Method and system for data fusion based on heterogeneous computing spatial domain

Abstract

The invention provides a data fusion method and system based on a heterogeneous computing spatial domain, and the method comprises the steps: a fusion point clustering step: correcting pixel scatter points in a projection plane to a projection grid where fusion points are located through an auxiliary network, and clustering the corresponding pixel scatter points to corresponding spherical grid angular points; a fusion point data structure filling step: storing the corrected fusion point data by adopting a two-dimensional array with the length of M * N, and recording pixel scatter information or pixel scatter information and fusion point position information by array elements; and a data fusion calculation step: fusing the pixel scatter point information and the fusion point position information in the array elements of the two-digit array into required information. According to the invention, spatial domain data fusion is realized by adopting a method of combining the CPU and the FPGA, correction measures are taken for errors between a projection plane and an earth curved surface, the errors are well eliminated, the accuracy of data fusion is improved, and the method is suitable for a satellite-borne embedded system.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing technology, and in particular to a method and system for spatial domain data fusion based on heterogeneous computing. Background Art

[0002] Satellites are equipped with multiple imaging systems that scan and image the earth in different ways. Pixels imaged at the same location, different imaging devices, different bands, and different times are clustered together to fuse information and generate different data products. This is a common application scenario. The images of each imaging system are generally transmitted back to the ground, and then pixel clustering and information fusion are performed. With the continuous increase in image data and the limitation of satellite communication bandwidth, clustering and fusion need to be processed in real time on the satellite. The clustering and data fusion algorithms of existing ground systems are often not applicable to satellite-borne embedded systems. In addition, since the earth's surface is a curved surface, the clustering of equal-view or equidistant grids requires multiple calculations of trigonometric and inverse trigonometric functions, resulting in poor real-time calculations.

[0003] Patent document CN112967179A discloses a wide-width imaging payload image remapping method and system, including: inputting original image data blocks within a period of time and imaging time, satellite three-axis attitude, and satellite position information. Performing pixel-by-pixel geolocation of the original image data. Calculating the geolocation of each pixel in the undistorted image. Establishing a local projection coordinate system. Mapping the three-dimensional position coordinates of the original image and the undistorted image to a plane. Performing two-dimensional linear interpolation in the plane to obtain the remapped image. Outputting the remapped image and the longitude and latitude of each pixel.

[0004] However, patent document CN112967179A does not take into account the mapping error caused by the curvature of the earth, which results in low accuracy of the fusion calculation. Summary of the invention

[0005] In view of the defects in the prior art, the object of the present invention is to provide a method and system for data fusion based on heterogeneous computing space domain.

[0006] A method for data fusion based on heterogeneous computing space domain provided by the present invention includes:

[0007] Fusion point clustering step: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points;

[0008] The fusion point data structure filling step: using a two-dimensional array with a length of M*N to store the corrected fusion point data, the array elements record the pixel scatter information or the pixel scatter information and the fusion point position information;

[0009] Data fusion calculation step: fusing the pixel scatter information and the fusion point position information in the array elements of the two-bit group into the required information.

[0010] Preferably, the fusion points are M rows and N columns of fusion points on the spherical grid;

[0011] The fusion points constitute (M-1)*(N-1) target grids, and the corner points of each target grid are spherical grid corner points;

[0012] The auxiliary network is established on the projection plane, and is composed of M rows and N columns of fusion points, which correspond to the target grid one by one and have the same center;

[0013] The modification includes the following sub-steps:

[0014] Step S1: Divide the projection plane into an upper half and a lower half with the point sequence L as the boundary;

[0015] Step S2: for the pixel scatter points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scatter points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid;

[0016] Step S3: Determine whether the pixel scatter points are above or below the straight line connecting the fusion points. If yes, execute step S4; if no, the correction ends.

[0017] Step S4: After correcting the pixel scatter points to the target network directly above or directly below, execute step S2.

[0018] Preferably, the step S3 comprises:

[0019] Step S3.1: Calculate the position of the pixel scatter point A, the formula is as follows:

[0020] A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p )

[0021] Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points;

[0022] Step S3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0,y0) and (x1,y1).

[0023] Preferably, the data type of the array elements is the same structure, and the structure includes an integer and an array;

[0024] The integer records the number of fused scattered pixels actually clustered to the fusion point;

[0025] The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0 ,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

[0026] Preferably, the computing logic of a single fusion point uses a pipeline approach to increase processing throughput;

[0027] The calculation of multiple fusion points uses parallel hardware circuits to increase the processing speed.

