Multi-sensor data association method and device, and electronic equipment

Abstract

The present invention discloses a multi-sensor data association method, device, and electronic device. The method includes: obtaining a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle; performing a preset matrix block algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an association cost matrix; according to the association cost matrix, using a minimum cost matching algorithm to obtain an association matching relationship between the fusion target and the target measurement; based on the association matching relationship, using the target measurement list received in the current cycle to update the fusion target list. The present invention solves the technical problem of high computational time complexity in the method of fusing the perception results of each sensor in the related art.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to a multi-sensor data association method and device, and electronic equipment. Background Art

[0002] For autonomous driving, the three major functional modules are generally composed of perception, decision-making and control. As the basis and premise of the three major functional modules, the perception module needs to achieve accurate perception of the surrounding environment. However, due to the limitations of sensor technology principles, different types of sensors perceive different target attributes and perception ranges. Different sensors of the same type have different perception areas due to different installation locations. Therefore, it is necessary to fuse the perception results of each sensor to obtain more accurate information about surrounding targets.

[0003] In related technologies, from the perspective of fusion objects, sensor fusion includes fusion at the data level, feature level, and target level; from the perspective of implementation paths, sensor fusion includes two methods: association first, then fusion, and direct fusion. The former mainly uses the method of global nearest neighbor (GNN) association plus Kalman filtering or DS evidence theory fusion, and the latter is mainly based on random finite set (RFS) fusion. Due to the incompleteness of RFS theory and the difficulty of implementation, the first fusion method - GNN method is often used in practical applications.

[0004] The implementation of the first fusion method mainly adopts the target-level post-fusion method, which generally includes two steps: target association and target fusion. The target association mainly adopts the global nearest neighbor (GNN) method based on Hungarian matching. However, the GNN method needs to calculate the association cost matrix between the target measurement and the fusion target in advance, and set a maximum matching distance to verify the validity of the Hungarian matching result. Assuming that the number of sensors is s, the number of target measurements of each sensor is m, and the number of fusion targets maintained by the fusion algorithm is n, the time complexity of calculating the association cost matrix is ​​O(s×m×n). When the number of targets is large or the number of sensors is large, the time complexity of the calculation is high, and it is difficult to implement in scenarios with high real-time requirements.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present invention provide a multi-sensor data association method and device, and an electronic device to at least solve the technical problem of high computational time complexity when fusing the perception results of each sensor in the related art.

[0007] According to one aspect of an embodiment of the present invention, there is provided a multi-sensor data association method, including: obtaining a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle; performing a preset matrix blocking algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an association cost matrix; according to the association cost matrix, using a minimum cost matching algorithm to obtain an association matching relationship between the fusion target and the target measurement; based on the association matching relationship, using the target measurement list received in the current cycle to update the fusion target list.

[0008] Optionally, a preset matrix blocking algorithm is executed on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate a step of an associated cost matrix, including: constructing a data fusion point set; obtaining a first position coordinate point of each of the fusion targets in the fusion target list and a second position coordinate point of each of the target measurements in the target measurement list; placing the first position coordinate point and the second position coordinate point into the data fusion point set in sequence; using a preset region partitioning algorithm to perform region partitioning on each position coordinate point in the data fusion point set to generate a region partitioning set; based on the region partitioning set, updating corresponding element values ​​in the associated cost matrix to obtain the associated cost matrix.

[0009] Optionally, before constructing the data fusion point set, the method further includes: constructing an initial state association cost matrix based on the target measurement list and the fusion target list; and initializing the element value of the association cost matrix to the maximum matching distance.

[0010] Optionally, a preset region division algorithm is used to divide regions for each position coordinate point in the data fusion point set, and the steps of generating a region division set include: respectively setting the length of the overlapping area in the first direction as one, the length of the overlapping area in the second direction as two, the number of regions to be divided in the first direction as one, and the number of regions to be divided in the second direction as two; calculating the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction in the data fusion point set; based on the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction, as well as the length of the overlapping area in the first direction as one, and the length of the overlapping area in the second direction as two, calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, and determining the region division set.

[0011] Optionally, after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, it also includes: calculating the length one of each sub-area in the first direction based on the number one of areas to be divided in the first direction, the minimum and maximum values ​​of the area position points after the overlapping area width is expanded in the first direction, and the length of the overlapping area in the first direction; calculating the length two of each sub-area in the second direction based on the number two of areas to be divided in the second direction, the minimum and maximum values ​​of the area position points after the overlapping area width is expanded in the second direction, and the length of the overlapping area in the second direction.

[0012] Optionally, after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, it also includes: defining an index one of a forward partition in the first direction, an index two of a reverse partition in the first direction, an index three of a forward partition in the second direction, and an index four of a reverse partition in the second direction; if there are unprocessed position points in the data fusion point set, determining the first direction coordinates and the second direction coordinates of the unprocessed position points; calculating the index one of the unprocessed position points based on the first direction coordinates of the unprocessed position points, the minimum value of all position points in the first direction, the length of the overlapping area in the first direction, and the length one of the sub-area in the first direction; and calculating the index one of the unprocessed position points based on the number one of the areas to be divided in the first direction, the number of all position points, and the length of the sub-area in the first direction. The second index of the unprocessed position point is calculated based on the maximum value of the point in the first direction, the first direction coordinate of the unprocessed position point, the length of the overlapping area in the first direction, and the first length of the sub-area in the first direction; the third index of the unprocessed position point is calculated based on the second direction coordinate of the unprocessed position point, the minimum value of all position points in the second direction, the length of the overlapping area in the second direction, and the second length of the sub-area in the second direction; the fourth index of the unprocessed position point is calculated based on the second number of areas to be divided in the second direction, the maximum value of all position points in the second direction, the second direction coordinate of the unprocessed position point, the length of the overlapping area in the second direction, and the second length of the sub-area in the second direction; determine the position point index of the unprocessed position point.

