A method and corresponding apparatus for constraint search based on data association

Abstract

The application discloses a method and corresponding device for constraint search based on data association. The method comprises the following steps: after determining a to-be-searched trajectory and associated data, initializing search parameters required for this search based on the length of the to-be-searched trajectory; searching forward search trajectory points and backward search trajectory points that are associated with the associated data from the to-be-searched trajectory by adopting a mode of dynamic adjustment of search step length and forward-backward bidirectional collaborative search according to the initialized dense search step length and sparse search step length; and merging the forward search trajectory points and the backward search trajectory points to obtain a constraint search result of this time. By the mode of forward-backward bidirectional collaborative search combined with dynamic adjustment of step length, it can be ensured that there is sufficient constraint in sensitive areas such as overlapping area edges, while the number of constraints is prevented from doubling due to too small step length, so that the calculation amount is effectively controlled while reasonable constraints are taken into account, and the calculation efficiency is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision and three-dimensional imaging, in particular to the technical field of point cloud reconstruction and image processing, and particularly relates to a method for constraint search based on data association and a corresponding device. BACKGROUND

[0002] In the field of autonomous driving, mobile robots, etc., comprehensive and high-precision three-dimensional point cloud reconstruction is a prerequisite for three-dimensional mapping, scene reconstruction, and surveying and mapping technical service projects. Taking large-scale scene reconstruction as an example, during project implementation, data collection of hundreds or even thousands of kilometers is often involved; the entire journey scene contains complex environments such as openness and occlusion, and needs to be collected by multiple vehicles or even vehicle teams or collected multiple times. Moreover, the same area is often collected multiple times, or the collection area is expanded to supplement the collection. As can be seen, three-dimensional point cloud reconstruction needs multiple and large amounts of data collection to comprehensively obtain important data in the scene.

[0003] However, multiple and large amounts of data collection inevitably lead to repeated data in overlapping areas. In order to ensure that the data in the overlapping area obtained by multiple collections is consistent in accuracy, without quality problems such as stratification and ambiguity, and to ensure high consistency between data, joint optimization is needed between multiple data, incremental data, and historical data. The key to joint optimization is to fully and efficiently find the constraints between data.

[0004] At present, data internal loop search and constraint search between data (such as between incremental data and historical data) mostly use equal interval search. However, equal interval search has the following problems: a larger search step will lead to a lack of constraints in sensitive areas such as the edge of the overlapping area, missing important constraints, and being unable to accurately search and locate the constraints; a smaller search step will lead to a doubling of the number of constraints, increasing the downstream calculation amount. SUMMARY

[0005] The present application provides a method for constraint search based on data association and a corresponding device to solve the problem that the constraint search scheme in the prior art cannot balance search positioning accuracy and an appropriate amount of search constraints without affecting the downstream calculation amount due to fixed search step.

[0006] The technical solution is as follows:

[0007] In a first aspect, a method for constraint search based on data association is provided, comprising:

[0008] determining a to-be-searched trajectory and associated data having a constraint relationship with the to-be-searched trajectory; wherein the constraint relationship includes one or more of position constraint, distance constraint, and scene constraint;

[0009] initialize search parameters required for the current search based on a length of the trajectory to be searched, wherein the search parameters include a dense search step and a sparse search step;

[0010] search for a forward search trajectory point and a backward search trajectory point associated with the constraint of the associated data from the trajectory to be searched according to the initialized dense search step and sparse search step, in a manner of dynamically adjusting search steps and bidirectional collaborative searching;

[0011] merge the forward search trajectory point and the backward search trajectory point to obtain a constraint search result of the current search.

[0012] In a possible implementation, the search parameters required for the current search are initialized based on the length of the trajectory to be searched, and specifically include:

[0013] the dense search step is initialized as a first proportion range of the length of the trajectory to be searched, and the first proportion range is [1 / 100, 1 / 200];

[0014] the sparse search step is initialized as a second proportion range of the length of the trajectory to be searched, and the second proportion range is [1 / 20, 1 / 50].

[0015] In a possible implementation, the forward search trajectory point and the backward search trajectory point associated with the constraint of the associated data are searched from the trajectory to be searched according to the initialized dense search step and sparse search step, in a manner of dynamically adjusting search steps and bidirectional collaborative searching, and specifically include:

[0016] perform down-sampling processing on the trajectory to be searched, and store the associated data as a k-d tree structure;

[0017] from a head end to a tail end of the down-sampled trajectory to be searched, dynamically adjust a step in a search process based on the initialized dense search step and sparse search step, to search for the forward search trajectory point associated with the k-d tree structure;

[0018] from the tail end to the head end of the down-sampled trajectory to be searched, dynamically adjust the step in the search process based on the initialized dense search step and sparse search step, to search for the backward search trajectory point associated with the k-d tree structure.

[0019] In a possible implementation, the search parameters further include a dense segment continuity threshold, and the dense segment continuity threshold is positively correlated with a constraint requirement degree of the current search.

[0020] Based on the initialized dense search step and sparse search step, the step in the search process is dynamically adjusted to search for a forward search trajectory point constrained by the k-d tree structure, and specifically comprises:

[0021] Each trajectory point in the down-sampled trajectory to be searched is traversed with the sparse search step;

[0022] It is judged whether a trajectory point constrained by the k-d tree structure is searched in the sparse traversal process;

[0023] If yes, the trajectory point searched this time is recorded as the start point of the dense segment, the search step is adjusted from the sparse search step to the dense search step, and the subsequent unsearched trajectory points are traversed. Otherwise, the operation of judging whether a trajectory point constrained by the k-d tree structure is searched in the sparse traversal process is returned;

[0024] It is judged whether a trajectory point constrained by the k-d tree structure is searched in the dense traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold;

[0025] If yes, the trajectory point searched this time is recorded to the dense segment. Otherwise, the search step is adjusted from the dense search step to the sparse search step, and the subsequent unsearched trajectory points are traversed;

[0026] The above operation is repeated until each trajectory point in the trajectory to be searched is traversed, and the trajectory points recorded by all dense segments are counted as the forward search trajectory point.

