Reference map acquisition method, device, equipment and storage medium

Abstract

The present application provides a method, apparatus, device, and storage medium for obtaining a baseline map, including: obtaining multiple submaps generated for different areas in a target scene; determining a scene information volume evaluation index corresponding to each of the multiple submaps based on the map data corresponding to each of the multiple submaps; and determining a target submap from the multiple submaps as a baseline map based on the scene information volume evaluation index. The baseline map is the starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene. In this solution, the scene information volume corresponding to each submap is measured by setting a scene information volume evaluation index. Furthermore, based on the scene information volume evaluation index, the submap containing the most scene information can be selected from the multiple submaps as the baseline map. When the multiple submaps are spliced ​​together, the other submaps can be well matched with the baseline map, thereby quickly splicing together a high-quality map.

Description

Technical Field

[0001] The present invention relates to the field of Internet technology, and in particular to a reference map acquisition method, device, equipment and storage medium. Background Art

[0002] With the development of artificial intelligence, robots are now able to perform different types of tasks in diverse scenarios. For example, sweeping robots can be used for floor cleaning, while warehouse robots can be used for cargo sorting. Robots are able to operate in various scenarios because they have pre-built maps of the current scene. Based on these maps, the robots can locate their specific position within the scene and perform path planning, enabling autonomous movement. Simultaneous Localization and Mapping (SLAM) is commonly used to construct scene maps.

[0003] When mapping a large environment, it's usually necessary to build maps for different areas of the target environment separately. From these maps, a base map is selected as the baseline. The target environment's map is then constructed by stitching other maps together with the base map. The choice of base map directly impacts the quality and speed of the stitching process. Summary of the Invention

[0004] Embodiments of the present invention provide a reference map acquisition method, apparatus, device, and storage medium for improving the quality and speed of map splicing based on the reference map.

[0005] In a first aspect, an embodiment of the present invention provides a method for obtaining a reference map, the method comprising:

[0006] Obtain multiple submaps generated for different areas in the target scene;

[0007] Determining scene information quantity evaluation indicators corresponding to the plurality of submaps respectively based on the map data corresponding to the plurality of submaps respectively;

[0008] According to the scene information evaluation index, a target submap is determined from the multiple submaps as a reference map, where the reference map is a starting submap when the multiple submaps are spliced ​​to generate a map of the target scene.

[0009] In a second aspect, an embodiment of the present invention provides a reference map acquisition device, the device comprising:

[0010] An acquisition module, used to acquire multiple sub-maps generated for different areas in the target scene;

[0011] The processing module is configured to determine, based on the map data corresponding to the multiple submaps, scene information quantity evaluation indicators corresponding to the multiple submaps; and based on the scene information quantity evaluation indicators, determine a target submap from the multiple submaps as a reference map, the reference map being a starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene.

[0012] In a third aspect, an embodiment of the present invention provides an electronic device comprising: an inertial measurement unit, a camera, a laser sensor, a memory, a processor, and a communication interface; wherein the memory stores executable code, and when the executable code is executed by the processor, the processor executes the reference map acquisition method as described in the first aspect.

[0013] In a fourth aspect, an embodiment of the present invention provides a non-transitory machine-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor can at least implement the baseline map acquisition method described in the first aspect.

[0014] In an embodiment of the present invention, when constructing a map corresponding to a target scene, the target scene is first divided into different regions, and then maps corresponding to each region, referred to as submaps, are constructed. Subsequently, a target submap is selected from multiple submaps as a baseline map, i.e., the starting submap for merging the multiple submaps. The multiple submaps are then spliced ​​together based on the baseline map to obtain a map corresponding to the target scene. In this solution, a scene information evaluation metric is set to measure the amount of scene information corresponding to each submap. Based on this metric, the submap containing the most scene information can be selected from the multiple submaps as the baseline map. Because the baseline map contains the most scene information and therefore covers the most target scenes, when the multiple submaps are spliced ​​together, the other submaps can better match the baseline map, resulting in a high-quality map that can be quickly spliced ​​together. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 A flowchart of a reference map acquisition method provided by an embodiment of the present invention;