[0013] Optionally, before constructing the data fusion point set, it also includes: determining an index partition value of the fusion target and the target measurement based on the total number of the fusion targets, wherein the index partition value is used to divide the fusion target and the target measurement in the data fusion point set.

[0014] Optionally, based on the region partition set, the corresponding element values ​​in the associated cost matrix are updated to obtain the step of the associated cost matrix, including: determining whether there is an unprocessed region in the region partition set; if there is an unprocessed region in the region partition set, extracting a region to be processed from the region partition set; obtaining an unprocessed position point index from the region to be processed; if the position point index is less than the index partition value, storing the fusion target index indicated by the position point index in the fusion target index list corresponding to the region to be processed; if the position point index is greater than or equal to the index partition value, storing the target measurement index indicated by the position point index in the target measurement index list corresponding to the region to be processed; and calculating the distance between the fusion target and the target measurement in the region to be processed according to the fusion target index list and the target measurement index list to update the corresponding elements in the associated cost matrix to obtain the associated cost matrix.

[0015] According to another aspect of an embodiment of the present invention, a multi-sensor data association device is also provided, including: an acquisition unit, used to acquire a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle; a generation unit, used to execute a preset matrix blocking algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an association cost matrix; a determination unit, used to obtain an association matching relationship between the fusion target and the target measurement according to the association cost matrix using a minimum cost matching algorithm; and an update unit, used to update the fusion target list based on the association matching relationship using the target measurement list received in the current cycle.

[0016] Optionally, the generation unit includes: a first construction subunit, used to construct a data fusion point set; a first acquisition subunit, used to acquire the first position coordinate point of each fusion target in the fusion target list and the second position coordinate point of each target measurement in the target measurement list; a storage subunit, used to put the first position coordinate point and the second position coordinate point into the data fusion point set in sequence; a first generation subunit, used to use a preset area partitioning algorithm to perform area partitioning on each position coordinate point in the data fusion point set to generate an area partitioning set; an update subunit, used to update the corresponding element values ​​in the association cost matrix based on the area partitioning set to obtain the association cost matrix.

[0017] Optionally, the generation unit also includes: a second construction subunit, used to construct an initial state association cost matrix based on the target measurement list and the fusion target list before constructing the data fusion point set; a first processing subunit, used to initialize the element value of the association cost matrix to the maximum matching distance.

[0018] Optionally, the first generating subunit includes: a setting module, used to respectively set the length of the overlapping area in the first direction, one, the length of the overlapping area in the second direction, the number of areas to be divided in the first direction, and the number of areas to be divided in the second direction; a first calculation module, used to calculate the minimum and maximum values ​​of all position coordinate points in all data fusion point sets in the first direction and the second direction; a second calculation module, used to calculate the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width based on the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction, as well as the length of the overlapping area in the first direction, one, and the length of the overlapping area in the second direction, so as to determine the area division set.

[0019] Optionally, the first generating subunit also includes: a third calculation module, which is used to calculate the length one of each sub-region in the first direction based on the number one of regions to be divided in the first direction, the minimum and maximum values ​​of the region position points after the width of the overlapping area is expanded in the first direction, and the length of the overlapping area in the first direction, after calculating the boundaries of the region after the width of the overlapping area is expanded around the region determined by the position points in the data fusion point set; and a fourth calculation module, which calculates the length two of each sub-region in the second direction based on the number two of regions to be divided in the second direction, the minimum and maximum values ​​of the region position points after the width of the overlapping area is expanded in the second direction, and the length of the overlapping area in the second direction.

[0020] Optionally, the first generating subunit further includes: a definition module, used to define index one of the forward partition in the first direction, index two of the reverse partition in the first direction, index three of the forward partition in the second direction, and index four of the reverse partition in the second direction after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width around each other; a first determination module, for determining the first direction coordinates and second direction coordinates of the unprocessed position points if there are unprocessed position points in the data fusion point set; a fifth calculation module, for calculating the index one of the unprocessed position points based on the first direction coordinates of the unprocessed position points, the minimum value of all position points in the first direction, the length of the overlapping area in the first direction, and the length one of the sub-area in the first direction; a sixth calculation module, for calculating the number of areas to be divided in the first direction. 1. Calculate the index 2 of the unprocessed position point based on the maximum value of all position points in the first direction, the first direction coordinates of the unprocessed position point, the length of the overlapping area in the first direction, and the length 1 of the sub-area in the first direction; the seventh calculation module is used to calculate the index 3 of the unprocessed position point based on the second direction coordinates of the unprocessed position point, the minimum value of all position points in the second direction, the length of the overlapping area in the second direction, and the length 2 of the sub-area in the second direction; the eighth calculation module is used to calculate the index 4 of the unprocessed position point based on the number 2 of areas to be divided in the second direction, the maximum value of all position points in the second direction, the second direction coordinates of the unprocessed position point, the length of the overlapping area in the second direction, and the length 2 of the sub-area in the second direction; the second determination module is used to determine the position point index of the unprocessed position point.

[0021] Optionally, the generating unit further includes: a determining subunit, used to determine, before constructing the data fusion point set, an index partitioning value of the fusion target and the target measurement based on the total number of the fusion targets, wherein the index partitioning value is used to split the fusion target and the target measurement in the data fusion point set.