[0027] In a possible implementation, when a trajectory point constrained by the k-d tree structure is not searched in the dense traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold, the method further comprises:

[0028] The search step is adjusted from the dense search step to a temporary search step starting from the trajectory point corresponding to the start point of the current dense segment, and the subsequent unsearched trajectory points are traversed. The temporary search step is smaller than the dense search step;

[0029] It is judged whether a trajectory point constrained by the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold;

[0030] If yes, the trajectory point searched this time is recorded to the dense segment. Otherwise, the search step is adjusted from the temporary search step to the sparse search step, and the subsequent unsearched trajectory points are traversed.

[0031] In a possible implementation, the association constraint relationship is a position constraint, and searching for the trajectory point associated with the k-d tree structure means that the position coordinates of the trajectory point in the trajectory to be searched and the trajectory point in the k-d tree structure are the same or satisfy a set proximity;

[0032] The association constraint relationship is a distance constraint, and searching for the trajectory point associated with the k-d tree structure means that the distance between the trajectory point in the trajectory to be searched and the trajectory point in the k-d tree structure is less than a set threshold value.

[0033] The association constraint relationship is a scene constraint, and searching for the trajectory point associated with the k-d tree structure means that the scene of the trajectory point in the trajectory to be searched and the trajectory point in the k-d tree structure is the same or satisfies a set similarity, or the scene features of the trajectory point in the trajectory to be searched and the trajectory point in the k-d tree structure are the same or satisfy a set proximity.

[0034] In a possible implementation, the forward search trajectory point and the backward search trajectory point are combined to obtain the constraint search result, and the combination specifically includes:

[0035] Any trajectory point in the forward search trajectory point and the backward search trajectory point that is separated by a distance less than the dense search step is removed, and the remaining trajectory points are combined to obtain the constraint search result; or

[0036] For any trajectory point in the forward search trajectory point and the backward search trajectory point that is in the dense segment, it is determined whether the distance between the current trajectory point and the nearest trajectory point is greater than the dense search step, and if yes, the trajectory point is not processed, otherwise, the trajectory point is removed. For any trajectory point in the forward search trajectory point and the backward search trajectory point that is not in the dense segment, it is determined whether the distance between the current trajectory point and the nearest trajectory point is greater than the sparse search step, and if yes, the trajectory point is not processed, otherwise, the trajectory point is removed. The remaining trajectory points are combined to obtain the constraint search result.

[0037] In a second aspect, a device for constraint search based on data association is provided, and the device includes:

[0038] A determination module is configured to determine a trajectory to be searched and associated data that has an association constraint relationship with the trajectory to be searched, wherein the association constraint relationship includes one or more of a position constraint, a distance constraint, and a scene constraint.

[0039] An initialization module is configured to initialize search parameters required for the current search based on the length of the trajectory to be searched, wherein the search parameters include a dense search step and a sparse search step.

[0040] The search module is configured to search, according to the initialized dense search step and sparse search step, the forward search track point and the backward search track point associated with the associated data constraint from the to-be-searched track by means of dynamic adjustment of search step and forward and backward bidirectional collaborative search.

[0041] The merging module is configured to merge the forward search track point and the backward search track point to obtain a constraint search result of this time.

[0042] In a third aspect, an electronic device is provided, comprising:

[0043] at least one processor; and

[0044] a memory connected with the at least one processor; wherein

[0045] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of the aspects and any possible implementation manner described above.

[0046] In a fourth aspect, a computer-readable storage medium is provided, and the storage medium stores at least one instruction, which is loaded and executed by a processor to implement the method of the aspects and any possible implementation manner described above.

[0047] In a fifth aspect, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the method of the aspects and any possible implementation manner described above.

[0048] The beneficial effects of the technical solutions provided in the present application at least include:

[0049] From the above technical solutions, it can be seen that after determining the to-be-searched track and the associated data, the present application can search in the sensitive areas such as the edge of the overlapping area with a dense search step (i.e. a short step) and search in the non-sensitive areas outside the overlapping area with a sparse search step (i.e. a long step) based on forward and backward bidirectional collaborative search. Through the combination of forward and backward bidirectional collaborative search and dynamic adjustment of the step, sufficient constraints can be ensured in the sensitive areas such as the edge of the overlapping area, while the number of constraints is prevented from being doubled due to an excessively small step, thereby effectively controlling the calculation amount while taking into account reasonable constraints, ensuring the calculation efficiency, ensuring the high consistency between data, and avoiding quality problems such as layering and blurring.

[0050] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative effort based on the embodiments in the present application shall fall within the protection scope of the present application.

[0052] Figure 1 is a step schematic diagram of a method for constraint search based on data association provided by an embodiment of the present application.

[0053] Figure 2 is a flow schematic diagram of constraint search of a to-be-searched trajectory C based on associated data ref provided by an embodiment of the present application.

[0054] Figure 3a is a result schematic diagram of equal-step search in the prior art.

[0055] Figure 3b is a result schematic diagram of forward search with dynamic step length adopted by the present application.

[0056] Figure 3c is a result schematic diagram of backward search with dynamic step length adopted by the present application.

[0057] Figure 3d is a result schematic diagram of combined processing of forward search trajectory points and backward search trajectory points provided by the present application.

[0058] Figure 4 is a structure block diagram of an apparatus for constraint search based on data association provided by still another embodiment of the present application.