[0028] A system based on heterogeneous computing space domain data fusion provided by the present invention includes:

[0029] Fusion point clustering module: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points;

[0030] Fusion point data structure filling module: uses a two-dimensional array with a length of M*N to store the corrected fusion point data, and the array elements record pixel scatter information or pixel scatter information and fusion point position information;

[0031] Data fusion calculation module: fuses the pixel scatter information and fusion point position information in the array elements of the two-bit group into the required information.

[0032] Preferably, the fusion points are M rows and N columns of fusion points on the spherical grid;

[0033] The fusion points constitute (M-1)*(N-1) target grids, and the corner points of each target grid are spherical grid corner points;

[0034] The auxiliary network is established on the projection plane, and is composed of M rows and N columns of fusion points, which correspond to the target grid one by one and coincide with the center;

[0035] The modification includes the following submodules:

[0036] Module M1: Divide the projection plane into the upper half and the lower half with the point sequence L as the boundary;

[0037] Module M2: for the pixel scatter points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scatter points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid;

[0038] Module M3: Determine whether the pixel scatter points are above or below the straight line connecting the fusion points. If yes, trigger module M4; if no, the correction ends.

[0039] Module M4: After correcting the pixel scatter points to the target network directly above or directly below, module M2 is triggered.

[0040] Preferably, the module M3 comprises:

[0041] Module M3.1: Calculate the position of pixel scatter point A, the formula is as follows:

[0042] A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p )

[0043] Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points;

[0044] Module M3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0,y0) and (x1,y1).

[0045] Preferably, the data type of the array elements is the same structure, and the structure includes an integer and an array;

[0046] The integer records the number of fused scattered pixels actually clustered to the fusion point;

[0047] The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0 ,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

[0048] Preferably, the computing logic of a single fusion point adopts a pipeline system to increase processing throughput;

[0049] The calculation of multiple fusion points uses parallel hardware circuits to increase the processing speed.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention adopts a method combining CPU and FPGA to realize spatial domain data fusion, corrects the mapping error caused by the influence of the earth's curved surface, improves the accuracy of data fusion, and is suitable for space-borne embedded systems.

[0052] 2. The clustering algorithm involved in the present invention is a memory-intensive calculation suitable for processing by a CPU processor. The processed result uses the data structure of the present invention to store necessary information, which is suitable for subsequent calculation-intensive data fusion calculations on FPGA. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0054] Figure 1 It is a schematic diagram of the working method of the present invention;

[0055] Figure 2 It is a cross-sectional relationship diagram of corner points in the present invention;

[0056] Figure 3 It is a top view of the projection plane in the present invention. DETAILED DESCRIPTION

[0057] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0058] The present invention uses a method combining CPU and FPGA to realize spatial domain data fusion. The steps executed by the CPU include: fusion point design step, projection plane establishment step, projection step, auxiliary grid establishment step, auxiliary grid clustering step, correction step, fusion point clustering step, and fusion point data structure filling step. The steps executed by the FPGA include data fusion calculation step.

[0059] According to a method for data fusion based on heterogeneous computing space domain provided by the present invention, Figure 1 As shown, including:

[0060] Fusion point clustering step: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points. The projection pixel scattered points are clustered to the fusion points in any neighborhood as needed. The auxiliary network is established on the projection plane, consisting of M rows and N columns of fusion points, which correspond to the target grid one by one and coincide with the center. The fusion points are M rows and N columns of fusion points on the spherical grid, and the fusion points constitute (M-1)*(N-1) target grids, and the corner points of each grid are the corner points of the spherical grid, such as 4 corner points of the grid, 12 corner points outside the 4 corner points, 20 corner points outside the 12 corner points, and so on. M and N are odd numbers. The correction includes the following sub-steps:

[0061] Step S1: Divide the projection plane into an upper half and a lower half using the point sequence L as a boundary.

[0062] Step S2: For the pixel scatter points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scatter points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid.

[0063] Step S3: Determine whether the pixel scatter points are above or below the straight line connecting the fusion points. If yes, execute step S4; if no, the correction ends. Step S3 includes:

[0064] Step S3.1: Calculate the position of the pixel scatter point A, the formula is as follows:

[0065] A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p )

[0066] Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points.

[0067] Step S3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0,y0) and (x1,y1).

[0068] Step S4: After correcting the pixel scatter points to the target network directly above or directly below, execute step S2.