[0022] Optionally, the updating subunit includes: a judgment module, judging whether there is an unprocessed area in the area division set; a first processing module, extracting a to-be-processed area from the area division set if there is an unprocessed area in the area division set; obtaining an unprocessed position point index from the to-be-processed area; a second processing module, storing the fusion target index indicated by the position point index into a fusion target index list corresponding to the to-be-processed area if the position point index is less than the index division value; a third processing module, storing the target measurement index indicated by the position point index into a target measurement index list corresponding to the to-be-processed area if the position point index is greater than or equal to the index division value; and a ninth calculation module, calculating the distance between the fusion target and the target measurement in the to-be-processed area according to the fusion target index list and the target measurement index list, so as to update the corresponding elements in the association cost matrix and obtain the association cost matrix.

[0023] In the present invention, the fusion target list maintained in the system of the previous cycle and the target measurement list received in the current cycle are first obtained, wherein each fusion target in the fusion target list represents an actual perception object, and the target measurement list contains the measurement information of all actual objects collected by multiple sensors in the current cycle, and then the preset matrix block algorithm is performed on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an association cost matrix, and then the association matching relationship between the fusion target and the target measurement is obtained according to the association cost matrix, and finally, based on the association matching relationship, the target measurement list received in the current cycle is used to update the fusion target list. In the present invention, a regional division method is introduced, and the association relationship between the fusion target and the target measurement is obtained by the association cost matrix and the minimum matching algorithm, so as to achieve the purpose of reducing the time complexity of calculating the association cost matrix, thereby achieving the technical effect of improving the efficiency of the data association algorithm, and then solving the technical problem of high computational time complexity when fusing the perception results of each sensor in the related technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0025] Figure 1 It is an optional schematic diagram of information after using a sensor to sense an object in the prior art;

[0026] Figure 2 is a schematic diagram of an optional regional division of collected data information according to the present invention;

[0027] Figure 3 is a schematic diagram of another optional regional division of a fusion target corresponding to a target point according to an embodiment of the present invention;

[0028] Figure 4 is a flow chart of an optional multi-sensor data association method according to an embodiment of the present invention;

[0029] Figure 5 is a flow chart of an optional data association algorithm based on region division according to an embodiment of the present invention;

[0030] Figure 6 is a flow chart of an optional cost matrix block calculation algorithm according to an embodiment of the present invention;

[0031] Figure 7 is a flow chart of an optional region partitioning algorithm according to an embodiment of the present invention;

[0032] Figure 8 is a schematic diagram of an optional multi-sensor data association device according to an embodiment of the present invention;

[0033] Fig. 9 It is a hardware structure block diagram of an electronic device (or mobile device) according to a data query method based on a blockchain network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0035] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0036] For autonomous driving, perception, decision-making and control constitute the three major functional modules. As the basis and premise of the three functional modules, the perception module needs to achieve accurate perception of the surrounding environment. Limited by the principles of sensor technology, different types of sensors perceive different target attributes and perception ranges. Different sensors of the same type have different perception areas due to different installation locations. Therefore, it is necessary to fuse the perception results of each sensor.

[0037] Figure 1 It is an optional schematic diagram of information after using a sensor to sense an object in the prior art, such as Figure 1 As shown, black and white (for illustration only; in the actual perception process, other color identifiers, such as yellow and orange, can also be used) represent the target points corresponding to the target measurements from the sensor and the fused targets of the fusion algorithm. Assuming that the number of target measurements is m and the number of fused targets is n, if the associated cost matrix of the target measurement and the fused target is calculated directly, the time complexity is O(m×n). Figure 1 The solution in has the technical problem of high computational time complexity when fusing the perception results of each sensor.

[0038] In order to solve the problem of high computational time complexity in fusing the sensor's perception results, this application introduces a region division scheme.

[0039] The present invention can be applied to various autonomous driving systems, software and products. It can perform correlation normalization on the object data sensed by multiple sensors, combine the fusion targets in the historical process and the correlation matching relationship between the target measurements obtained after the multiple sensors currently sense and collect the actual objects, and update the fusion target list in real time, thereby providing corresponding data support for the subsequent display and delineation of the fusion targets.

[0040] Figure 2 is a schematic diagram of an optional regional division of collected data information according to the present invention, such as Figure 2 As shown, it can be Figure 1 The entire area in is divided into Figure 2 The four regions shown in the figure can significantly reduce the computational time complexity when the targets are evenly distributed in space. The time complexity is reduced to O(4×(m / 4)×(n / 4)), that is, O((m×n) / 4). When divided into k regions, the time complexity is reduced to O((m×n) / k). If we use Figure 2 The area division method in may divide the target measurement and fusion target that originally correspond to the same real target into two different areas, resulting in that the two target measurement and fusion targets cannot be associated with each other because the distance between them becomes infinite.

[0041] In order to address the problem that the target measurement and fusion targets cannot be associated because the distance between them becomes infinite, the embodiment of the present invention further introduces a new area division method. Figure 3 is a schematic diagram of another optional regional division of the fusion target corresponding to the target point according to an embodiment of the present invention, such as Figure 3 As shown in FIG. 1 , the region division method used in the present invention is schematically illustrated. When introducing the region division, a concept of an overlapping region is introduced. That is, when performing region division, there is a certain intersection between different sub-regions. The target measurement and fusion target located in the overlapping region belong to each sub-region where the overlapping region is located, thereby solving the problem of Figure 2 In order to ensure that there is no associated matching relationship between target measurements and fused targets belonging to different sub-regions and not located in the overlapping area, the width of the overlapping area should not be less than the maximum matching distance. Since the overlapping area is very small compared to the sub-region, the time complexity can be approximately equal to that of the regional division method without overlapping areas.

[0042] By adopting a new region partitioning method, the time complexity of calculating the association cost matrix in the data association algorithm is reduced. When performing region partitioning, a region partitioning method based on overlapping areas is adopted to avoid missed associations that may be introduced by region partitioning. When generating partitioned areas with overlapping areas, a region partitioning algorithm that combines forward partitioning with reverse partitioning is adopted.