[0059] Figure 5 is a block diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0060] The exemplary embodiments of the present application will be described below with reference to the accompanying drawings, including various details in order to understand the present application. These should be considered as merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Also, in order to be clear and concise, the description below omits the description of well-known functions and structures.

[0061] Obviously, the described embodiments are only some embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort shall fall within the protection scope of the present application.

[0062] It should be noted that the terminal device involved in the embodiments of the present application can include, but is not limited to, a mobile phone, a personal digital assistant (Personal Digital Assistant, PDA), a wireless handheld device, a tablet computer (Tablet Computer), and the like. The display device can include, but is not limited to, a personal computer, a television, and the like.

[0063] In addition, the term "and / or" in this paper is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the existence of A alone, the existence of A and B together, and the existence of B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are in an "or" relationship.

[0064] In view of the problem that the search constraint scheme in the prior art cannot balance the search positioning accuracy and the appropriate search constraint without affecting the downstream calculation amount due to the fixed search step, the present application proposes a constraint search scheme with dynamically adjusted search step, thereby improving the calculation speed and ensuring the necessary condition of sufficient constraint in sensitive areas such as the edge of the overlap area. The main inventive concept of the present application is that, after determining the to-be-searched trajectory and the associated data, based on the forward and backward collaborative search, the search can be performed with a dense search step (i.e., a short step) in sensitive areas such as the edge of the overlap area, and the search can be performed with a sparse search step (i.e., a long step) in non-sensitive areas outside the overlap area. Through the forward and backward collaborative search combined with the dynamically adjusted step, sufficient constraint in sensitive areas such as the edge of the overlap area can be ensured, while avoiding the doubling of the number of constraints due to too small a step, thereby effectively controlling the calculation amount while taking into account reasonable constraints, ensuring the calculation efficiency, ensuring the high consistency between data, and avoiding quality problems such as layering and blurring.

[0065] The application scenario of the present application scheme can include:

[0066] 1. Autonomous driving: through high-precision three-dimensional point cloud reconstruction, more accurate environmental information is provided for the autonomous driving system, which helps the vehicle to better plan the driving path, avoid collision, and improve the driving safety.

[0067] 2. Three-dimensional mapping of mobile robots: efficient three-dimensional mapping in complex environments ensures that the robot can accurately understand the surrounding environment, thereby making safer and more effective navigation decisions.

[0068] 3. Surveying and mapping field: in large-scale scene reconstruction, through joint optimization of multiple data, incremental data, and historical data, the high consistency between data is ensured, and quality problems such as layering and blurring are avoided.

[0069] Reference Figure 1As shown in the figure, it is a schematic diagram of steps of a method for constraint search based on data association provided by an embodiment of the present application. The execution subject of the method can be a device for constraint search based on data association, which can be an electronic device such as a computer, a tablet computer, a smart terminal, a smart wearable device, etc. with functions of data calculation, processing, storage, etc., or a software module or component integrated in these electronic devices, etc.

[0070] As shown in the figure, the method for constraint search based on data association can include the following steps: Figure 1

[0071] Step 102: determining a to-be-searched trajectory and associated data having a constraint relationship with the to-be-searched trajectory; wherein the constraint relationship includes one or more of a position constraint, a distance constraint, and a scene constraint.

[0072] In the present application, the to-be-searched trajectory can be relatively complete scene data collected in three-dimensional point cloud reconstruction, for example, multi-circle scene data collected at one time for a park. The associated data can be the to-be-searched trajectory itself, or historical scene data or incremental scene data collected at other times, etc. For example, historical scene data collected at a historical time for the above-mentioned park.

[0073] The constraint relationship in the present application is the key to search. Only the data having a constraint relationship with the to-be-searched trajectory can be identified as associated data. The constraint relationship can be a position association, for example, having the same or similar position, or a distance association, for example, being very close, or a scene association, for example, being located in the same or similar scene. In summary, any data having a similar constraint relationship can be associated data of the to-be-searched trajectory.

[0074] Step 104: initializing search parameters required for this search based on the length of the to-be-searched trajectory, wherein the search parameters include a dense search step and a sparse search step.

[0075] In the present application, the dense search step is initialized as a first proportional range of the length of the to-be-searched trajectory, and the first proportional range is [1 / 100, 1 / 200]. The dense search step is used to control the fine granularity in the trajectory search process. A shorter step can improve the search accuracy, but will increase the calculation amount. Therefore, the dense search step can be set to 1 / 100 to 1 / 200 of the total length of the trajectory, and the minimum step is ensured to be not less than 1m, so as to ensure the effectiveness and calculation feasibility of the search.

[0076] ​The sparse search step is initialized as a second proportion range of the length of the trajectory to be searched, and the second proportion range is [1 / 20, 1 / 50]. The sparse search step is used to accelerate the search process of the long-distance trajectory segment in the non-sensitive area, and avoid wasting time and effort (computing power) by performing short-step search in the non-sensitive area. Therefore, the sparse search step is generally set to 1 / 20 to 1 / 50 of the total length of the trajectory to balance the search efficiency and coverage.

[0077] Step 106: According to the initialized dense search step and sparse search step, the forward search trajectory point and the backward search trajectory point associated with the associated data association constraint are searched from the trajectory to be searched by using the sparse and dense dynamic adjustment search step and the front and back bidirectional collaborative search mode.

[0078] In the scheme of the present application, the sparse and dense dynamic adjustment search step and the front and back bidirectional collaborative search mode can be used to search the positioning trajectory point. For example, the sparse and dense dynamic adjustment search step is used for forward search to obtain the forward search trajectory point, and the sparse and dense dynamic adjustment search step is used for backward search to obtain the backward search trajectory point. For another example, in one search process, the sparse and dense dynamic adjustment search step is used for search, and each time the forward search trajectory point is searched, the forward and backward search is performed based on the trajectory point, and the trajectory point is recorded. Generally, the search is ended after a certain number of searches, and the forward search is continued until the search is completed, and the forward search trajectory point and the backward search trajectory point in the search process are counted.