[0069] Steps for filling the fusion point data structure: Use a two-dimensional array of length M*N to store the corrected fusion point data. The array elements record the pixel scatter information or the pixel scatter information and the fusion point position information. The data type of the array elements is the same structure, which contains an integer and an array. The integer records the number of fused scattered pixels actually clustered to the fusion point. The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0 ,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

[0070] Data fusion calculation step: The pixel scattered point information and the fusion point position information in the array element of the two-bit group are fused into the required information. The calculation logic of a single fusion point adopts the pipeline method to increase the processing throughput. The calculation of multiple fusion points adopts parallel hardware circuits to increase the processing speed.

[0071] Furthermore, the present invention also includes a fusion point design step, a projection plane establishment step, a projection step, an auxiliary grid establishment step and an auxiliary grid clustering step.

[0072] Fusion point design steps: Design M rows and N columns of fusion points on the spherical area that are applicable to the following steps. The M rows and N columns of fusion points form (M-1)*(N-1) grids. The corner points of each grid are fusion points. The corner point profile relationship diagram is as follows: Figure 3 As shown in . Each row of N fusion points is on a section that contains the center of the sphere, and the M sections corresponding to M rows intersect on the same sphere diameter. In the M sections, the distribution pattern of the N points is the same, as shown in Figure 2 As shown. The M-1 angles formed by the M sections are equal. The N-1 angles formed by the N fusion points and the center of the sphere are equal; or the N-1 angles formed by the N fusion points and a point S outside the spherical surface in the section are equal, and the distances between the S point and the first point and the Nth point among the N fusion points are equal, and the distances between the S points in different sections and the center of the sphere are equal. The spherical surface is a celestial spherical surface, and the satellite orbit flies along the column direction. The total field of view angle in the vertical track direction is determined according to the vertical track distribution range of the pixels to be clustered, and the angle between the two adjacent fusion points and the satellite is determined according to the set number of fusion points N in the vertical track direction, and all the above angles divide the total field of view angle equally. At every set time interval Δt, the fusion point data is generated, and the total field of view angle, the angles on both sides of the vertical track direction, and the number of fusion points in each generation remain unchanged. The total field of view angle in the vertical track direction is symmetrical along both sides of the track, and the range envelopes the vertical track distribution range of the pixels to be clustered. The number of fusion points M is an odd number, and the sub-satellite point is used as the fusion point.

[0073] Projection plane establishment step: The projection plane passes through the center of the sphere and is perpendicular to the line connecting the center of the sphere and the center of the grid in the sphere. According to the distribution range of the pixels to be clustered in the along-track direction, accumulate M rows of fusion point data, select the fusion point at the middle position of the M fusion points under the satellite in the along-track direction, and make a projection plane. The plane is perpendicular to the line connecting the above-selected fusion point and the center of the earth. Establish a two-dimensional coordinate system in the above-mentioned projection plane, with the along-track direction as the Y axis and the vertical track direction as the horizontal axis X. In the projection plane establishment step, the projection plane passes through the center of the earth, and the normal vector points to the side with the satellite. The two-dimensional coordinate system of the projection plane uses the center of the earth as the origin, the flight direction as the Y axis, and the Y axis rotated 90° clockwise as the X axis. N is an odd number, and N rows of fusion points envelop the distribution range of the pixels to be clustered in the along-track direction.

[0074] Projection step: project the fusion points onto the projection plane to form a grid in the projection plane. These fusion points are called projection fusion points, and the area where they are located is the projection grid area; project the pixel scattered points onto the projection plane, and these points are called projection pixel scattered points. Project the fusion points of the M rows and N columns of the grid onto the projection plane. The M rows of fusion points form (M-1)*(N-1) target grids, and record the coordinates of the fusion points in the above two-dimensional coordinate system. Project the pixels to be clustered onto the projection plane, and record the coordinates of the pixels to be clustered in the above two-dimensional coordinate system. In the projection step, the coordinates of the pixels to be clustered in the above two-dimensional coordinate system are stored in an M*N two-dimensional array. The elements of the array are structure objects that record the two-dimensional coordinates. The steps for calculating the projection coordinates in the two-dimensional coordinate system are:

[0075] Step 1: In the Earth's rotating coordinate system, calculate the coordinates of the unit vectors of the X-axis and Y-axis of the two-dimensional coordinate system, respectively, and record them as (x xn ,y xn ,z xn ) and (x yn ,y yn ,z yn ).