[0043] The present application is described in detail below in conjunction with various embodiments.

[0044] Embodiment 1

[0045] According to an embodiment of the present invention, a method embodiment of multi-sensor data association is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0046] Figure 4 is a flow chart of an optional multi-sensor data association method according to an embodiment of the present invention. Figure 4 As shown, the method comprises the following steps:

[0047] Step S401, obtaining a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle;

[0048] Step S402, executing a preset matrix block algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate a correlation cost matrix;

[0049] Step S403, according to the association cost matrix, a minimum cost matching algorithm is used to obtain an association matching relationship between the fusion target and the target measurement;

[0050] Step S404: based on the association matching relationship, the target measurement list received in the current period is used to update the fusion target list.

[0051] Through the above steps, the fusion target list maintained in the system in the previous cycle and the target measurement list received in the current cycle are first obtained, wherein each fusion target in the fusion target list represents an actual perception object, and the target measurement list contains the measurement information of all actual objects collected by multiple sensors in the current cycle, and then the preset matrix block algorithm is performed on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an association cost matrix, and then the association matching relationship between the fusion target and the target measurement is obtained according to the association cost matrix, and finally, based on the association matching relationship, the target measurement list received in the current cycle is used to update the fusion target list. In this embodiment, a regional division method is introduced, and the association relationship between the fusion target and the target measurement is obtained by the association cost matrix and the minimum matching algorithm, so as to achieve the purpose of reducing the time complexity of calculating the association cost matrix, thereby achieving the technical effect of improving the efficiency of the data association algorithm, and then solving the technical problem of high computational time complexity when fusing the perception results of each sensor in the related technology.

[0052] The embodiments of the present invention are described in detail below in combination with the above-mentioned implementation steps.

[0053] Step S401, obtaining a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle.

[0054] In this embodiment, the fusion targets included in the fusion target list refer to: target objects obtained after the information fusion of sensors has been completed in the historical process, such as vehicles on the road, trees beside the road, buildings beside the road, etc.; and the target measurements in the target measurement list refer to: information of each actual object object collected by each sensor in the current time period, using target measurement to characterize the actual object. Since different or multiple sensors may collect information on the same object or the same object, and the collection angles and collection times are different, the measurement information of the actual objects collected is also different. Therefore, it is necessary to first associate the objects and then fuse them into the previous fusion target list to facilitate subsequent autonomous driving planning, road modeling and other tasks.

[0055] Combine the following Figure 5 The above steps are schematically explained:

[0056] Figure 5 is a flow chart of an optional data association algorithm based on region division according to an embodiment of the present invention. Figure 5 As shown in step 1 in , when dividing the area, you can first obtain a list of all fusion targets maintained in the system in the previous association period (defined as fusion_tracks in this embodiment) and a list of target measurements received in the current period (defined as sensor_objects in this embodiment).

[0057] Step S402, executing a preset matrix block algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate a correlation cost matrix;

[0058] like Figure 5 In step 2, after obtaining the fusion list maintained in the previous cycle and the target measurement list received in the current cycle, the "cost matrix block calculation algorithm" can be executed to generate an association cost matrix (defined as association_mat in this embodiment).

[0059] The following is a detailed description of how to generate the association cost matrix.

[0060] As an optional implementation of this embodiment, before constructing the data fusion point set, it also includes: determining the index partition value of the fusion target and the target measurement based on the total number of fusion targets, wherein the index partition value is used to divide the fusion target and the target measurement in the data fusion point set.

[0061] Optionally, a preset matrix blocking algorithm is executed on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate the step of an associated cost matrix, including: constructing a data fusion point set; obtaining the first position coordinate point of each fusion target in the fusion target list and the second position coordinate point of each target measurement in the target measurement list; placing the first position coordinate point and the second position coordinate point into the data fusion point set in sequence; using a preset region partitioning algorithm to perform region partitioning on each position coordinate point in the data fusion point set to generate a region partitioning set; based on the region partitioning set, updating the corresponding element values ​​in the associated cost matrix to obtain the associated cost matrix.

[0062] Figure 6 is a flowchart of an optional cost matrix block calculation algorithm according to an embodiment of the present invention, combined with Figure 6 The steps of generating the association cost matrix are schematically illustrated, such as Figure 6 As shown, including:

[0063] Set the target measurement list sensor_objects and the fusion target list fusion_tracks;

[0064] Initialize the element value of the association cost matrix association_mat formed by the target measurement list and the fused target list to the maximum matching distance;

[0065] Define the index partition value of the fusion target and target measurement (indicate the index partition point, for example, the fusion target list has 10, the target measurement table has 10, and there are a total of 20 points 0-19, then the data 10 is the index partition value) track_meas_split is the number of fusion targets;

[0066] Put the position coordinate points of the fusion target into the fusion target and target measurement point set list points (corresponding to the above-mentioned data fusion point set) in the order of their positions in the fusion target list, and put the position coordinate points of the target measurement into the fusion target and target measurement point set list points in the order of their positions in the target measurement list, that is, put the above-mentioned first position coordinate point and the second position coordinate point into the data fusion point set in sequence;

[0067] Call the "region partitioning algorithm" for points to generate region partitioning results zones (corresponding to the above region partitioning set); according to the above region partitioning results zones, the corresponding element values ​​in the associated cost matrix can be updated to obtain the above associated cost matrix.

[0068] In this embodiment, when performing region division, a region division method based on overlapping areas is adopted, thereby avoiding missed associations that may be introduced by region division.