[0017] Figure 2A flowchart of another reference map acquisition method provided by an embodiment of the present invention;

[0018] Figure 3 A schematic diagram of a common view feature point provided by an embodiment of the present invention;

[0019] Figure 4 A schematic structural diagram of a reference map acquisition device provided by an embodiment of the present invention;

[0020] Figure 5 This is a schematic structural diagram of an electronic device provided in this embodiment. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0023] Figure 1 A flowchart of a reference map acquisition method provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method includes the following steps:

[0024] 101. Obtain multiple submaps generated for different areas in the target scene.

[0025] 102. Determine scene information quantity evaluation indicators corresponding to the plurality of sub-maps, respectively, based on the map data corresponding to the plurality of sub-maps.

[0026] 103. Determine a target submap from the multiple submaps as a reference map based on scene information volume evaluation indicators corresponding to the multiple submaps, wherein the reference map is a starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene.

[0027] The reference map acquisition method provided by the embodiment of the present invention can be applied to a robot, or can be applied to a robot client installed on a smart phone, a laptop computer or a server. The robot client can interact with the robot through network communication to exchange information and instructions.

[0028] When mapping a target scene (i.e., building a map), it is usually necessary to first use a perception device (such as an inertial measurement unit, a laser sensor, a camera, etc.) to perceive the target scene, and then build a map based on the scene information obtained by the perception device (such as point cloud data, visual images, etc.).

[0029] In actual applications, when mapping some larger target scenes (such as outdoor scenes), it is difficult to obtain all the scene information of the entire target scene at one time, and there is a defect that if an error occurs during the mapping process, the map needs to be rebuilt. To this end, the target scene is usually divided into different areas first, and then the corresponding scene information is collected by sensing devices for different areas respectively, and sub-maps of each area are constructed based on the corresponding scene information of each area. Finally, the sub-maps of different areas are spliced ​​to obtain a map of the target scene. In this embodiment, the maps corresponding to different areas are called sub-maps to distinguish them from the map corresponding to the target scene.

[0030] To facilitate the splicing of multiple submaps, when dividing the target scene into regions, there is a common view area between the two different regions. The common view area refers to an area that can be perceived by the sensing devices in the two regions at the same time, and can also be understood as an overlapping area between the two regions. It can be understood that if there is a common view area X between region A and region B, then the submap Ma corresponding to region A and the submap Mb corresponding to region B both contain partial maps corresponding to the common view area X, so that the submap Ma and the submap Mb can be spliced ​​based on the partial map corresponding to the common view area X. For example, the spatial relative transformation between the submap Ma and the submap Mb can be calculated by constructing a linear geometric constraint of the partial map corresponding to the common view area X, and finally the splicing of the submap Ma and the submap Mb can be achieved through the relative transformation.

[0031] When stitching together multiple submaps corresponding to different areas of the target scene, a starting submap must be determined from the submaps, referred to as the baseline map in this embodiment. The remaining submaps are then stitched together with the baseline map to create a map of the target scene. The selection of the baseline map affects the efficiency and quality of the stitching. For example, if a submap is randomly selected from multiple submaps as the baseline map and covers less of the target scene, the remaining submaps may not be able to be stitched together with it, ultimately leading to stitching failure or poor map accuracy.

[0032] To improve the efficiency of stitching multiple submaps and the accuracy of the resulting map, this solution, after obtaining multiple submaps generated for different areas of the target scene, determines a scene information evaluation index for each submap based on the map data corresponding to each submap. This index measures the amount of scene information corresponding to each submap. Furthermore, based on this index, the submap containing the most scene information is selected from the multiple submaps as the baseline map. It is understood that the more scene information a baseline map contains, the more target scenes and corresponding common view areas it covers. This ensures that when stitching multiple submaps together, the other submaps can better match the baseline map, resulting in a faster, high-quality map.

