization algorithm such as particle swarm and the like. However, the existing method is mainly limited in that after the region is decomposed, the obtained imaging strips are all strips parallel to the satellite subsatellite point track, and the agile satellite with the in-motion imaging capability can decompose and obtain the imaging strips which are not parallel to the satellite subsatellite point track according to the shape and size of the region target, so that the existing multi-strip splicing task planning method is difficult to exert the efficiency of the in-motion imaging mode.

Disclosure of Invention

The invention provides an agile satellite single-track multi-strip splicing imaging task planning preprocessing technology based on preprocessing operations such as imaging time window cutting and normalization, non-tracking strip segmentation and the like, establishes a single-track multi-strip splicing task optimization mathematical model, and provides a technical scheme for agile satellite single-track multi-strip splicing imaging task planning.

The invention provides a method for planning a multi-strip splicing task for imaging in a single track of an agile satellite, which comprises the following steps of:

step S1, aiming at an imaging task area, firstly establishing an external rectangle of the area based on the principle of rotating a clamping shell, and then segmenting the external rectangle according to the width requirement to obtain a plurality of strips;

step S2, obtaining the coverage rate corresponding to each strip in the step S1;

step S3, finding the coordinates of the middle points of the starting edge and the ending edge of the strip obtained in the step S1, and calculating the imaging time window of the starting end point and the ending end point of each strip;

step S4, the satellite attitude motion is reduced into the plane motion of the camera pointing point, the plane motion constraint of the satellite camera pointing point in the multi-strip splicing imaging process is constructed, and the constraint condition of the multi-strip splicing imaging task planning mathematical model is determined;

step S5, cutting and imaging time normalization operation are carried out on the imaging time window obtained in the step S3, and decision variables of the multi-strip splicing task planning mathematical model are determined;

s6, constructing an imaging multi-strip splicing task planning mathematical model in the single track of the agile satellite, and determining the quantitative relation between a model decision variable and a target function and a constraint condition;

and step S7, solving by adopting a PSO optimization algorithm to obtain an imaging multi-strip splicing task planning scheme, and realizing the maximum coverage of an imaging task area.

Further, in step S1, the bounding rectangle of the region is created based on the principle of rotating the card shell to realize an optimal bounding rectangle for determining the target region, so that the number of bands for dividing the same region target is reduced, by the following method,

step S1.1, inputting n vertexes of a convex polygon P according to a clockwise sequence, and calculating end points of all four polygons, wherein the end points are marked as xminP, xmaxP, yminP and ymaxP;

s1.2, constructing four tangent lines of the P through four points, and determining two 'clamping shell' sets;

s1.3, if one or two tangent lines coincide with one edge, calculating the area of a rectangle determined by the four tangent lines, and storing the area as a current minimum value, otherwise, defining the current minimum value as infinity;

step S1.4, rotating the line clockwise until one line is superposed with one edge of the polygon;

s1.5, calculating the area of a new rectangle, comparing the area with the current minimum value, updating if the area is smaller than the current minimum value, and storing rectangle information for determining the minimum value;

step S1.6, repeating the step S1.4 and the step S1.5 until the rotation angle of the tangent is more than 90 degrees;

and S1.7, outputting the vertex coordinates of the minimum circumscribed rectangle.

In step S5, moreover, the cropping operation is performed on the imaging time window as follows,

the cutting operation needs to process the starting time and the ending time of the imaging time windows of the starting edge and the ending edge of all the strips, and the imaging time windows of two continuous points i and i +1 are respectively made to be [ T_{i_s}，T_{i_e}]And [ T_{i+1_s}，T_{i+1_e}]Wherein T is_{i_s}Represents the starting time, T, of the ith point imaging time window_{i_e}Representing the end time, T, of the ith point imaging time window_{i+1_s}Denotes the starting time, T, of the (i + 1) th imaging time window_{i+1_e}Represents the termination time of the (i + 1) th imaging time window;

firstly, carrying out a first type of cutting operation, based on the starting time of cutting adjacent time windows, sequentially starting from a first point, comparing the starting time of imaging time windows of two adjacent points, and if T exists_{i_s}＞T_{i+1_s}Then let T_{i_s}＝T_{i+1_s}Until the imaging time window starting time of all points is finished in sequence; then, the second kind of cutting operation is carried out, the end time of the adjacent time window is cut, the reverse order starts from the last point, if T exists_{i_e}＞T_{i+1_e}Then let T_{i_e}＝T_{i+1_e}And until the cutting of all the points at the end time of the imaging time window is finished in the reverse order.

In step S5, based on the clipping of the imaging time window, the imaging time of each point is compressed to the [0,1] interval according to the order of the observation points and the start time of the imaging time window of the adjacent point, and becomes a normalized time coefficient, and the elimination of the invalid search space is realized on the basis of maintaining the observation timing of the point;

the corresponding imaging instant normalization operation is implemented as follows,

1) according to the normalization coefficient s corresponding to the imaging time of the first point₁Recovering s by the following formula₁Corresponding toImaging time t₁：

t₁＝s₁*(T_{1_e}-T_{1_s})

2) According to the calculated t₁With the start time T of the imaging time window of the second point_{2_}s sequence, recovering t₂：

If t₁≤T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-T_{2_s})；

If t₁＞T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-t₁)；

3) And so on, repeatedly executing the previous substep 2), by comparing the imaging time t of the ith point_iAnd the starting time T of the (i + 1) th point imaging time window_{i+1_s}The order of the imaging time t of i +1 is recovered_i+1Until the imaging time of all points is recovered;

wherein, T_i-sDenotes the starting time, T, of the ith imaging time window_i-eIndicating the end time, t, of the ith imaging time window_iFor a random time of the i-th imaging time window, s_iIs the normalization coefficient corresponding to the ith time window, and the range is 0 to 1.