[0079] Optionally, when the forward search trajectory point and the backward search trajectory point associated with the associated data association constraint are searched from the trajectory to be searched by using the sparse and dense dynamic adjustment search step and the front and back bidirectional collaborative search mode according to the initialized dense search step and sparse search step, the trajectory to be searched can be down-sampled, and the associated data can be stored as a k-d tree structure; from the head end to the tail end of the down-sampled trajectory to be searched: based on the initialized dense search step and sparse search step, the step in the search process is dynamically adjusted to search the forward search trajectory point associated with the k-d tree structure; from the tail end to the head end of the down-sampled trajectory to be searched: based on the initialized dense search step and sparse search step, the step in the search process is dynamically adjusted to search the backward search trajectory point associated with the k-d tree structure.

[0080] The downsampling processing on the to-be-searched trajectory can reduce the number of trajectory points, reduce the search times, and improve the search efficiency. The association data is stored as a k-d tree structure, and specifically, the association data can be stored in the k-d tree structure for facilitating search and traversal. By performing forward search and backward search on the to-be-searched trajectory respectively, more trajectory points can be covered to the greatest extent, and error accumulation caused by single-direction search can be avoided, that is, missing search of trajectory points having an association constraint relationship can be avoided, and the search hit rate and efficiency are improved.

[0081] Further, the search parameter further includes a dense segment continuous threshold value, which is positively correlated with the constraint requirement degree this time. For example, the constraint requirement for a sensitive area is higher, and the dense segment continuous threshold value can be adjusted to be larger, and vice versa, which can be adjusted to be smaller. The dense segment continuous threshold value is used to control the switching condition of the search mode. When the number of steps of the continuous dense search reaches the threshold value, the sparse search mode can be switched to improve the efficiency. The dense segment continuous threshold value can be set as a fixed value (such as 4 or 8), or adjusted according to actual application requirements.

[0082] Optionally, when searching for the forward search trajectory point or the backward search trajectory point associated with the k-d tree structure in the search process based on the initialized dense search step length and the sparse search step length, the following operations can be referred to: traversing each trajectory point in the to-be-searched trajectory after downsampling at the sparse search step length; judging whether a trajectory point associated with the k-d tree structure is searched in the sparse traversal process; if yes, recording the searched trajectory point this time as the start point of a dense segment, adjusting the search step length from the sparse search step length to the dense search step length, and traversing the subsequent unsearched trajectory points; otherwise, returning to the operation of judging whether a trajectory point associated with the k-d tree structure is searched in the sparse traversal process; judging whether a trajectory point associated with the k-d tree structure is searched in the dense traversal process, and whether the number of trajectory points recorded in the dense segment reaches the dense segment continuous threshold value; if yes, recording the searched trajectory point this time to the dense segment; otherwise, adjusting the search step length from the dense search step length to the sparse search step length, and continuing to traverse the subsequent unsearched trajectory points; repeating the above operations until each trajectory point in the to-be-searched trajectory is traversed, and recording the trajectory points recorded in all dense segments as the forward search trajectory point. In this way, by dynamically adjusting the search step length, the long and short step lengths are adaptively switched according to the distribution characteristics of the trajectory overlapping area, the high precision is maintained in the sensitive area (such as the overlapping area and the boundary), and the search efficiency is improved in the non-sensitive area (outside the overlapping area).

[0083] When no trajectory point associated with the constraint of the k-d tree structure is searched in the dense traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold, the search step can be adjusted from the dense search step to a temporary search step, and the subsequent unsearched trajectory points can be continuously traversed from the trajectory point corresponding to the start point of the current dense segment. The temporary search step is smaller than the dense search step. It is determined whether the trajectory point associated with the constraint of the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold. If yes, the searched trajectory point is recorded in the dense segment. Otherwise, the search step is adjusted from the temporary search step to a sparse search step, and the subsequent unsearched trajectory points are continuously traversed. In the dense segment, the search precision in the dense segment can be improved by further reducing the search step to supplement the trajectory points with the associated constraint in the dense segment.

[0084] Optionally, the associated constraint relationship is a position constraint, and the searched trajectory point associated with the constraint of the k-d tree structure means that the position coordinates of the trajectory point in the to-be-searched trajectory are the same as or satisfy a set proximity degree with the position coordinates of the trajectory point in the k-d tree structure. For example, the position coordinates of the trajectory point 1 in the to-be-searched trajectory are the same as the position coordinates of the trajectory point I in the k-d tree structure, or the horizontal coordinates are the same.

[0085] The associated constraint relationship is a distance constraint, and the searched trajectory point associated with the constraint of the k-d tree structure means that the distance between the trajectory point in the to-be-searched trajectory and the trajectory point in the k-d tree structure is less than a set threshold. For example, the distance between the trajectory point 1 in the to-be-searched trajectory and the trajectory point I in the k-d tree structure is 0 or close to 0, or is less than 10 cm. The set threshold is only an example and is not limited.

[0086] The associated constraint relationship is a scene constraint, and the searched trajectory point associated with the constraint of the k-d tree structure means that the scene of the trajectory point in the to-be-searched trajectory is the same as or satisfies a set similarity degree with the scene of the trajectory point in the k-d tree structure, or the scene features of the trajectory point in the to-be-searched trajectory are the same as or satisfy a set proximity degree with the scene features of the trajectory point in the k-d tree structure. For example, the scene image of the trajectory point 1 in the to-be-searched trajectory is the same as or similar to the scene image of the trajectory point I in the k-d tree structure. Or, the scene features extracted from the scene image of the trajectory point 1 in the to-be-searched trajectory are the same as or similar to the scene features extracted from the scene image of the trajectory point I in the k-d tree structure. The set threshold is only an example and is not limited.