[0033] In the specific implementation process, parameters that can describe the amount of scene information contained in the sub-map can be set, and based on each parameter, the scene information amount evaluation index corresponding to each of the multiple sub-maps can be determined. Figure 2 Provide explanation.

[0034] Figure 2 A flowchart of another reference map acquisition method provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the method includes the following steps:

[0035] 201. Acquire multiple submaps generated for different regions of the target scene, each submap corresponding to map data. The map data includes: two-dimensional feature points corresponding to visual images collected for different regions of the target scene, three-dimensional spatial points corresponding to the two-dimensional feature points in the target scene, movement data during the collection of the visual images, and size information of the corresponding submaps. The visual images are used to generate the submaps.

[0036] In practical applications, maps can be categorized in a variety of ways. For example, based on the type of data used to generate the map, maps can be divided into visual maps, point cloud maps, and so on. It is understood that different types of maps correspond to different parameters that describe the amount of scene information contained in the map. For ease of understanding, this embodiment uses a visual submap as an example, but this is not limiting.

[0037] For any area Y in the target scene, when generating a submap (visual map) of area Y, the camera first acquires a video stream of area Y. Then, multiple key frames used to generate the submap are determined from the video stream, which is the visual image in this embodiment. Two-dimensional feature points are then extracted from the visual image, and the corresponding three-dimensional spatial points (also called map points) in the target scene are determined. Finally, combined with the motion data when the visual image was collected, a submap of area Y is generated, which corresponds to size information. The motion data includes: a motion path and pose data matching the motion path. The pose data includes rotation information and translation information.

[0038] Based on the above submap generation process, in this embodiment, the 2D feature points corresponding to the visual image, the 3D spatial points corresponding to the 2D feature points in the target scene, the movement data during the visual image acquisition, and the size information of the corresponding submap are referred to as the submap's map data. The map data is used to determine the values ​​of parameters that describe the amount of scene information contained in the submap.

[0039] In this embodiment, the parameters used to describe the amount of scene information contained in the sub-map include: the number of common view feature points, the number of common view space points, the length and area of ​​the moving path.

[0040] 202. Determine the number of common-view feature points corresponding to the first submap based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to the second submap. The common-view feature points are two-dimensional feature points contained in both the first submap and the second submap. The first submap is any one of the multiple submaps, and the second submap is any submap of the multiple submaps except the first submap.

[0041] Among them, the common view feature points correspond to the aforementioned common view areas. The more common view feature points there are between sub-maps, the more common view areas the sub-maps correspond to, and the more scene information the sub-maps contain.

[0042] For ease of understanding, first combine Figure 3 , an example is given to illustrate the process of determining the number of common view feature points between two sub-maps.

[0043] Figure 3 Schematic diagram of a common view feature point provided by an embodiment of the present invention. Figure 3 As shown in the figure, there are two submaps: submap M1 and submap M2. There are two 3D space points P1 and P2 in the target scene. P1 and P2 are located in the common view area of ​​the corresponding areas of submap M1 and submap M2. Among them, P1 is used to generate the visual images of submap M1 and submap M2 respectively ( Figure 3In the visual images used to generate submaps M1 and M2, P2 corresponds to a two-dimensional feature point F1. F1 and F2 may be located in different visual images. For example, visual image 1 contains F1, while visual image 2 contains both F1 and F2.

[0044] Based on the above assumptions, according to the two-dimensional feature points corresponding to submap M1 and the two-dimensional feature points corresponding to submap M2, it can be determined that submap M1 and submap M2 both contain two-dimensional feature points F1 and F2. Therefore, F1 and F2 are the common view feature points of submap M1 and submap M2. Figure 3 As shown, the two-dimensional feature points corresponding to the submap M1 include 2 F1 and 2 F2, and the two-dimensional feature points corresponding to the submap M2 include 2 F1 and 3 F2. Therefore, the number of common view feature points between the submap M1 and the submap M2 is O 12 It is 9, including 4 F1 and 5 F2.