In step S6, a mathematical model is constructed with the imaging time normalization coefficients corresponding to the start and end observation times of each strip as decision variables and the satellite attitude mobility, coverage and imaging completion time as objective functions, where n is the number of strips after the region is decomposed, each strip has 2 end points, and the mathematical model is formally expressed as:

Maximize：f(s₁，s₂，s₃，......，s_2n)

Define：IFcov(s₁，s₂，s₃，......，s_2n)＞cov(s′₁，s₂′，s₃′，......，s_2n′)

THENf(s₁，s₂，s₃，......，s_2n)＞f(s₁′，s₂′，s₃′，......，s_2n′)

ELSEIFcov(s₁，s₂，s₃，......，s_2n)＝cov(s₁′，s₂′，s₃′，......，s_2n′)&&

time(s₁，s₂，s₃，......，s_2n)＜time(s₁′，s₂′，s₃′，......，s_2n′)

Subjectto：T_O-begin≥t_o-begin

v_instrip-m≤v_max(m＝1，2，3......n)

T_outstrip-p≥t_outstrip-p(p＝1，2，3......n-1)

s_i∈[0，1]，i＝1，2，......，2n

abstraction of the model's objective function into a logical expression f(s)₁，s₂，s₃，......，s_2n) It covers two criteria: coverage indicator cov(s)₁，s₂，s₃，......，s_2n) And imaging task completion time(s)₁，s₂，s₃，......，s_2n) (ii) a Wherein the corresponding normalized time coefficient is denoted as s₁，s₂，s₃，......，s_2nThe objective function needs to take the maximum value, s₁′，s₂′，s₃′，......，s_2n' is another scheme of 2n corresponding normalized time coefficients corresponding to n point targets;

the constraint conditions meet the judgment of the attitude constraint relationship, the attitude mobility of the initial section, the strip internal push-broom and the strip switching is converted into three groups of constraints, one is the initial section, and the judgment T is used for judging_O-begin≥t_o-beginWhether or not the above-mentioned conditions are satisfied,determining whether the attitude maneuver process of the satellite from the initial state to the push-broom first endpoint meets the attitude maneuver capability constraint; secondly, the strip is internally pushed and swept, and the pushing and sweeping speed v of each strip m is judged_instrip-mWhether or not v is satisfied_instrip-m≤v_maxDetermining whether the push-broom of the mth strip can be realized; thirdly, a band switching section, by judging T_outstrip-p≥t_outstrip-pDetermining whether the p +1 th stripe can be pushed and swept continuously after the p stripe pushing and sweeping is finished;

wherein, the consumption time of the random emergence from the source point to the starting point of the first stripe is T_O-beginThe shortest time is t_O-beginThe consumption time of switching random between the p-th stripe is T_outstrippThe shortest time is t_outstri_p-p，v_maxIs the maximum speed.

Furthermore, in step S7, the updating rule is reset based on the standard PSO algorithm, including introducing the center particle P_centerAt position x thereof_centerThe method is the average value of historical optimal positions of all particles of a current population, when the speed and the positions of the particles are updated, two particles are randomly selected from the population, the historical optimal values of the objective functions of the two particles are compared, and the particle with the high value is P_winParticles of low value are P_loseTo P only_losePosition x of_loseAnd velocity v_loseAnd (6) updating.

The invention provides an imaging multi-strip splicing task planning system in an agile satellite single track, which is used for realizing the imaging multi-strip splicing task planning method in the agile satellite single track.

Compared with the prior art, the method has the following advantages:

according to the method, the non-tracking segmentation is carried out on the imaging area of the agile satellite, the circumscribed rectangle of the area is established based on the principle of rotating the card shell, and then the circumscribed rectangle strip is decomposed according to the satellite width. In the construction of a task planning model, the normalization time coefficient of the observation time is used as a decision variable, and the preprocessing method of the imaging time window clipping and the imaging time normalization effectively overcomes the problem of low solving efficiency caused by directly modeling the imaging time as the decision variable.

Drawings

FIG. 1 is a schematic diagram of a minimum bounding rectangle constructed according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of an imaging time window according to an embodiment of the present invention;

fig. 3 is a schematic diagram of satellite imaging window prediction according to an embodiment of the present invention, in which fig. 3a is a schematic diagram of a satellite characteristic cone, and fig. 3b is a schematic diagram of a start point and an end point of a satellite imaging time window;

FIG. 4 is a schematic view of a prior art single-rail multi-strip imaging mode;

FIG. 5 is a simplified schematic of a single-track multiple-strip model according to an embodiment of the present invention;

fig. 6 is a schematic diagram of whether a uniform velocity segment exists in a modeling process according to an embodiment of the present invention, where fig. 6a is a schematic diagram of a case of a non-uniform velocity segment, and fig. 6b is a schematic diagram of a case of a uniform velocity segment;

FIG. 7 is a diagram illustrating the violation of dot sequence caused by random imaging moments in the prior art;

FIG. 8 is a diagram illustrating a prior art search dead space;

FIG. 9 is a diagram illustrating imaging time window cropping, in accordance with an embodiment of the present invention;

fig. 10 is a schematic diagram before and after decomposition of an experimental area according to an embodiment of the present invention, in which fig. 10a is the schematic diagram before decomposition, and fig. 10b is the schematic diagram after decomposition.

Fig. 11 is a flowchart of the steps of a method for planning the imaging multi-band splicing task in the single track of the agile satellite according to the embodiment of the invention.

Detailed Description

The technical scheme of the invention is explained in detail in the following by combining the drawings and the embodiment.

The invention provides an agile satellite single-track multi-strip splicing imaging task planning preprocessing technology based on preprocessing operations such as imaging time window cutting and normalization, non-tracking strip segmentation and the like, establishes a single-track multi-strip splicing task optimization mathematical model, provides an agile satellite single-track multi-strip splicing imaging task planning method, and improves satellite imaging efficiency.

As shown in fig. 11, an embodiment provides an agile satellite single-track multi-strip stitching imaging task planning method, including the following steps:

the agile satellite can perform attitude maneuver along three axes, and the flexibility enables the satellite to have a longer observation time window for a target and be more advantageous in single-track multi-strip splicing imaging. The single-track multi-strip splicing imaging is specific to regional targets, and because the satellite camera is limited in width, coverage of the regional targets with a large area cannot be achieved through single-strip imaging, the regional targets need to be subjected to strip division, and the regional targets are continuously imaged for multiple times by using the flexible attitude capability of an agile satellite.