[0087] Step 108: merging the forward search trajectory point and the backward search trajectory point to obtain a current constraint search result.

[0088] In the scheme of the present application, the forward search trajectory points and the backward search trajectory points can be combined to obtain the current constraint search result in the following manner.

[0089] Manner one: any trajectory point in the forward search trajectory points and the backward search trajectory points with a distance less than the dense search step is removed, and the remaining trajectory points are combined to obtain the current constraint search result.

[0090] Manner two: for any trajectory point in the forward search trajectory points and the backward search trajectory points in the dense segment, it is determined whether the distance between the current trajectory point and the nearest trajectory point is greater than the dense search step, if yes, no processing is performed, otherwise, the trajectory point is removed; for any trajectory point in the forward search trajectory points and the backward search trajectory points outside the dense segment, it is determined whether the distance between the current trajectory point and the nearest trajectory point is greater than the sparse search step, if yes, no processing is performed, otherwise, the trajectory point is removed; the remaining trajectory points are combined to obtain the current constraint search result.

[0091] Both the above-mentioned manner one and manner two can avoid the trajectory points after combination being too dense, wherein, the manner one is superior to the manner two, in the manner one, whether the trajectory points searched in the dense segment or the trajectory points searched outside the dense segment, the problem of the trajectory points being too dense or even increasing the calculation amount of a certain trajectory segment due to the small interval between the trajectory points can be avoided, thereby effectively ensuring the reasonable distribution between the searched trajectory points. In the manner two, the interval between the trajectory points searched in the dense segment is not less than the dense search step, and the interval between the trajectory points searched outside the dense segment is not less than the sparse search step, thereby avoiding the dense trajectory points in the dense segment being too dense and the sparse trajectory points outside the dense segment being too dense.

[0092] Referring to FIG. 1, Figure 2 FIG. 1 is a flow diagram of a constraint search process of a to-be-searched trajectory C based on associated data ref provided by an embodiment of the present application.

[0093] Step 202: input the to-be-searched trajectory C and the associated data ref.

[0094] Step 204: down-sample the to-be-searched trajectory C, and store the associated data ref in a k-d tree.

[0095] Step 206: determine the associated function of the down-sampled to-be-searched trajectory C and the associated factor in the k-d tree.

[0096] The associated factor herein can be understood as an associated constraint relationship, and the associated function can be a specific relationship type. For example, the associated factor can be a distance relationship, and the associated function can be that the distance is less than a set threshold.

[0097] Step 208: Based on the forward search, each trajectory point in the down-sampled trajectory C to be searched is traversed with a large step size, and the correlation factor search of the k-d tree is performed.

[0098] Step 210: Determine whether the correlation factor is searched.

[0099] If yes, step 212 is performed, otherwise, jump back to step 208. Meanwhile, step 210 has a branch, that is, when the trajectory point searched this time is the last trajectory point, whether the correlation factor is searched or not, the current search needs to be ended.

[0100] Step 212: The current trajectory point searched is taken as the starting point of the dense segment, and the search step size is switched to a small step size.

[0101] Step 214: Continue the subsequent traversal with a small step size to perform the correlation factor search of the k-d tree.

[0102] Step 216: Determine whether the correlation factor is searched and the number of trajectories in the dense segment does not reach the dense segment continuity threshold.

[0103] If yes, record the trajectory point searched to the dense segment, and jump to step 214, otherwise, jump to step 208. Step 216 has a branch, that is, when the trajectory point searched this time is the last trajectory point, whether the correlation factor is searched or not, and whether the number of trajectories in the dense segment reaches the dense segment continuity threshold, the current forward search needs to be ended.

[0104] Step 218: End the current forward search to obtain the forward search trajectory point.

[0105] Step 220: Based on the backward search, each trajectory point in the down-sampled trajectory C to be searched is traversed with a large step size, and the correlation factor search of the k-d tree is performed.

[0106] Step 222: Determine whether the correlation factor is searched.

[0107] If yes, step 224 is performed, otherwise, jump back to step 220. Meanwhile, step 222 has a branch, that is, when the trajectory point searched this time is the last trajectory point, whether the correlation factor is searched or not, the current backward search needs to be ended.

[0108] Step 224: The current trajectory point searched is taken as the starting point of the dense segment, and the search step size is switched to a small step size.

[0109] Step 226: Continue the subsequent traversal with a small step size to perform the correlation factor search of the k-d tree.

[0110] Step 228: Determine whether the correlation factor is searched and the number of trajectories in the dense segment does not reach the dense segment continuity threshold.

[0111] If yes, record the searched trajectory points into the dense segment, and jump to step 226, otherwise jump to step 220. Step 228 has a branch, that is, when the searched trajectory point is the last trajectory point, whether the correlation factor is searched or not and whether the number of trajectory points in the dense segment reaches the dense segment continuity threshold value, the current backward search needs to be ended.

[0112] Step 230: end the current backward search, and obtain the backward search trajectory points.

[0113] Step 232: eliminate any trajectory point with a distance less than the dense search step, and combine the remaining trajectory points to obtain the current constraint search result.