[0076] Step 2: Note that the coordinates of the point to be projected in the Earth's rotating coordinate system are (x r0 ,y r0 ,z r0 ), the projection point coordinates on the projection plane (x p ,y p ) is calculated as follows:

[0077] x p =x r0 x xn +y r0 y xn +z r0 z xn

[0078] y p =xr0 x yn +y r0 y yn +z r0 z yn .

[0079] Auxiliary grid establishment steps: Establish an affine auxiliary grid, the row length of the auxiliary grid is equal to the row length of the projection grid near the center of the projection grid area or the row length of the projection grid at the corresponding position near the center line of the projection grid area; the column length of the auxiliary grid is equal to the column length of the projection grid near the center of the projection grid area or the column length of the projection grid at the corresponding position near the center line of the projection grid area. Let the generation time of the fusion point of the first row be t1, the generation time of the fusion point of the Mth row be t2, and define t m =(t1+t2) / 2. The distance between adjacent fusion points in the Y-axis direction of the auxiliary grid is equal to t m -Δt / 2 time and t m +Δt / 2, the distance between the subsatellite point and the projection point on the projection plane is denoted as d. Generate t according to the method in the fusion point generation step. m The point data is fused at every moment and projected onto the auxiliary plane to form a point sequence L consisting of N points. This point sequence corresponds one-to-one with the fused point sequence of each row of the auxiliary grid, and the corresponding points have the same X coordinates. The point sequence L is stored in a one-dimensional array of length M, and the array elements are structure objects that record one-dimensional coordinates.

[0080] Auxiliary grid clustering step: cluster the projected pixel scatter points to the auxiliary grid. By calculating the ratio of the y coordinate in the projected coordinate of the pixel to d, determine which row of the grid the pixel belongs to, and by comparing the X coordinate value in the point column L, determine which column the pixel is in. The X coordinate value in the point column L is compared using a binary search algorithm.

[0081] The present invention also provides a system based on heterogeneous computing space domain data fusion. The system based on heterogeneous computing space domain data fusion can be realized by executing the process steps of the method based on heterogeneous computing space domain data fusion, that is, those skilled in the art can understand the method based on heterogeneous computing space domain data fusion as a preferred implementation of the system based on heterogeneous computing space domain data fusion.

[0082] A system based on heterogeneous computing space domain data fusion provided by the present invention includes:

[0083] Fusion point clustering module: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points. The fusion point is the M-row and N-column fusion point on the spherical grid. The fusion point constitutes (M-1)*(N-1) target grids, and the corner point of each target grid is the spherical grid corner point. The auxiliary network is established on the projection plane, consisting of M rows and N columns of fusion points, corresponding to the target grid one by one and the center coincides. The correction includes the following submodules: Module M1: Divide the projection plane into the upper half and the lower half with the point column L as the boundary. Module M2: For the pixel scattered points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scattered points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid. Module M3: Determine whether the pixel scattered points are above or below the straight line connecting the fusion points. If so, trigger module M4; if not, the correction ends. The module M3 includes:

[0084] Module M3.1: Calculate the position of pixel scatter point A, the formula is as follows:

[0085] A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p )

[0086] Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points.

[0087] Module M3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0, y0) and (x1, y1). Module M4: After correcting the pixel scatter points to the target network just above or just below, module M2 is triggered.

[0088] Fusion point data structure filling module: A two-dimensional array of length M*N is used to store the corrected fusion point data. The array elements record the pixel scatter information or the pixel scatter information and the fusion point position information. The data type of the array elements is the same structure, which contains an integer and an array. The integer records the number of fused scattered pixels actually clustered to the fusion point. The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

[0089] Data fusion calculation module: fuses the pixel scattered point information and fusion point position information in the array elements of the two-bit group into the required information. The calculation logic of a single fusion point uses a pipeline system to increase the processing throughput; the calculation of multiple fusion points uses parallel hardware circuits to increase the processing speed.

[0090] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0091] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for data fusion based on heterogeneous computing space domain, characterized in that: include: Fusion point clustering step: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points; The fusion point data structure filling step: using a two-dimensional array with a length of M*N to store the corrected fusion point data, the array elements record the pixel scatter information or the pixel scatter information and the fusion point position information; Data fusion calculation step: fusing the pixel scatter information and the fusion point position information in the array elements of the two-bit group into the required information.