When a traditional imaging divides a regional target, the direction of a division strip is required to be parallel to the track of the subsatellite point, and the method is called an along-track strip division method. The method corresponds to a passive push-broom imaging mode of an imaging satellite, and cannot well exert the flexible attitude adjustment capability of an agile satellite, so that the completion efficiency of regional target observation tasks is low. The method for dividing the observation strips along any direction is called as a non-tracking strip dividing method, can reduce the number of strips for dividing the same region target, can reduce the time consumed in the process of posture conversion in observation, and improves the observation efficiency. Therefore, the invention provides a non-tracing stripe dividing method based on a rotary card shell.

When the non-tracing-band division is carried out, the invention actually selects a plurality of rectangular bands to cover the target area, and the rectangular bands are parallel to each other. Therefore, the invention actually needs to determine an optimal circumscribed rectangle of the target area, so that the number of strips for dividing the target in the same area is reduced, the time consumed in the process of posture conversion can be reduced in observation, and the observation efficiency is improved. Therefore, the main problem when the invention performs non-tracing band segmentation is to determine the minimum bounding rectangle of the region target, and here, the invention adopts a non-tracing band segmentation method based on a rotating card shell in computer graphics, and for convenience of reference, the following description is provided.

Referring to fig. 1, there is one side of the minimum bounding rectangle of the convex polygon P coinciding with one side of the polygon P. Therefore, for any polygon, a convex hull is established first, and there is no change for the convex polygon, and for the concave polygon, the convex hull of the concave polygon is determined, and then the convex polygon is processed.

In practice, consider a convex polygon having two pairs of tangent lines tangent to four endpoints in the x and y directions. The four lines have defined a circumscribed rectangle of the polygon. However, unless the polygon has a horizontal or vertical side, the area of the rectangle cannot be included in the minimum area. However, it is possible to rotate the wire until the conditions are met. This process is the core of the following algorithm:

(1) suppose that n vertices of one convex polygon P are input in a clockwise order. The endpoints of all four polygons are calculated and are denoted xminP, xmAXP, yminP, ymaxP.

(2) Four tangents to P are constructed through the four points. They determined two "card shell" sets.

(3) If one (or two) tangent lines coincide with one edge, the area of the rectangle determined by the four tangent lines is calculated and saved as the current minimum. Otherwise define the current minimum as infinity.

(4) The lines are rotated clockwise until one of the tangent lines coincides with one of the polygon edges.

(5) The area of the new rectangle is calculated and compared to the current minimum. If the current minimum value is smaller than the current minimum value, updating and storing the rectangle information for determining the minimum value.

(6) And repeating the step 4 and the step 5 until the angle of the tangent line is rotated by more than 90 degrees.

(7) And outputting the vertex coordinates of the minimum bounding rectangle.

Obtaining a minimum external rectangle for the target area according to the external rectangle of any polygon, and then dividing the external rectangle according to a preset fixed width to obtain coordinates of four vertexes of each strip (along long edges or short edges, specifically according to actual conditions). Because the minimum circumscribed rectangle is established, the conditions of minimum number of decomposed strips and relatively less gesture maneuver consumption can be realized.

solving all the rectangular strips according to the vertex coordinates of all the rectangular strips in the step S1 to obtain a covering polygon;

and solving all the rectangular strips to obtain a covering polygon according to the vertex coordinates of all the rectangular strips in a vector polygon logical operation mode.

The unit of the vertex coordinates of the four vertexes of each rectangular strip is mm;

rounding the vertex coordinates of all the rectangular strips;

solving all the rectangular strips according to the rounded vertex coordinates of all the rectangular strips to obtain a covering polygon; the embodiment of the invention adopts Vatti algorithm to realize the logic operation of complex polygon, the Vatti algorithm is the prior art, and the invention is not described in detail;

and dividing the area of the effective coverage polygon by the area of the target area to obtain the coverage rate.

The area of the effective coverage polygon is obtained according to the following steps:

dividing the effective coverage polygon into a plurality of triangles, solving the area of each triangle one by adopting a vector product method, and summing the areas of all the triangles to obtain the area of the effective coverage polygon; wherein, the calculation formula of the area of the effective coverage polygon is as follows:

wherein x is_i、y_iIs the plane coordinate of the ith vertex, q is the number of the vertices of the effective coverage polygon, and S is the area of the effective coverage polygon, wherein i and q are positive integers.

Step S3, finding the coordinates of the middle points of the starting edge and the ending edge of the strip obtained in step S1, and finding the imaging time window of the starting end point and the ending end point of each strip;

the imaging time window forecast determines the starting point and the ending point of a time period in which the satellite can image the target in a future time period, and is the basis of the imaging task planning of the agile satellite. The imaging time window is schematically shown in fig. 2, and there are time windows 1, 2, 3 between the start (Begin) and the End (End), which form gaps 1, 2, 3, 4.

The target can be observed in the process of orbital motion of the optical imaging satellite, and the real-time maximum observable range is an area with a fixed size. The maximum observation range of a conventional non-agile satellite is determined by its field of view, and for agile satellites, the maximum observation range is related to its attitude maneuver maximum angle and field of view angle, as shown in fig. 3 a. For an object, the time period for which the object is observed may be the time window for the object.

In view of the fact that most of the existing mainstream agile remote sensing satellites adopt linear array sensors, the maximum observable range of the satellites can be described by adopting a characteristic cone, the center line of the cone points to the center of the earth from an imaging center, and the half cone angle is the maximum observable angle of the satellites. In the imaging time window, the target G is located inside the characteristic cone, namely the corresponding included angle beta is less than or equal to Kmax, otherwise, the target G is located outside the characteristic cone.

For agile satellites, the maximum view angle K of the satellite_maxFrom the instantaneous field of view (IFOV) of the sensor and the attitude maximum manoeuvrability (maximum manoeuvrability) L of the satellite_maxCo-determined, formula 2:

K_max＝L_max+IFOV/2 (2)

from the above, the method for determining the point target imaging time window comprises: when the time t1 when the point target enters the characteristic cone region and the time t2 when the point target exits the characteristic cone region are determined, the imaging time window of the point target is (t1, t2), namely, beta is not more than Kmax (t1, t 2). As can be seen from FIG. 3b, the relationship is satisfied when the target enters or exits the characteristic cone

Formula 3:

wherein the content of the first and second substances,

SO: a vector of the satellite pointing to the earth's center;

SG: a vector of satellites to ground targets;

GO: a vector of ground targets to geocentric;

beta: the included angle between the vector SO and the vector SG;

k_max: maximum viewing angle.