[0114] In combination Figure 3a , the result schematic diagram of the prior art equal step search is shown, Figure 3b , the result schematic diagram of the dynamic step forward search adopted by the present application is shown, Figure 3c , the result schematic diagram of the dynamic step backward search adopted by the present application is shown, Figure 3d , the result schematic diagram of the combined processing of the forward search trajectory points and the backward search trajectory points adopted by the present application is shown. The straight line is the to-be-searched trajectory C, the blank rhombus is the unsearched trajectory point, the black rhombus is the searched trajectory point, and the circular region is the correlation data. The to-be-searched trajectory and the correlation data used are the same, and by comparison, after the forward search and the backward search, the searched trajectory points are more, and the correlation data is more complete. Figure 3b , and Figure 3c , the different search results can be obtained through the forward and backward double-direction collaborative dynamic step processing, and then the two search results are combined and processed to obtain the final search result shown in Figure 3d , compared with the search result of Figure 3a , at least the following effect advantages exist:

[0115] - reduce the total amount of invalid constraints: through the long and short step double-direction search method, the short step is used for searching in the sensitive areas such as the edge of the overlapping area, and the long step is used for searching in the non-sensitive areas such as the outside of the overlapping area, so that the total amount of invalid constraints is effectively reduced, and the calculation amount is significantly reduced;

[0116] - adaptive adjustment of step: the search step can be adaptively adjusted according to the characteristics of the current search area. In the dense area, a small step is used to ensure sufficient constraints; in the sparse area, a large step is switched to reduce unnecessary calculation. This adaptive adjustment mechanism makes the search process more efficient and avoids the limitations of fixed step search;

[0117] - Improve data consistency and accuracy: Ensure the constraints of the edge of the overlap area, the edge of the overlap area and other sensitive areas are crucial to the consistency and accuracy of the data. The traditional method often cannot provide enough constraints in these areas due to the fixed search step, resulting in data layering, ambiguity and other quality problems. The present application ensures that sensitive areas such as the edge of the overlap area have enough constraints through short-step search, thereby significantly improving the consistency and accuracy of the data.

[0118] Therefore, the present application uses an innovative front-back bidirectional sparse adaptive step search method to significantly reduce the total amount of constraints while ensuring that sensitive areas such as the edge of the overlap area have enough constraints, thereby improving the calculation efficiency and ensuring the consistency and accuracy of the data. This method is particularly suitable for large-scale scene reconstruction in the fields of autonomous driving, mobile robots and surveying, and can significantly improve the performance and application effect of these fields.

[0119] It should be noted that for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the order of the described actions, because according to the present application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0120] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0121] Figure 4 The structure block diagram of the device for constraint search based on data association provided by an embodiment of the present application is shown in FIG. 1. Figure 4The device 400 for constraint search based on data association in the embodiment can include a determination module 401, an initialization module 402, a search module 403 and a merging module 404. The determination module 401 is configured to determine a to-be-searched trajectory and associated data having a constraint relationship with the to-be-searched trajectory. The constraint relationship includes at least one or more of a position constraint, a distance constraint and a scene constraint. The initialization module 402 is configured to initialize a search parameter required for this search based on a length of the to-be-searched trajectory. The search parameter includes a dense search step and a sparse search step. The search module 403 is configured to search for a forward search trajectory point and a backward search trajectory point associated with the associated data from the to-be-searched trajectory according to the initialized dense search step and sparse search step by using a dynamic adjustment search step and a forward-backward bidirectional collaborative search mode. The merging module 404 is configured to merge the forward search trajectory point and the backward search trajectory point to obtain a constraint search result.

[0122] It should be noted that part or all of the device for constraint search based on data association in the embodiment can be an application located at a local terminal, or can be a plug-in or a software development kit (SDK) and the like located in the application at the local terminal, or can be a processing engine located in a network side server, or can be a distributed system located at the network side, and the embodiment does not particularly limit this.

[0123] It can be understood that the application can be a native application (nativeApp) installed on the local terminal, or can be a web application (webApp) of a browser on the local terminal, and the embodiment does not limit this.

[0124] Optionally, in a possible implementation manner of the embodiment, when initializing the search parameter required for this search based on the length of the to-be-searched trajectory, the initialization module 402 is specifically configured to initialize the dense search step as a first proportional range of the length of the to-be-searched trajectory, and the first proportional range is [1 / 100, 1 / 200]; and initialize the sparse search step as a second proportional range of the length of the to-be-searched trajectory, and the second proportional range is [1 / 20, 1 / 50].

[0125] Optionally, in a possible implementation manner of the embodiment, when the search module 403 searches the forward search track points and the backward search track points associated with the constraint of the associated data from the to-be-searched track according to the initialized dense search step and sparse search step, the dynamic adjustment of the search step and the forward and backward collaborative search, the to-be-searched track is specifically processed by downsampling, and the associated data is stored as a k-d tree structure; from the head end to the tail end of the to-be-searched track after downsampling: based on the initialized dense search step and sparse search step, the step in the search process is dynamically adjusted to search for the forward search track points associated with the k-d tree structure; from the tail end to the head end of the to-be-searched track after downsampling: based on the initialized dense search step and sparse search step, the step in the search process is dynamically adjusted to search for the backward search track points associated with the k-d tree structure.

[0126] Optionally, in a possible implementation manner of the embodiment, the search parameters further include a dense segment continuity threshold, and the dense segment continuity threshold is positively correlated with the constraint requirement degree; when the search module 403 dynamically adjusts the step in the search process based on the initialized dense search step and sparse search step to search for the forward search track points associated with the k-d tree structure, the dense search step is specifically used to traverse each track point in the to-be-searched track after downsampling; it is judged whether the track point associated with the k-d tree structure is searched in the sparse traversal process; if yes, the searched track point is recorded as the start point of the dense segment, the search step is adjusted from the sparse search step to the dense search step, and the subsequent unsearched track points are traversed; otherwise, the operation of judging whether the track point associated with the k-d tree structure is searched in the sparse traversal process is returned; it is judged whether the track point associated with the k-d tree structure is searched in the dense traversal process, and the number of track points recorded in the dense segment does not reach the dense segment continuity threshold; if yes, the searched track point is recorded to the dense segment; otherwise, the search step is adjusted from the dense search step to the sparse search step, and the subsequent unsearched track points are continuously traversed; the above operation is repeated until each track point in the to-be-searched track is traversed, and the track points recorded in all dense segments are counted as the forward search track points.