[0069] Optionally, a preset region division algorithm is used to divide regions for each position coordinate point in the data fusion point set, and the steps of generating a region division set include: respectively setting the length of the overlapping area in the first direction as one, the length of the overlapping area in the second direction as two, the number of regions to be divided in the first direction as one, and the number of regions to be divided in the second direction as two; calculating the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction in the data fusion point set; based on the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction, as well as the length of the overlapping area in the first direction as one, and the length of the overlapping area in the second direction as two, calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, and determining the region division set.

[0070] In the embodiment of the present invention, the first direction may be set as the x direction; and the second direction may be set as the y direction.

[0071] Figure 7 is a flowchart of an optional area division algorithm according to an embodiment of the present invention, combined with Figure 7 The preset region division algorithm is used to explain the steps of region division for each position coordinate point in the data fusion point set.

[0072] Specify the length of the overlapping area in the x (corresponding to the first direction) overlap_len_x (corresponding to the length of the upper overlapping area in the first direction, one), the length of the overlapping area in the y (corresponding to the second direction) overlap_len_y (corresponding to the length of the overlap in the second direction, two), the number of area divisions in the x direction partition_num_x (corresponding to the number of areas to be divided in the first direction, one), and the number of area divisions in the y direction partition_num_y (corresponding to the number of areas to be divided in the second direction, two).

[0073] Calculate the minimum and maximum values ​​x_min, x_max, y_min, y_max of all target points (corresponding to the above position coordinate points) in the point set (corresponding to the above data fusion point set) in the x and y directions.

[0074] Calculate the boundaries of the area after expanding the area determined by the target point in the point set by the width of the overlapping area around it: x_min = x_min-overlap_len_x, x_max = x_max+overlap_len_x, y_min = y_min-overlap_len_y, y_max = y_max+overlap_len_y.

[0075] As another optional aspect of this embodiment, after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, it also includes: calculating the length one of each sub-area in the first direction based on the number one of areas to be divided in the first direction, the minimum and maximum values ​​of the area position points after the overlapping area width is expanded in the first direction, and the length of the overlapping area in the first direction; and calculating the length two of each sub-area in the second direction based on the number two of areas to be divided in the second direction, the minimum and maximum values ​​of the area position points after the overlapping area width is expanded in the second direction, and the length of the overlapping area in the second direction.

[0076] Combination Figure 7 The above method is schematically illustrated to calculate the lengths of the sub-areas in the x and y directions, partition_len_x (corresponding to the above length one) = ((x_max-x_min)-(partition_num_x-1)*overlap_len_x) / partition_num_x, partition_len_y (corresponding to the above length two) = ((y_max-y_min)-(partition_num_y-1)*overlap_len_y) / partition_num_y.

[0077] An optional embodiment, after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width on all sides, also includes: defining index one of the forward partition in the first direction, index two of the reverse partition in the first direction, index three of the forward partition in the second direction, and index four of the reverse partition in the second direction; if there are unprocessed position points in the data fusion point set, determining the first direction coordinates and the second direction coordinates of the unprocessed position points; calculating the index one of the unprocessed position points based on the first direction coordinates of the unprocessed position points, the minimum value of all position points in the first direction, the length of the overlapping area in the first direction, and the length one of the sub-area in the first direction; based on the number of areas to be divided in the first direction 1. Calculate the index 2 of the unprocessed position point based on the maximum value of all position points in the first direction, the first direction coordinates of the unprocessed position point, the length of the overlapping area in the first direction, and the length 1 of the sub-area in the first direction; calculate the index 3 of the unprocessed position point based on the second direction coordinates of the unprocessed position point, the minimum value of all position points in the second direction, the length of the overlapping area in the second direction, and the length 2 of the sub-area in the second direction; calculate the index 4 of the unprocessed position point based on the number 2 of areas to be divided in the second direction, the maximum value of all position points in the second direction, the second direction coordinates of the unprocessed position point, the length of the overlapping area in the second direction, and the length 2 of the sub-area in the second direction; determine the position point index of the unprocessed position point.

[0078] Combine the following Figure 7 , the above implementation mode is schematically illustrated:

[0079] Define the index of the forward partition in the x direction (corresponding to the first direction) as x_forward_idx (corresponding to the index one), the index of the reverse partition in the x direction as x_reverse_idx (corresponding to the index two), the index of the forward partition in the y direction as y_forward_idx (corresponding to the index three), and the index of the reverse partition in the y direction as y_reverse_idx (corresponding to the index four);

[0080] In this embodiment, when generating the divided regions with overlapping areas, a region division algorithm combining forward division and reverse division is adopted, which can improve the efficiency of calculating the association cost matrix.

[0081] If there are unprocessed points in the point set points (corresponding to the above data fusion points), then take a point point (corresponding to the above unprocessed position point) from the point set points.

[0082] Note that the x-coordinate of point is point_x (corresponding to the first-direction coordinate), and the y-coordinate of point is point_y (corresponding to the second-direction coordinate), and calculate respectively:

[0083] x_forward_idx=point_x-x_min_) / (overlap_len_x+partition_len_x), x_reverse_idx=partition_num_x-1-(x_max_-point_x) / (overlap_len_x+partition_len_x), y_forward_idx=point_y-y_min_) / (overlap_len_y+partition_len_y), y_reverse_idx=partition_num_y-1-(y_max_-point_y) / (overlap_len_y+partition_len_y);

[0084] The current target point point belongs to the area whose indices (corresponding to the position point indices of the above-mentioned unprocessed position points) are (y_forward_idx, x_forward_idx), (y_forward_idx, x_reverse_idx), (y_reverse_idx, x_forward_idx), (y_reverse_idx, x_reverse_idx).