The point target object under study has the transformation from an inertial system (ECI) to an earth-fixed system (ECEF), wherein the transformation matrix is as follows:

r_ECEF＝W^-1(t)*R^-1(t)*^-1(t)*r_ECI (4)

in the above formula, r_ECEFIs the position of the geocentric geostationary coordinate system, r_ECIIs the position of the geocentric inertial coordinate system; q (t), W (t), R (t) respectively correspond to polar shift, autorotation and nutation conversion matrixes, the matrix element solving method can be implemented by adopting the prior art, and refer to IERS Conventions 2003 (IERS Technical Note number 32), which is not repeated in the invention.

And S4, the satellite attitude motion is reduced into the plane motion of the camera pointing point, the plane motion constraint of the satellite camera pointing point in the multi-strip splicing imaging process is constructed, and the constraint condition of the multi-strip splicing imaging task planning model is determined.

The single-track multi-strip splicing imaging mode is one of effective methods for realizing imaging in a large-area target of an agile satellite. When the agile satellite with the imaging capability in motion performs large-area target imaging, the area target is divided into a plurality of strips, and then the strips are spliced after being subjected to push-broom imaging for multiple times, so that large-area coverage can be realized, and fig. 4 is a schematic diagram of a single-track multi-strip imaging mode.

In order to simplify the complex constraint relation of angular displacement, angular velocity and angular velocity in the imaging process, the method takes the factors of satellite attitude maneuver, orbital motion, earth rotation and the like as the entry points of modeling, converts the attitude maneuver constraint of the satellite in the space into the plane motion constraint of the points, converts the angular velocity and the angular acceleration into the maximum velocity and the acceleration of the pointing point of the camera, and simplifies the constraint condition by utilizing the theory of plane motion.

As shown in fig. 5, after the polygon (shaded portion) of the target area is decomposed, three rectangular strips are obtained, single-track multi-strip coverage is realized, and the satellite adjusts the track of the pointing point of the camera (i.e. the broken line segment formed by a-B-C-D-E-F-G-H-I-J-K) through imaging in motion, so as to realize full coverage of the three strips.

In fig. 5, the point O is the starting point of the camera pointing direction, and B, C, F, G, J, K is the midpoint of each strip, which can be obtained from the polygon corner coordinates after decomposition. The broken line segments of the track of the pointing points of the camera can be divided into three types, the OAB segment is an initial segment, the BC, FG and JK segments are in-band push-scan segments, and the CDEF and GHIJ segments are band switching segments.

An assumption is made about the motion trajectory of the camera pointing point:

maximum angular velocity omega of camera pointing point_maxAnd maximum acceleration a_max: the maximum velocity v can be known from the conversion relationship between the linear velocity and the angular velocity_max＝ω_maxR, wherein R is the satellite altitude.

A push-broom section in the strip: for uniform motion, take segment BC as an example, corresponding speed

Wherein L is_BCIs the length of BC segment, t_BCTime is consumed for segment BC. Without loss of generality, let the velocity in the mth band be v_instrip-mN, n is the number of strips, if v is 1, 2, 3_instrip-m≤v_maxThen it can be done if v_instrip-m＞v_maxThen it cannot be done.

An initial stage: taking the OAB segment as an example, the OAB segment is divided into two linear motion segments OA and AB. Velocity v of starting point O ₀0, A is on the reverse extension of BC, and AB is the uniform acceleration process, speed v _A0, velocity v at point B_B＝v_BC. According to the assumptions, the shortest time for completing the OAB section under the maximum mobility of the satellite needs to be determined, if the time is less than the shortest time, the OAB section cannot be time-mobile, and if the time is more than or equal to the shortest time, the OAB section is feasible.

Firstly, the coordinate of the point A is determined, only the distance from the point A to the point B needs to be determined because the point A is on the reverse extension line of the BC, and the time is consumed in the section AB because the point AB is in uniform acceleration motion

The OA distance and the coordinates of the point A can be obtained.

Secondly, after the coordinates of the point A are determined, the distance d of the OA can be obtained_OA. The OA section may have two cases of "speed up first and then uniform speed and then speed down" or "speed up first and then speed down". If it is

At this time, no uniform speed section exists; if it is not

Both cases are shown in fig. 6a and 6b, respectively.

Third, obtaining the shortest time consumption t of OAB section_OB. If the OAB period consumption time is less than t_OBIt cannot be completed, if the OAB segment consumption time is greater than or equal to t_OBThen this can be done. Without loss of generality, let the consumption time from the source point to the starting point of the first stripe be T_O-beginThe shortest time is^t _O-beginIf T is_O-begin＜t_O-beginIt cannot be completed if T_O-begin≥t_O-beginThen this can be done.

A band switching section: taking the CDEF section as an example, the point D is on the BC extension line, the point E is on the reverse extension line of EF, the CD uniform deceleration section and the EF uniform acceleration sectionVelocity v at point D _D0, E point velocity v _E0, C, F point velocity v_C＝v_BC，v_F＝v_FGDetermined by the speed of the constant-speed push-broom section. According to the above assumptions, the shortest time for completing the CDEF segment under the maximum mobility capability of the satellite needs to be determined, if the time is less than the shortest time, it indicates that the CDEF segment cannot complete the mobility, and if the time is greater than or equal to the shortest time, it indicates that the CDEF segment is feasible.

In the first step, point coordinates are determined D, E. The method is consistent with the method for determining the coordinates of the point A, firstly, the lengths of the CD and the EF are determined, the finishing time of the CD and the EF can be determined because the maximum acceleration of the two paths is accelerated to the maximum speed, and the coordinates of the two points are obtained after the distance between the CD and the EF is obtained.