[0127] Optionally, in a possible implementation manner of the embodiment, when the search module 403 does not search the trajectory point of the constraint associated with the k-d tree structure in the dense traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold, the method further includes: starting from the trajectory point corresponding to the start point of the current dense segment, adjusting the search step from the dense search step to a temporary search step, and continuing to traverse the subsequent unsearched trajectory points; the temporary search step is smaller than the dense search step; determining whether the trajectory point of the constraint associated with the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold; if yes, recording the searched trajectory point in the current search to the dense segment; otherwise, adjusting the search step from the temporary search step to a sparse search step, and continuing to traverse the subsequent unsearched trajectory points.

[0128] Optionally, in a possible implementation manner of the embodiment, when the search module 403 does not search the trajectory point of the constraint associated with the k-d tree structure in the dense traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold, the method further includes: starting from the trajectory point corresponding to the start point of the current dense segment, adjusting the search step from the dense search step to a temporary search step, and continuing to traverse the subsequent unsearched trajectory points; the temporary search step is smaller than the dense search step; determining whether the trajectory point of the constraint associated with the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold; if yes, recording the searched trajectory point in the current search to the dense segment; otherwise, adjusting the search step from the temporary search step to a sparse search step, and continuing to traverse the subsequent unsearched trajectory points.

[0129] Optionally, in a possible implementation manner of the embodiment, when the search module 403 does not search the trajectory point of the constraint associated with the k-d tree structure in the dense traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold, the method further includes: starting from the trajectory point corresponding to the start point of the current dense segment, adjusting the search step from the dense search step to a temporary search step, and continuing to traverse the subsequent unsearched trajectory points; the temporary search step is smaller than the dense search step; determining whether the trajectory point of the constraint associated with the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded in the dense segment does not reach the dense segment continuous threshold; if yes, recording the searched trajectory point in the current search to the dense segment; otherwise, adjusting the search step from the temporary search step to a sparse search step, and continuing to traverse the subsequent unsearched trajectory points.

[0130] In this embodiment, after the to-be-searched trajectory and the associated data are determined, the search parameters required for this search can be initialized based on the length of the to-be-searched trajectory, and the forward search trajectory points and the backward search trajectory points that are associated with the associated data and are constrained can be searched from the to-be-searched trajectory by using the way of dynamically adjusting the search step length and the way of forward and backward collaborative search according to the initialized dense search step length and sparse search step length. The forward search trajectory points and the backward search trajectory points are combined to obtain the constraint search result of this time. Through the way of forward and backward collaborative search combined with dynamically adjusting the step length, it can be ensured that there are sufficient constraints in sensitive areas such as the edge of the overlapping area, while the number of constraints is prevented from being doubled due to too small step length, so as to effectively control the calculation amount while taking into account reasonable constraints, and ensure the calculation efficiency.

[0131] One embodiment of the present application provides a computer readable storage medium, the storage medium stores at least one instruction, the at least one instruction is loaded and executed by a processor to implement the method for constraint search based on data association as described above.

[0132] One embodiment of the present application provides an electronic device, the electronic device includes a processor and a memory, the memory stores at least one instruction, the instruction is loaded and executed by the processor to implement the method for constraint search based on data association as described above.

[0133] In the technical solution of the present application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in the technical solution comply with relevant laws and regulations and do not violate public order and good customs.

[0134] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present application described and / or claimed in this document.

[0135] As Figure 5As shown, the electronic device 500 includes a computing unit 501 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 502 or a computer program loaded into a random access memory (RAM) 503 from a storage unit 508. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0136] A plurality of components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, and the like; an output unit 507, such as various types of displays, a speaker, and the like; a storage unit 508, such as a magnetic disk, an optical disk, and the like; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0137] The computing unit 501 can be various general-purpose and / or special-purpose processing components having processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, and the like. The computing unit 501 performs various methods and processes described above, such as the method of performing a constrained search based on data association. For example, in some embodiments, the method of performing a constrained search based on data association can be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the method of performing a constrained search based on data association described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the method of performing a constrained search based on data association by any other appropriate means, such as by means of firmware.

[0138] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0139] Program code for carrying out methods of the present application can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces a means for implementing the functions / acts specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, or entirely on a remote machine or server.

[0140] In the context of the present application, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0141] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0142] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0143] The computer system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server is generally established by computer programs running on the respective computers and having a client-server relationship to each other. The servers can be cloud servers, servers of a distributed system, or servers combined with a blockchain.

[0144] It should be understood that various forms of flow shown above can be used, with steps reordered, added, or removed. For example, steps recited in the present disclosure can be performed in parallel, in series, or in a different order, without limitation, as long as the desired results of the technology of the present disclosure are achieved.

[0145] The specific embodiments described above are not intended to limit the scope of the present application. Those skilled in the art will understand that various modifications, combinations, sub-combinations, and alternatives of the specific embodiments described above can be made without departing from the spirit and scope of the present application. Accordingly, any modification, equivalent substitution, and improvement not described above are intended to be included within the scope of the present application.

Claims

1. A method for constraint search based on data association, characterized in that, Comprise: Determine the trajectory to be searched and the associated data associated with the trajectory to be searched; wherein the associated constraint relationship at least contains one or more of the following: position constraint, distance constraint and scene constraint; Based on the length of the trajectory to be searched, initialize the search parameters required for this search, wherein the search parameters include: dense search step and sparse search step; According to the initialized dense search step and sparse search step, search the forward search trajectory point and the backward search trajectory point associated with the associated data from the trajectory to be searched by adopting the way of dynamic adjustment of sparse and dense search step and forward and backward bidirectional collaborative search; The forward search trajectory point and the backward search trajectory point are combined to obtain the constraint search result of this time.