[0045] During the specific implementation process, when determining the common view feature points of the first submap and the second submap, as an optional method, a distance calculation can be performed based on the descriptor corresponding to each two-dimensional feature point of the first submap and the descriptor corresponding to the two-dimensional feature point of the second submap, such as calculating the Euclidean distance or the Hamming distance. If the distance corresponding to the two descriptors is less than a set threshold, the two-dimensional feature points corresponding to the two descriptors are determined to be common view feature points.

[0046] In practical applications, the number of the second sub-map may be one or more.

[0047] For example, assuming that the target scene is divided into n different areas, n sub-maps are generated accordingly, namely sub-map M1, sub-map M2, ..., sub-map Mn, where n is an integer greater than 1.

[0048] If n is equal to 2, the number of the second submap is one, and the number of common view feature points corresponding to the first submap is the number of common view feature points between the first submap and the second submap. For example, if the first submap is submap M1 and the second submap is submap M2, then the number of common view feature points corresponding to submap M1 is O. 1s That is, the number of common view feature points O between submap M1 and submap M2 12 , that is, O 1s =O 12 Similarly, the number of common view feature points corresponding to submap M2 is O 2s =O 21 , where O 12 =O 21 .

[0049] If n is greater than 2, then the number of second submaps is multiple, i.e., (n-1), and the number of common view feature points corresponding to the first submap is the sum of the number of common view feature points between the first submap and each second submap. In the specific implementation, the number of common view feature points between the first submap and each second submap is first determined based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to each second submap. Then, the sum of the number of common view feature points between the first submap and each second submap is determined as the number of common view feature points of the first submap.

[0050] For example, if the first submap is submap M1, the second submap corresponds to submap M2, submap M3, ..., submap Mn. When determining the number O of common view feature points corresponding to submap M1, 1s When , first calculate the number of common view feature points O between submap M1 and submap M2 based on the two-dimensional feature points corresponding to submap M1 and the two-dimensional feature points corresponding to submap M2. 12 , according to the two-dimensional feature points corresponding to submap M1 and the two-dimensional feature points corresponding to submap M3, calculate the number of common view feature points O between submap M1 and submap M3 13 , ..., according to the two-dimensional feature points corresponding to the submap M1 and the two-dimensional feature points corresponding to the submap Mn, calculate the number of common view feature points O between the submap M1 and the submap Mn 1n After that, the number of common view feature points corresponding to the submap M1 is determined. 1s is the sum of the number of common view feature points corresponding to submap M1, submap M2, submap M3, ..., submap Mn, that is, O 1s =O 12 +O 13 +...+O 1n .

[0051] 203. Determine the number of common view space points corresponding to the first submap based on the three-dimensional space points corresponding to the common view feature points corresponding to the first submap.

[0052] It is understandable that the process of shooting with a camera is actually the process of converting three-dimensional space into a two-dimensional visual image, and the two-dimensional feature points in the visual image all have corresponding three-dimensional space points in the three-dimensional space.

[0053] The common view area is characterized from a two-dimensional perspective using common view feature points in step 202, and from a three-dimensional perspective using common view space points in step 203. It is understood that for the same common view area between two submaps, the common view feature points and common view space points between the two submaps should match, because the common view feature points and common view space points represent the same area.

[0054] Since the common view feature points between any two submaps have already been determined in step 202, and the map data contains the corresponding 3D spatial points in the target scene for the 2D feature points corresponding to each submap, the 3D spatial points corresponding to the common view feature points between any two submaps can be determined, and the determined 3D spatial points are used as the common view spatial points between the two submaps. The greater the number of common view spatial points between the submaps, the more common view areas the submaps correspond to, and the more scene information the submaps contain.

[0055] For example, still Figure 3 As an example, submap M1 and submap M2 in Figure 3 Corresponding assumptions, after determining that the common view feature points between submap M1 and submap M2 are two-dimensional feature points F1 and F2, it is further determined that the three-dimensional space point corresponding to the two-dimensional feature point F1 is P1, and the three-dimensional space point corresponding to the two-dimensional feature point F2 is P2. Thus, the common view space points between submap M1 and submap M2 are three-dimensional space points P1 and P2, and the number of common view space points between submap M1 and submap M2 is L. 12 is 2.