Second, determine the DE length. The DE section is similar to the AB section, and the symbols indicate similar meanings, two cases may occur: first, the slave D is first accelerated by the maximum acceleration a_maxAcceleration to maximum velocity v_maxThen, at a constant speed for a certain distance, and finally, at-a_maxUniformly decelerating to E; the second is that the process from D to E has no uniform speed process, and the maximum acceleration a is firstly applied from D_maxAcceleration to maximum velocity v_maxThen with-a_maxUniformly decelerates to E. Two cases were determined as follows:

if it is not

Then

If it is not

Then

If the CDEF period consumption time is less than t_CD+t_DE+t_EFIt cannot be completed if the CDEF period is longer than t_CD+t_DE+t_EFThen this can be done. Let the consumption time of switching random between the p-th stripes be T_outstrip-p(p 1, 2, 3.. n-1) and the shortest time is t_outstrip-pIf T is_outstrip-p＜t_outstrip-pThen it cannot be done; if T is_outstrip-p≤t_outstrip-pThen this can be done.

And step S5, performing cutting and imaging time normalization operation on the imaging time window obtained in the step S3, and determining decision variables of the multi-strip splicing task planning mathematical model.

When planning a satellite imaging task, generally, the imaging time is considered to be selected as a decision variable, a mathematical model is constructed, and an optimization algorithm is adopted to solve the mathematical model. The core of the imaging multi-band splicing task planning in motion is to further determine the starting time and the ending time of push-broom imaging of each band on the basis of band division. When the task planning mathematical model is constructed, if the observation time of the starting point (the midpoint of the starting edge) and the ending point (the midpoint of the ending edge) of each imaging strip is directly used as a decision variable, and an optimization algorithm is adopted to solve the problems that two defects exist:

first, the imaging time window due to the presence of multiple points mayThe method can have a plurality of conditions of phase separation, intersection and the like, and in the process of random initialization and operation by adopting an evolutionary algorithm, a large number of solutions which violate the imaging point sequence exist in the population, and if the solutions are not controlled, the searching cannot be carried out in a feasible solution range. As shown in FIG. 7, t₁For the first point at a random time of its imaging time window, t₂Is a random moment of the first point in its imaging time window, t₃If the observation time is directly adopted as a decision for the random time of the first point in the imaging time window, the condition of violating the point sequence occurs.

Secondly, in the process of continuously observing a plurality of points in real time by imaging in motion, the correlation of the observation time exists, and the transitivity exists due to mutual constraint. If the starting time and the ending time of the imaging time window are directly used as the imaging time to be used as the upper and lower bounds of the decision variable of the optimization model for searching, the efficiency of model solving is directly influenced. As shown in fig. 8, t₁Not only restricts the subsequent t₂And the feasible value range of (d), and imaging the subsequent point at time t₃、 t₁、t_nEtc. also has an important effect. As the number of points increases, the constraint effect is continuously transmitted, and if the imaging moment is directly used as a decision variable, a great number of invalid solutions are undoubtedly caused to occur. Particularly, in the population initialization process, if no control is added, the situation that all initial populations are invalid solutions even occurs.

In order to solve the above two problems, the present invention proposes to perform a cropping operation on the imaging time window first. The cropping operation requires processing of the start and end times of the imaging time window for the middle of the start and end edges of all the strips. Let the imaging time windows of two consecutive points i, i +1 be [ T ] respectively_{i_s}，T_{i_e}]And [ T_{i+1_s}，T_{i+1_e}]Wherein T is_{i_s}Represents the starting time, T, of the ith point imaging time window_{i_e}Representing the end time, T, of the ith point imaging time window_{i+1_s}Denotes the starting time, T, of the (i + 1) th imaging time window_{i+1_e}Indicating the termination instant of the (i + 1) th imaging time window. The basic idea of cutting is: firstly, a first type of cutting operation is carried out, namely cutting is carried out based on the starting time of adjacent time windows, the starting time of imaging time windows of two adjacent points is compared from a first point in sequence, and if T exists_{i_s}＞T_{i+1_s}Then let T_{i_s}＝T_{i+1_s}Until the imaging time window starting time of all points is finished in sequence; then, the second kind of cutting operation is carried out, the end time of the adjacent time window is cut, the reverse order starts from the last point, if T exists_{i_e}＞T_{i+1_e}Then let T_{i_e}＝T_{i+1_e}And until the cutting of all the points at the end time of the imaging time window is finished in the reverse order.

As shown in FIG. 9, after the above-mentioned cropping operation is completed, five cases may occur in the imaging time window of two adjacent points, where [ T ] is_{1_s}，T_{1_e}]、[T_{2_s}，T_{2_e}]Imaging time windows of adjacent first and second points, respectively. 1) If the starting time and the ending time of the first point time window are both earlier than the starting time of the second point time window, the cutting condition can not be triggered, and the cutting of the time window is not carried out; 2) if the first point time window is positioned between the starting time and the ending time of the second point time window at the ending time, and the starting time of the first imaging time window is not later than the starting time of the second point imaging time, the cutting condition is not established, and the cutting of the time window is not carried out; 3) if the first point time window is positioned in the starting time and the ending time of the second point time window at the ending time, and the starting time of the first point imaging time window is positioned in the second point imaging time window, executing a first type of cutting operation, and cutting the starting time of the second point imaging time window; 4) if the ending time of the first point imaging time window is not earlier than the ending time of the second imaging time and the starting time of one point imaging time window is not later than the starting time of the second imaging time, executing a second type of cutting operation and cutting the ending time of the first point imaging time window; 5) if the end of the imaging time window at the first point is not earlier than the end of the imaging time window at the second pointAnd if the starting time of the first point imaging time window is positioned in the second point imaging time window, executing a first type of cutting operation, and cutting the starting time of the second point imaging time window.

Further, the invention provides a normalization method of the imaging time. The imaging time normalization is based on the cutting of the imaging time window, the imaging time of each point is compressed to the [0,1] interval according to the sequence of the observation points and the initial time of the imaging time window of the adjacent point to form a normalization time coefficient, and the elimination of the invalid search space is realized on the basis of keeping the observation time sequence of the points.

The method for restoring the time normalization coefficient to the corresponding imaging moment is introduced as follows, and the idea is as follows:

1) firstly, according to the normalization coefficient s corresponding to the imaging time of the first point₁Recovering s by the following formula₁Corresponding imaging time t₁：

t₁＝s₁*(T_{1_e}-T_{1_s})

2) Then, based on the calculated t₁With the start time T of the imaging time window of the second point_{2_s}Restoring the sequence of t₂：

If t₁≤T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-T_{2_s})；

If t₁＞T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-t₁)；

3) Then, by analogy, repeatedly executing the previous substep 2) by comparing the imaging time t of the ith point_iAnd the starting time T of the (i + 1) th point imaging time window_{i+1_s}The order of the imaging time t of i +1 is recovered_i+1And the recovery of all the point imaging moments is completed.