2. The method of claim 1, wherein, Based on the length of the trajectory to be searched, initialize the search parameters required for this search, specifically including: The dense search step is initialized to the first proportion range of the length of the trajectory to be searched, and the first proportion range is [1 / 100, 1 / 200]; The sparse search step is initialized to the second proportion range of the length of the trajectory to be searched, and the second proportion range is [1 / 20, 1 / 50].

3. The method of claim 1, wherein, According to the initialized dense search step and sparse search step, search the forward search trajectory point and the backward search trajectory point associated with the associated data from the trajectory to be searched by adopting the way of dynamic adjustment of sparse and dense search step and forward and backward bidirectional collaborative search, specifically including: Downsample the trajectory to be searched, and store the associated data as a k-d tree structure; From the head to the tail of the downsampled trajectory to be searched: based on the initialized dense search step and sparse search step, dynamically adjust the step in the search process to search the forward search trajectory point associated with the k-d tree structure; From the tail to the head of the downsampled trajectory to be searched: based on the initialized dense search step and sparse search step, dynamically adjust the step in the search process to search the backward search trajectory point associated with the k-d tree structure.

4. The method of claim 3, wherein, The search parameters further include: dense segment continuous threshold, which is positively correlated with the constraint requirement degree of this time; Based on the initialized dense search step and sparse search step, dynamically adjust the step in the search process to search the forward search trajectory point associated with the k-d tree structure, specifically including: Traverse each trajectory point in the downsampled trajectory to be searched with the sparse search step; Determine whether a trajectory point associated with the k-d tree structure is searched in the sparse traversal process; If yes, record the searched trajectory point as the starting point of the dense segment, adjust the search step from the sparse search step to the dense search step, and traverse the subsequent unsearched trajectory points; otherwise, return to the operation of determining whether a trajectory point associated with the k-d tree structure is searched in the sparse traversal process; Determine whether a trajectory point associated with the k-d tree structure is searched in the dense traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold; If yes, record the searched trajectory point to the dense segment; otherwise, adjust the search step from the dense search step to the sparse search step, and continue to traverse the subsequent unsearched trajectory point. Repeat the above operation until each trajectory point in the to-be-searched trajectory is traversed, and count the trajectory points recorded by all dense segments as the forward search trajectory point.

5. The method of claim 4, wherein, When no trajectory point associated with the constraint of the k-d tree structure is searched in the dense traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold, the method further comprises: Start from the trajectory point corresponding to the start point of the current dense segment, adjust the search step from the dense search step to the temporary search step, and continue to traverse the subsequent unsearched trajectory point; wherein the temporary search step is smaller than the dense search step; Determine whether a trajectory point associated with the constraint of the k-d tree structure is searched in the current traversal process, and the number of trajectory points recorded by the dense segment does not reach the dense segment continuous threshold; If yes, record the searched trajectory point to the dense segment; otherwise, adjust the search step from the temporary search step to the sparse search step, and continue to traverse the subsequent unsearched trajectory point.

6. The method of claim 4, wherein, If the associated constraint relationship is a position constraint, searching for a trajectory point associated with the constraint of the k-d tree structure means that the position coordinates of the trajectory point in the to-be-searched trajectory are the same as or satisfy a set proximity degree with the trajectory point in the k-d tree structure; If the associated constraint relationship is a distance constraint, searching for a trajectory point associated with the constraint of the k-d tree structure means that the distance between the trajectory point in the to-be-searched trajectory and the trajectory point in the k-d tree structure is less than a set threshold; If the associated constraint relationship is a scene constraint, searching for a trajectory point associated with the constraint of the k-d tree structure means that the scene of the trajectory point in the to-be-searched trajectory is the same as or satisfies a set similarity degree with the trajectory point in the k-d tree structure; or the scene feature of the trajectory point in the to-be-searched trajectory is the same as or satisfies a set proximity degree with the trajectory point in the k-d tree structure.

7. The method according to any one of claims 1 to 4, 6, wherein Merge the forward search trajectory point and the backward search trajectory point to obtain the current constraint search result, specifically including: Remove any trajectory point in the forward search trajectory point and the backward search trajectory point that is separated by a distance less than the dense search step, and merge the remaining trajectory points to obtain the current constraint search result; or For any trajectory point in the forward search trajectory point and the backward search trajectory point that is in the dense segment, determine whether the distance between the current trajectory point and the nearest trajectory point is greater than the dense search step, if yes, do not process, otherwise, remove the trajectory point; for any trajectory point in the forward search trajectory point and the backward search trajectory point that is not in the dense segment, determine whether the distance between the current trajectory point and the nearest trajectory point is greater than the sparse search step, if yes, do not process, otherwise, remove the trajectory point; merge the remaining trajectory points to obtain the current constraint search result.

8. An apparatus for constraint search based on data association, the apparatus comprising: The method comprises: A determination module is configured to determine a to-be-searched trajectory and associated data having an associated constraint relationship with the to-be-searched trajectory; wherein the associated constraint relationship includes one or more of a position constraint, a distance constraint, and a scene constraint. An initialization module is configured to initialize search parameters required for the current search based on a length of the trajectory to be searched, wherein the search parameters include a dense search step and a sparse search step; A search module is configured to search for a forward search trajectory point and a backward search trajectory point associated with the associated data constraint from the trajectory to be searched in a manner of dynamic adjustment of search step and bidirectional collaborative search according to the initialized dense search step and sparse search step; A merging module is configured to merge the forward search trajectory point and the backward search trajectory point to obtain a result of the current constraint search.

9. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are configured to cause the computer to perform the method according to any one of claims 1-7.

10. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-7.

Images