[0085] Optionally, the step of updating the corresponding element values in the association cost matrix based on the region division set to obtain the association cost matrix includes: determining whether there is an unprocessed region in the region division set; if there is an unprocessed region in the region division set, extracting a region to be processed from the region division set; obtaining an index of an unprocessed position point from the region to be processed; if the position point index is less than the index division value, storing the fusion target index indicated by the position point index into the fusion target index list corresponding to the region to be processed; if the position point index is greater than or equal to the index division value, storing the target measurement index indicated by the position point index into the target measurement index list corresponding to the region to be processed; calculating the distance between the fusion target and the target measurement within the region to be processed according to the fusion target index list and the target measurement index list, so as to update the corresponding elements in the association cost matrix to obtain the association cost matrix.

[0086] Combined with, such as Figure 6 , the above optional implementation manners are described.

[0087] If there is an unprocessed region in the division region set zones, take a current region zone from the division region set zones;

[0088] If there is an unprocessed target point index in the current region zone, take an index i of a target point from the current region zone (corresponding to the above position point index);

[0089] If i < track_meas_split, that is, the position point index is less than the index division value, store the fusion target index i into the fusion target index list track_l_idx corresponding to the current region; otherwise, store the target measurement index i - track_meas_split indicated by the position point index into the target measurement index list meas_l_idx corresponding to the current region;

[0090] Calculate the distance between the fusion target and the target measurement within the current region according to the index list track_l_idx and the target measurement index meas_l_idx, and update the corresponding element association_mat[track_l_idx][meas_l_idx] in the association cost matrix, and then continue to judge whether there is an unprocessed target index in the current region set zone until there is no unprocessed region in the region division set.

[0091] Return the association cost matrix association_mat.

[0092] Step S403, according to the association cost matrix, using the minimum cost matching algorithm to obtain the association matching relationship between the fusion target and the target measurement.

[0093] Step S404: based on the association matching relationship, the target measurement list received in the current period is used to update the fusion target list.

[0094] Combination Figure 5 The updating of the fusion target list is schematically explained. Step 3, according to the association cost matrix association_mat, the minimum cost matching algorithm is used to obtain the association matching relationship between the fusion target (generating the actual perceived target) and the target measurement (the only target object detected by multiple sensors).

[0095] Step 4: Update the fusion target of the previous cycle with the measurement of the current cycle, and then end.

[0096] Through the above embodiment, in order to solve the problem of high computational complexity of the association cost matrix in the above GNN, the space occupied by all target measurements and fused targets in a frame of data is divided into several sub-regions with overlapping areas, so that there is no association matching relationship between target measurements and fused targets belonging to different sub-regions, thereby converting the global cost matrix for calculating the entire spatial area into the local cost matrix for calculating each sub-region, thereby reducing the time complexity of calculating the association cost matrix, and further improving the efficiency of the data association algorithm.

[0097] The present invention is described below in conjunction with another optional embodiment.

[0098] Embodiment 2

[0099] This embodiment provides a multi-sensor data association device, and each implementation unit included in the data association device corresponds to each implementation step in the above-mentioned embodiment 1.

[0100] This embodiment provides a multi-sensor data association device. It should be noted that the multi-sensor data association device of the embodiment of the present invention can be used to execute the multi-sensor data association method provided by the embodiment of the present invention. The multi-sensor data association device provided by the embodiment of the present invention is introduced below.

[0101] Figure 8 is a schematic diagram of another optional multi-sensor data association device according to an embodiment of the present invention, such as Figure 8 As shown, it includes: an acquisition unit 81, a generation unit 82, a determination unit 83, and an update unit 84, wherein:

[0102] An acquisition unit 81 is used to acquire a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle;

[0103] A generating unit 82, configured to execute a preset matrix block algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate a correlation cost matrix;

[0104] A determination unit 83 is used to obtain the correlation matching relationship between the fusion target and the target measurement by using a minimum cost matching algorithm according to the correlation cost matrix;

[0105] The updating unit 84 is configured to update the fusion target list by using the target measurement list received in the current period based on the association matching relationship.

[0106] The above-mentioned multi-sensor data association device can be used to obtain the fusion target list maintained in the system in the previous cycle and the target measurement list received in the current cycle through the acquisition unit 81, wherein each fusion target in the fusion target list represents the actual perception object, and the target measurement list contains the measurement information of all actual objects collected by multiple sensors in the current cycle, and then the fusion target in the fusion target list and the target measurement in the target measurement list are executed by the generation unit 82 to generate the association cost matrix, and then the determination unit 83 uses the minimum cost matching algorithm to obtain the association matching relationship between the fusion target and the target measurement according to the association cost matrix, and finally the update unit 84 uses the target measurement list received in the current cycle to update the fusion target list based on the association matching relationship. In this embodiment, the region division method is introduced, and the association relationship between the fusion target and the target measurement is obtained by the association cost matrix and the minimum matching algorithm, so as to achieve the purpose of reducing the time complexity of calculating the association cost matrix, thereby achieving the technical effect of improving the efficiency of the data association algorithm, and then solving the technical problem of high calculation time complexity when fusing the perception results of each sensor in the related technology.

[0107] Optionally, the generation unit includes: a first construction subunit, used to construct a data fusion point set; a first acquisition subunit, used to acquire the first position coordinate point of each fusion target in the fusion target list and the second position coordinate point of each target measurement in the target measurement list; a storage subunit, used to put the first position coordinate point and the second position coordinate point into the data fusion point set in sequence; a first generation subunit, used to use a preset area partitioning algorithm to perform area partitioning on each position coordinate point in the data fusion point set to generate an area partitioning set; an update subunit, used to update the corresponding element values ​​in the association cost matrix based on the area partitioning set to obtain the association cost matrix.

[0108] Optionally, the generation unit also includes: a second construction subunit, used to construct an initial state association cost matrix based on the target measurement list and the fusion target list before constructing the data fusion point set; a first processing subunit, used to initialize the element value of the association cost matrix to the maximum matching distance.