As a specific example, the original imaging time window, the clipped imaging time window and the normalized selected time are sequentially arranged from left to right. Wherein s is₁＝0.05，s₂＝0.2，s₃When the imaging time of the first point is first restored to 0.25: t is t₁＝s₁*(T_{1_e}-T_{1_s}) (ii) a When restoring the imaging time window of the second point, due to t₁＞T_{2_s}，t₂＝T_{2_s}+s₂*(T_{2_e}-t₁) (ii) a When the imaging time of the third point is restored, t is caused₂≤T_{2_s}Then t is₃＝T_{3_s}+s₃*(T_{3_e}-T_{3_s})。

Through the normalization operation, a normalization coefficient s for imaging time of all points can be constructed_iMapping relation with imaging time, and comparing s_iAs a decision variable of the imaging task planning model. And in the model solving process, when the time consumption of the scheme is calculated and whether the speed and acceleration constraints are met or not is judged, the normalization coefficient is restored to the imaging moment.

through the steps S1 and S2, the division of the strips and the calculation method of the coverage rate of each strip are determined; through the steps S3 and S4, the time window constraint and mobility constraint conditions of the core in the task planning modeling process are given; in step S5, an imaging time window cropping and imaging time instant normalization factor is provided. On the basis, the step constructs a mathematical model which takes the imaging time normalization coefficient corresponding to the start and stop observation time of each strip as a decision variable, takes the satellite attitude mobility as a constraint condition and takes the coverage rate and the imaging completion time as an objective function, and the mathematical model is formally expressed as follows:

the optimization model is as follows:

Maximize：f(s₁，s₂，s₃，......，s_2n)

THENf(s₁，s₂，s₃，......，s₂n)＞f(s₁′，s₂′，s₃′，......，s_2n′)

Subjectto：T_O-begin≥t_O-begin

v_instrip-m≤v_max(m＝1，2，3......n)

T_outstrip-p≥t_outstrip-p(p＝1，2，3......n-1)

s_i∈[0，1]，i＝1，2，......，2n (6)

in the following judgment, Define represents "definition", if represents "if", then represents "then", else represents "if", and else represents "subject to" subject limited ".

Where n is the number of bands after the region is decomposed, each band has 2 endpoints (the starting edge midpoint and the ending edge midpoint), and there are 2n endpoints in total. The corresponding normalized time coefficient is denoted as s₁，s₂，s₃，......，s_2n， s₁′，s₂′，s₃′，......，s_2n' is another scheme of 2n corresponding normalized time coefficients for n point targets.

Decision variables: time normalization coefficient s corresponding to the imaging moments of the first two endpoints of each strip₁，s₂，s₃，......，s_2n。

Calculating an objective function: abstraction of the model's objective function into a logical expression f(s)₁，s₂，s₃，......，s_2n) It covers two criteria: coverage indicator cov(s)₁，s₂，s₃，......，s_2n) And imaging task completion time(s)₁，s₂，s₃，......，s_2n). The optimal strategy is to compare the coverage rates firstly, compare the coverage rates corresponding to the two task schemes, and take the scheme with the large coverage rate as the optimal strategy; if the coverage rates are the same, comparing the total consumed time, and taking the scheme with short imaging completion time as the best.

1) And (3) calculating a coverage rate index in an objective function: firstly, restoring the imaging time of each strip endpoint according to the step S5, then determining which strip imaging times can meet the gesture mobility constraint to realize coverage according to the step S4, and then calculating the coverage rate of the effectively covered strip according to the step S2;

2) the imaging task completion time can be reduced according to step S5, the imaging time of the breakpoint of each strip is obtained, and the number of the strip effectively covered is obtained according to step S4 (if it is determined that the k strips can be pushed and swept), the task completion time is t_2k。

And (3) judging whether the constraint conditions meet the requirements:

1) imaging time window constraint: through the step S1, it can be ensured that any normalized coefficient is within the imaging time window after being restored;

2) posture constraint relation judgment, namely converting the posture mobility of the initial section, the strip internal pushing and sweeping and the strip switching into three groups of constraints respectively through step S4, wherein one group is the initial section and the three groups of constraints are obtained through judgment T_O-begin≥t_O-beginWhether or not toIf so, determining whether the attitude maneuver process from the initial state to the push-broom first end point of the satellite meets the attitude maneuver capability constraint; secondly, the strip is internally pushed and swept, and the pushing and sweeping speed v of each strip m is judged_instrip-mWhether or not v is satisfied_instrip-m≤v_maxWhether the push-broom of the mth strip can be realized or not can be known; thirdly, a band switching section, by judging T_outstrip-p≥t_outstrip-pIt can be seen whether the p +1 th stripe is continuously swept after the p-th stripe is swept.

Wherein, the consumption time of the random emergence from the source point to the starting point of the first stripe is T_O-beginThe shortest time is t_O-beginThe consumption time of switching random between the p-th stripe is T_outstrip-pThe shortest time is t_outstrip-p，v_maxIs the maximum speed.

Step S7, solving by adopting an improved PSO optimization algorithm to obtain an imaging multi-strip splicing task planning scheme, and realizing the maximum coverage of an imaging task area;

in specific implementation, the solution can be realized by adopting a PSO algorithm. In consideration of the standard PSO algorithm, each particle updates the position thereof according to the individual historical extreme value and the population optimal value, and the advancing direction of the particle has great limitation. If the population optimal value is only a local optimal solution, the whole population still approaches to the solution, so that a better space cannot be explored. The fundamental reason for this stagnation is that the particles simply look at the population optimum and cannot learn to the surrounding peers; the method further provides that on the basis of a standard PSO algorithm, the updating rule is reset, and the specific method comprises the following steps:

introduction of the central particle P_centerAt position x thereof_centerIs the average value of the historical optimal positions of all the particles of the current population, when the speed and the position of the particles are updated, two particles are randomly selected in the population firstly, the historical optimal values of the fitness (objective function) of the two particles are compared, the particle with the high value is P_winParticles of low value are P_loseTo P only_losePosition x of_loseAnd velocity v_loseAnd (6) updating. The following formula:

v_lose(t+1)＝c₁v_lose(t)+c₂·(x_win(t)-x_lose(t))+c₃·(x_center(t)-x_lose(t))

x_lose(t+1)＝x_lose(t)+v_lose(t+1)(7)

where t represents the algebra of the population iteration, v_lose(t+1)、v_lose(t) represents a particle P_loseSpeed of t +1 th generation and t th generation, x_lose(t+1)、x_lose(t) represents the second particle P_losePosition of t +1 th generation, t generation, x_win(t) represents a particle P_winPosition of (a), x_center(t) represents the center particle P of the t-th generation species group_centerPosition of (a), (b) c₁、c₂、c₃Is [0,1]]The random number of (2). By updating, the inferior particle tracks the historical optimal position P of the superior particle_winAnd a central particle P_centerTo update its speed and position.

Performing multiple iterations by using the improved PSO algorithm to determine an optimal solution for the imaging time of the end point of each imaging strip, wherein the steps specifically comprise:

the first step, initializing the particles in the population: a number N of particle populations are generated, each particle representing an imaging task plan. Wherein, the particle position represents a decision variable (i.e. a time normalization coefficient for imaging an end point of each strip) of the mathematical model constructed in the step S6, an initial value is randomly valued in a value range [0,1] thereof, and the initial velocity of the particle is set to zero; generating a historical optimal particle swarm with the scale of N, and assigning the speed, the position and the fitness value of each particle to the historical optimal value of the corresponding particle; wherein N is the population scale of the particle swarm, and the value can be preset according to the requirement in specific implementation.

Secondly, according to the method of step S5, restoring the imaging scheme corresponding to each particle (i.e. the imaging time of each strip end point);

and thirdly, evaluating the fitness value of each particle in the population. The method specifically comprises the following steps: according to the method of step S4, each party is judgedWhether or not the time series of imaging the plurality of strips satisfies the condition, and for the strips satisfying the constraint condition, the imaging coverage yield cov (S) is calculated by the coverage calculation method described in the step S2₁，s₂，s₃，......，s_2n) (ii) a The task completion time calculation method described in step S6 obtains the imaging completion time (S) of the protocol₁，s₂，s₃，......，s_2n)；

Step four, randomly selecting two particles in the population according to a certain probability, comparing the fitness values of the two particles, updating the position and the speed of the particle with a lower fitness value by adopting the updating method of the formula (7), and replacing the historical optimal value corresponding to the particle with the updated particle;

and fifthly, repeatedly executing the step G to the last four steps, and sequencing the fitness values of the particle swarm formed by the G iteration to obtain the particles with the optimal fitness values in the particle swarm. Wherein G is the set population iteration number. In specific implementation, the value can be preset according to the requirement.

And sixthly, restoring the position of the optimal particle to the imaging time of each strip end point to obtain an optimal imaging task scheme.

Improved algorithm introduces P in updating speed and position of inferior particle by changing evolution strategy of particle_centerThe method enhances the exchange learning among the particles, improves the diversity of the particles and overcomes the phenomenon of prematurity caused by local extremum.

In specific implementation, the above process can adopt computer software technology to realize automatic operation process. The system device of the corresponding operation flow should also be within the protection scope of the present invention. An example is proposed, in which the satellite orbit-related parameters and the satellite attitude maneuver are shown in table 1, the region coordinates are shown in table 2, and the agile satellite performs a multi-slice stitching imaging scheme on the region shown in table 3.

TABLE 1 satellite parameters

TABLE 2 region vertex coordinates

Step a1, aiming at the target area, a minimum external rectangle of the area is generated by adopting a method of rotating the card shell, the coordinates of four vertexes of the external rectangle are respectively (30.1728, 115.409), (30.2857, 116.615), (31.1147, 116.516), (31.001, 115.3) (anticlockwise from the first vertex at the lower left corner), and the external rectangle is banded according to the width (here, the given value is 15.4 km). The divided region is decomposed into 6 strips, and the coordinates of the 6 strips are as follows: the first strip: (30.1728, 115.409), (30.2857, 116.615), (30.4239, 116.598), (30.3109, 115.391); second strip: (30.3109, 115.391), (30.4239, 116.598), (30.562, 116.582), (30.4489, 115.373); third strip: (30.4489, 115.373), (30.562, 116.582), (30.7002, 116.566), (30.5869, 115.355); fourth strip: (30.5869, 115.355), (30.7002, 116.566), (30.8384, 116.549), (30.7249, 115.337); fifth strip: (30.7249, 115.337), (30.8384, 116.549), (30.9765, 116.533), (30.863, 115.319); the sixth strip: (30.863, 115.319), (30.9765, 116.533), (31.1147, 116.516), (31.001, 115.3) (counterclockwise from the first vertex in the lower left corner). The schematic diagrams before and after the area decomposition are respectively shown in fig. 10a and fig. 10 b.

Step a2, solving the coverage rate corresponding to each strip in step a1, wherein the coverage rates corresponding to six strips are as follows: 5.7%, 19.9%, 23.3%, 21.2%, 19.7%, 10.2%.

Step a3, cutting and normalizing the imaging time window of the starting point and the ending point of the 6 strips separated from the target area.

Step a3, establishing a single-track multi-strip splicing imaging mathematical model.

Step a4, solving using a modified PSO optimization algorithm.

The imaging moments solved are shown in the following table:

TABLE 3 optimization results

The optimization result with the coverage rate of 100% is completed by using 6 strips for the area 2, and the corresponding optimization time is shown in the table above, so that the optimization effect is good.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.