2. The method for data fusion based on heterogeneous computing space domain according to claim 1 is characterized in that: The fusion points are M rows and N columns of fusion points on the spherical grid; The fusion points constitute (M-1)*(N-1) target grids, and the corner points of each target grid are spherical grid corner points; The auxiliary network is established on the projection plane, and is composed of M rows and N columns of fusion points, which correspond to the target grid one by one and coincide with the center; The modification includes the following sub-steps: Step S1: Divide the projection plane into an upper half and a lower half with the point sequence L as the boundary; Step S2: for the pixel scatter points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scatter points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid; Step S3: Determine whether the pixel scatter points are above or below the straight line connecting the fusion points. If yes, execute step S4; if no, the correction ends. Step S4: After correcting the pixel scatter points to the target network directly above or directly below, execute step S2.

3. The method for data fusion based on heterogeneous computing space domain according to claim 2 is characterized in that: The step S3 comprises: Step S3.1: Calculate the position of the pixel scatter point A, the formula is as follows: A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p ) Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points; Step S3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0,y0) and (x1,y1).

4. The method for data fusion based on heterogeneous computing space domain according to claim 1 is characterized in that: The data type of the array elements is the same structure, and the structure contains an integer and an array; The integer records the number of fused scattered pixels actually clustered to the fusion point; The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0 ,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

5. The method for data fusion based on heterogeneous computing space domain according to claim 1 is characterized in that: The computing logic of a single fusion point uses pipeline method to increase processing throughput; The calculation of multiple fusion points uses parallel hardware circuits to increase the processing speed.

6. A system based on heterogeneous computing space domain data fusion, characterized in that: include: Fusion point clustering module: Through the auxiliary network, the pixel scattered points in the projection plane are corrected to the projection grid where the fusion point is located, and the corresponding pixel scattered points are clustered to the corresponding spherical grid corner points; Fusion point data structure filling module: uses a two-dimensional array with a length of M*N to store the corrected fusion point data, and the array elements record pixel scatter information or pixel scatter information and fusion point position information; Data fusion calculation module: fuses the pixel scatter information and fusion point position information in the array elements of the two-bit group into the required information.

7. The system based on heterogeneous computing space domain data fusion according to claim 6 is characterized in that: The fusion points are M rows and N columns of fusion points on the spherical grid; The fusion points constitute (M-1)*(N-1) target grids, and the corner points of each target grid are spherical grid corner points; The auxiliary network is established on the projection plane, and is composed of M rows and N columns of fusion points, which correspond to the target grid one by one and coincide with the center; The modification includes the following submodules: Module M1: Divide the projection plane into the upper half and the lower half with the point sequence L as the boundary; Module M2: for the pixel scatter points in the upper half or the lower half, find the corresponding target network according to the auxiliary network where the pixel scatter points are located, and take out the coordinates of the corresponding upper or lower fusion points in the target grid; Module M3: Determine whether the pixel scatter points are above or below the straight line connecting the fusion points, and if so, trigger module M4; If not, the correction ends; Module M4: After correcting the pixel scatter points to the target network directly above or directly below, module M2 is triggered.

8. The system based on heterogeneous computing space domain data fusion according to claim 7 is characterized in that: The module M3 comprises: Module M3.1: Calculate the position of pixel scatter point A, the formula is as follows: A=(x0-x1)(y1-x p )-(y0-y1)(x1-x p ) Among them, (x0, y0), (x1, y1) represent the coordinates of the scattered pixels to be clustered, (x p ,y p ) represents the coordinates of the left and right fusion points; Module M3.2: Determine whether the pixel scatter point A is a negative value. If so, then (x p ,y p ) is below the line connecting (x0,y0) and (x1,y1); if not, p ,y p ) is above the line connecting (x0,y0) and (x1,y1).

9. The system based on heterogeneous computing space domain data fusion as claimed in claim 6 is characterized in that: The data type of the array elements is the same structure, and the structure contains an integer and an array; The integer records the number of fused scattered pixels actually clustered to the fusion point; The array records the scattered pixel information to be fused (x r0 ,y r0 ,z r0 ,I), where x r0 ,y r0 and z r0 is the scattered point coordinate, I is the scattered point gray value, and the scattered point pixels to be fused are filled starting from the low address of the array.

10. The system based on heterogeneous computing space domain data fusion according to claim 6 is characterized in that: The computing logic of a single fusion point uses a pipeline system to increase processing throughput; The calculation of multiple fusion points uses parallel hardware circuits to increase the processing speed.

Images