[0109] Optionally, the first generating subunit includes: a setting module, used to respectively set the length 1 of the overlapping area in the first direction, the length 2 of the overlapping area in the second direction, the number 1 of the areas to be divided in the first direction, and the number 2 of the areas to be divided in the second direction; a first calculation module, used to calculate the minimum and maximum values ​​of all position coordinate points in all data fusion point sets in the first direction and the second direction; a second calculation module, used to calculate the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width based on the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction, as well as the length 1 of the overlapping area in the first direction and the length 2 of the overlapping area in the second direction, and determine the area division set.

[0110] Optionally, the first generating subunit also includes: a third calculation module, which is used to calculate the length one of each sub-region in the first direction based on the number one of regions to be divided in the first direction, the minimum and maximum values ​​of the region position points after the width of the overlapping area is expanded in the first direction, and the length of the overlapping area in the first direction, after calculating the boundaries of the region after the width of the overlapping area is expanded around the region determined by the position points in the data fusion point set; and a fourth calculation module, which calculates the length two of each sub-region in the second direction based on the number two of regions to be divided in the second direction, the minimum and maximum values ​​of the region position points after the width of the overlapping area is expanded in the second direction, and the length of the overlapping area in the second direction.

[0111] Optionally, the first generating subunit further includes: a defining module, for defining an index one of a forward partition in the first direction, an index two of a reverse partition in the first direction, an index three of a forward partition in the second direction, and an index four of a reverse partition in the second direction after calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width respectively around the area; a first determining module, for determining the first direction coordinates and the second direction coordinates of the unprocessed position points if there are unprocessed position points in the data fusion point set; a fifth calculating module, for calculating the index one of the unprocessed position points based on the first direction coordinates of the unprocessed position points, the minimum value of all position points in the first direction, the length of the overlapping area in the first direction, and the length one of the sub-area in the first direction; a sixth calculating module, for calculating the index one of the unprocessed position points based on the area to be divided in the first direction. The seventh calculation module is used to calculate the index of the unprocessed position point based on the first number of domains, the maximum value of all position points in the first direction, the first direction coordinates of the unprocessed position point, the length of the overlapping area in the first direction, and the first length of the sub-area in the first direction; the seventh calculation module is used to calculate the index of the unprocessed position point based on the second direction coordinates of the unprocessed position point, the minimum value of all position points in the second direction, the length of the overlapping area in the second direction, and the second length of the sub-area in the second direction; the eighth calculation module is used to calculate the index of the unprocessed position point based on the second number of areas to be divided in the second direction, the maximum value of all position points in the second direction, the second direction coordinates of the unprocessed position point, the length of the overlapping area in the second direction, and the second length of the sub-area in the second direction; the second determination module is used to determine the position point index of the unprocessed position point.

[0112] Optionally, the generating unit further includes: a determining subunit, used to determine the index partitioning value of the fusion target and the target measurement based on the total number of fusion targets before constructing the data fusion point set, wherein the index partitioning value is used to divide the fusion target and the target measurement in the data fusion point set.

[0113] Optionally, the updating subunit includes: a judgment module, which judges whether there is an unprocessed area in the area division set; a first processing module, which extracts a to-be-processed area from the area division set if there is an unprocessed area in the area division set; and obtains the unprocessed position point index from the to-be-processed area; a second processing module, which stores the fusion target index indicated by the position point index into a fusion target index list corresponding to the to-be-processed area if the position point index is less than the index division value; a third processing module, which stores the target measurement index indicated by the position point index into a target measurement index list corresponding to the to-be-processed area if the position point index is greater than or equal to the index division value; and a ninth calculation module, which calculates the distance between the fusion target and the target measurement in the to-be-processed area according to the fusion target index list and the target measurement index list, so as to update the corresponding elements in the association cost matrix and obtain the association cost matrix.

[0114] The above-mentioned multi-sensor data association may also include a processor and a memory. The above-mentioned acquisition unit 81, generation unit 82, determination unit 83, update unit 84, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0115] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the problem of high computational time complexity in the method of fusing the perception results of each sensor in the related art can be solved by adjusting the kernel parameters.

[0116] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0117] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-mentioned credit system operation monitoring methods by executing the executable instructions.

[0118] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium comprising a stored computer program, wherein when the computer program is executed

[0119] Fig. 9 FIG. 1 is a hardware structure block diagram of an electronic device (or mobile device) for a method for monitoring operation of a credit system according to an embodiment of the present invention. Fig. 9 As shown, the electronic device may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), and a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a keyboard, a power supply and / or a camera. It can be understood by those skilled in the art that Fig. 9 The structure shown is only for illustration and does not limit the structure of the above electronic device. Fig. 9 More or fewer components as shown, or with Fig. 9 Different configurations are shown.

[0120] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0121] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0122] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0123] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0124] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0125] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0126] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-sensor data association method, characterized in that: include: Obtaining a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle; The preset matrix block algorithm is executed on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an associated cost matrix, including: constructing a data fusion point set; obtaining a first position coordinate point of each fusion target in the fusion target list and a second position coordinate point of each target measurement in the target measurement list; placing the first position coordinate point and the second position coordinate point into the data fusion point set in sequence; using a preset region partitioning algorithm to perform region partitioning on each position coordinate point in the data fusion point set to generate a region partitioning set; based on the region partitioning set, updating corresponding element values ​​in the associated cost matrix to obtain an associated cost matrix; The steps of using a preset area division algorithm to divide each position coordinate point in the data fusion point set into regions, and generating a region division set include: respectively setting a length of one of the overlapping area in the first direction, a length of two of the overlapping area in the second direction, a number of one of the regions to be divided in the first direction, and a number of two of the regions to be divided in the second direction; calculating the minimum and maximum values ​​of all the position coordinate points in the data fusion point set in the first direction and the second direction; based on the minimum and maximum values ​​of all the position coordinate points in the first direction and the second direction, as well as the length of one of the overlapping area in the first direction and the length of two of the overlapping area in the second direction, calculating the boundaries of the region after the region determined by the position points in the data fusion point set is expanded by an overlapping area width around the region, and determining the region division set; According to the association cost matrix, a minimum cost matching algorithm is used to obtain an association matching relationship between the fusion target and the target measurement; Based on the association matching relationship, the fusion target list is updated by using the target measurement list received in the current period.