Claims

1. An imaging multi-strip splicing task planning method in single track of an agile satellite is characterized by comprising the following steps:

in step S5, the cropping operation is performed on the imaging time window as follows,

the cutting operation needs to process the starting time and the ending time of the imaging time windows of the starting edge and the ending edge of all the strips, and the imaging time windows of two continuous points i and i +1 are respectively made to be [ T_{i_s},T_{i_e}]And [ T_{i+1_s},T_{i+1_e}]Wherein T is_{i_s}Represents the starting time, T, of the ith point imaging time window_{i_e}Representing the end time, T, of the ith point imaging time window_{i+1_s}Denotes the starting time, T, of the (i + 1) th imaging time window_{i+1_e}Represents the termination time of the (i + 1) th imaging time window;

firstly, carrying out a first type of cutting operation, based on the starting time of cutting adjacent time windows, sequentially starting from a first point, comparing the starting time of imaging time windows of two adjacent points, and if T exists_{i_s}>T_{i+1_s}Then let T_{i_s}＝T_{i+1_s}Until the imaging time window starting time of all points is finished in sequence; then, the second kind of cutting operation is carried out, the end time of the adjacent time window is cut, the reverse order starts from the last point, if T exists_{i_e}>T_{i+1_e}Then let T_{i_e}＝T_{i+1_e}Until the cutting of all the point imaging time window ending time is finished in the reverse order;

in step S5, based on the clipping of the imaging time window, the imaging time of each point is compressed into the [0,1] interval to become a normalized time coefficient according to the order of the observation points and the initial time of the imaging time window of the adjacent point, and the elimination of the invalid search space is realized on the basis of maintaining the observation time sequence of the points;

1) according to the normalization coefficient s corresponding to the imaging time of the first point₁Recovering s by the following formula₁Corresponding imaging time t₁：

t₁＝s₁*(T_{1_e}-T_{1_s})

2) According to the calculated t₁With the start time T of the imaging time window of the second point_{2_s}Restoring the sequence of t₂：

If t₁≤T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-T_{2_s})；

If t₁>T_{2_s}If so, let t₂＝T_{2_s}+s₂*(T_{2_e}-t₁)；

wherein, T_i-sDenotes the starting time, T, of the ith imaging time window_i-eIndicating the end time, t, of the ith imaging time window_iFor a random time of the i-th imaging time window, s_iIs the normalization coefficient corresponding to the ith time window, and the range is 0 to 1;

in step S6, a mathematical model is constructed that takes the imaging time normalization coefficients corresponding to the start and end observation times of each strip as a decision variable, takes the satellite attitude mobility, the coverage rate, and the imaging completion time as objective functions, and sets n as the number of strips after the region is decomposed, each strip has 2 end points, and the mathematical model is formally expressed as:

Maximize：f(s₁,s₂,s₃,......,s_2n)

Define：IFcov(s₁,s₂,s₃,......,s_2n)＞cov(s₁′,s₂′,s₃′,......,s_2n′)

THEN f(s₁,s₂,s₃,......,s_2n)＞f(s₁′,s₂′,s₃′,......,s_2n′)

ELSE IFcov(s₁,s₂,s₃,......,s_2n)＝cov(s₁′,s₂′,s₃′,......,s_2n′)&&

time(s₁,s₂,s₃,......,s_2n)＜time(s₁′,s₂′,s₃′,......,s_2n′)

THENf(s₁,s₂,s₃,......,s_2n)＞f(s₁′,s₂′,s₃′,......,s_2n′)

Subjectto：T_O-begin≥t_O-begin

v_instrip-m≤v_max m＝1，2，3……n

T_outstrip-p≥t_outstrip-p p＝1，2，3……n-1

s_i∈[0,1]，i＝1,2,......,2n

wherein Maximize represents the maximum value, the first row indicates that the objective function needs to obtain the maximum value, in the subsequent judgment, Define represents "Define", IF represents "IF", THEN represents "THEN", ELSE IF represents "ELSE IF", and Subjectto represents "subject limited";

abstraction of the model's objective function into a logical expression f(s)₁,s₂,s₃,......,s_2n) It covers two criteria: coverage indicator cov(s)₁,s₂,s₃,......,s_2n) And imaging task completion time(s)₁,s₂,s₃,......,s_2n) (ii) a Wherein the corresponding normalized time coefficient is denoted as s₁,s₂,s₃,......,s_2nObjective function ofIt is necessary to obtain a maximum value, s₁′,s₂′,s₃′,......,s_2n' is another scheme of 2n corresponding normalized time coefficients corresponding to n point targets;

the constraint conditions meet the judgment of the attitude constraint relationship, the attitude mobility of the initial section, the strip internal push-broom and the strip switching is converted into three groups of constraints, one is the initial section, and the judgment T is used for judging_O-begin≥t_O-beginWhether the attitude maneuver capability constraint is met or not is determined, and whether the attitude maneuver process from the initial state to the push-broom first end point of the satellite meets the attitude maneuver capability constraint or not is determined; secondly, the strip is internally pushed and swept, and the pushing and sweeping speed v of each strip m is judged_instrip-mWhether or not v is satisfied_instrip-m≤v_maxDetermining whether the push-broom of the mth strip can be realized; thirdly, a band switching section, by judging T_outstrip-p≥t_outstrip-pDetermining whether the p +1 th stripe can be pushed and swept continuously after the p stripe pushing and sweeping is finished;

wherein, the consumption time of the random emergence from the source point to the starting point of the first stripe is T_O-beginThe shortest time is t_O-beginThe consumption time of switching random between the p-th stripe is T_outstrip-pThe shortest time is t_outstrip-p，v_maxIs the maximum speed;

2. The method for planning the imaging multi-band splicing task in the single track of the agile satellite according to claim 1, wherein the method comprises the following steps: in step S1, the bounding rectangle of the region is established based on the principle of rotating the card shell to realize an optimal bounding rectangle of the target region, so that the number of bands for dividing the same region target is reduced, which is realized as follows,

3. The method for planning the imaging multi-band splicing task in the single track of the agile satellite according to the claim 1 or 2, wherein: in step S7, based on the standard PSO algorithm, the updating rule is reset, including introducing the central particle P_centerAt position x thereof_centerThe method is the average value of historical optimal positions of all particles of a current population, when the speed and the positions of the particles are updated, two particles are randomly selected from the population, the historical optimal values of the objective functions of the two particles are compared, and the particle with the high value is P_winParticles of low value are P_loseTo P only_losePosition x of_loseAnd velocity v_loseAnd (6) updating.

4. An imaging multi-strip splicing task planning system in single track of an agile satellite is characterized in that: the method for realizing the imaging multi-strip splicing task planning in the single track of the agile satellite according to any one of claims 1 to 3.