2. The data association method according to claim 1, characterized in that: Before building the data fusion point set, it also includes: Based on the target measurement list and the fusion target list, construct an association cost matrix of an initial state; Initialize the element value of the association cost matrix to the maximum matching distance.

3. The data association method according to claim 1, characterized in that: After calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlap area width around each other, the method further includes: Calculate the length of each sub-region in the first direction based on the number of regions to be divided in the first direction, the minimum and maximum values ​​of the region position points after the width of the overlapping region is expanded in the first direction, and the length of the overlapping region in the first direction. Based on the number of regions to be divided in the second direction, the minimum and maximum values ​​of the region position points after the width of the overlapping area is expanded in the second direction, and the length of the overlapping area in the second direction, the length of each sub-area in the second direction is calculated.

4. The data association method according to claim 3, characterized in that: After calculating the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlap area width around each other, the method further includes: Define index one of the forward partition in the first direction, index two of the reverse partition in the first direction, index three of the forward partition in the second direction, and index four of the reverse partition in the second direction; If there are unprocessed position points in the data fusion point set, determining the first direction coordinates and the second direction coordinates of the unprocessed position points; Calculate an index one of the unprocessed position point based on the first direction coordinate of the unprocessed position point, the minimum value of all position points in the first direction, the length of the overlapping area in the first direction, and the length one of the sub-area in the first direction; Calculate the index 2 of the unprocessed position point based on the number 1 of regions to be divided in the first direction, the maximum value of all position points in the first direction, the first direction coordinates of the unprocessed position point, the length of the overlapping area in the first direction, and the length 1 of the sub-region in the first direction; Calculate the third index of the unprocessed position point based on the second direction coordinate of the unprocessed position point, the minimum value of all position points in the second direction, the length of the overlapping area in the second direction, and the second length of the sub-area in the second direction; Calculate the index of the unprocessed position point based on the second number of regions to be divided in the second direction, the maximum value of all position points in the second direction, the second direction coordinate of the unprocessed position point, the length of the overlapping area in the second direction, and the second length of the sub-region in the second direction; Determine the position point index of the unprocessed position point.

5. The data association method according to claim 1, characterized in that: Before building the data fusion point set, it also includes: Based on the total number of the fusion targets, an index partition value of the fusion target and the target measurement is determined, wherein the index partition value is used to partition the fusion target and the target measurement in the data fusion point set.

6. The data association method according to claim 5, characterized in that: Based on the region partition set, the step of updating the corresponding element values ​​in the association cost matrix to obtain the association cost matrix includes: Determining whether there is an unprocessed area in the area division set; If there is an unprocessed area in the area division set, extracting a to-be-processed area from the area division set; obtaining an unprocessed position point index from the to-be-processed area; If the position point index is less than the index division value, storing the fusion target index indicated by the position point index into the fusion target index list corresponding to the area to be processed; If the location point index is greater than or equal to the index division value, storing the target measurement index indicated by the location point index into the target measurement index list corresponding to the area to be processed; According to the fusion target index list and the target measurement index list, the distance between the fusion target and the target measurement in the area to be processed is calculated to update the corresponding elements in the association cost matrix to obtain the association cost matrix.

7. A multi-sensor data association device, characterized in that: include: An acquisition unit is used to acquire a fusion target list maintained in the system in the previous cycle and a target measurement list received in the current cycle, wherein each fusion target in the fusion target list represents an actual perceived object, and the target measurement list contains measurement information of all actual objects collected by multiple sensors in the current cycle; A generating unit is used to execute a preset matrix block algorithm on the fusion targets in the fusion target list and the target measurements in the target measurement list to generate an associated cost matrix; the generating unit includes: a first constructing subunit, used to construct a data fusion point set; a first acquiring subunit, used to acquire a first position coordinate point of each fusion target in the fusion target list and a second position coordinate point of each target measurement in the target measurement list; a storing subunit, used to sequentially place the first position coordinate point and the second position coordinate point into the data fusion point set; a first generating subunit, used to adopt a preset region partitioning algorithm to perform region partitioning on each position coordinate point in the data fusion point set to generate a region partitioning set; an updating subunit, used to update the associated cost matrix based on the region partitioning set. The first generating subunit comprises: a setting module, used to respectively set the length of the overlapping area in the first direction, the length of the overlapping area in the second direction, the number of areas to be divided in the first direction, and the number of areas to be divided in the second direction; a first calculating module, used to calculate the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction in all data fusion point sets; a second calculating module, used to calculate the boundaries of the area after the area determined by the position points in the data fusion point set is expanded by an overlapping area width based on the minimum and maximum values ​​of all position coordinate points in the first direction and the second direction, as well as the length of the overlapping area in the first direction and the length of the overlapping area in the second direction, so as to determine the area division set; A determination unit, configured to obtain an association matching relationship between the fusion target and the target measurement by using a minimum cost matching algorithm according to an association cost matrix; An updating unit is configured to update the fusion target list by using the target measurement list received in a current period based on the association matching relationship.

8. An electronic device, characterized in that: The invention comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the multi-sensor data association method according to any one of claims 1 to 6.

Images