[0056] When determining the number of common view space points corresponding to any submap (i.e., the first submap), it is still assumed that the target scene is divided into n different areas, and n submaps are generated accordingly, namely submap M1, submap M2, ..., submap Mn, where n is an integer greater than 1.

[0057] If n is equal to 2, the number of the second submap is one, and the number of common view space points corresponding to the first submap is the number of common view space points between the first submap and the second submap. For example, if the first submap is submap M1 and the second submap is submap M2, then the number of common view space points corresponding to submap M1 is L 1s That is, the number of common view points L between submap M1 and submap M2 12 , that is, L 1s =L 12 Similarly, the number of common view space points L corresponding to the submap M2 is 2s =L 21 , where L 12 =L 21 .

[0058] If n is greater than 2, then there are multiple second submaps, namely (n-1), and the number of common view feature points corresponding to the first submap is the sum of the number of common view space points between the first submap and each second submap. In the specific implementation, the number of common view space points between the first submap and each second submap is first determined based on the three-dimensional spatial points corresponding to the common view feature points between the first submap and each second submap. Then, the sum of the number of common view space points between the first submap and each second submap is determined as the number of common view space points corresponding to the first submap.

[0059] For example, if the first submap is submap M1, the second submap corresponds to submap M2, submap M3, ..., submap Mn. When determining the number L of common view feature points corresponding to submap M1, 1s When , first calculate the number of common view space points L between submap M1 and submap M2 based on the three-dimensional space points corresponding to the common view feature points between submap M1 and submap M2 12 , calculate the number of common view space points L between submap M1 and submap M3 based on the three-dimensional space points corresponding to the common view feature points between submap M1 and submap M3 13 , ..., according to the 3D space points corresponding to the common view feature points between submap M1 and submap Mn, calculate the number of common view space points L between submap M1 and submap Mn 1n After that, determine the number of common view space points L corresponding to the submap M1 1s is the sum of the number of common view space points corresponding to submap M1, submap M2, submap M3, ..., submap Mn, that is, L 1s =L 12 +L 13 +...+L 1n .

[0060] 204. Determine the length of the movement path corresponding to the first submap based on the movement data corresponding to the first submap; and determine the area corresponding to the first submap based on the size information corresponding to the first submap.

[0061] The movement path can be understood as the movement path taken when collecting visual images used to generate the first submap. If the visual images used to generate the first submap are collected by a camera configured on the robot, the movement path of the robot when collecting visual images used to generate the first submap is the movement path corresponding to the first submap.

[0062] It is understandable that the longer the moving path of the submap is, the more scene information is collected, and the larger the amount of scene information corresponding to the submap is. The larger the area of ​​the submap is, the more target scenes are covered and the more scene information is contained.

[0063] 205. Determine scene information quantity evaluation indicators corresponding to the multiple submaps according to the number of common view feature points, the number of common view space points, the length and area of ​​the moving paths corresponding to the multiple submaps.

[0064] When determining the scene information evaluation indicators corresponding to multiple sub-maps, the multiple parameters corresponding to the multiple sub-maps (i.e., the number of common view feature points, the number of common view space points, the length and area of ​​the moving path) are comprehensively considered. By determining the scores corresponding to different parameters, the scene information evaluation indicators corresponding to the multiple sub-maps are determined.

[0065] Specifically, the score of the number of common view feature points of the first submap is determined based on the ratio of the number of common view feature points corresponding to the first submap to the total number of common view feature points, where the total number of common view feature points is the sum of the numbers of common view feature points corresponding to the multiple submaps.

[0066] For ease of understanding, for example, it is still assumed that the target scene is divided into n different areas, and n sub-maps are generated accordingly, namely sub-map M1, sub-map M2, ..., sub-map Mn, where n is an integer greater than 1.

[0067] The number of common view feature points corresponding to the submap MX (the value of X ranges from 1 to n) can be expressed as follows:

[0068] S OX = O xs / O num (1)

[0069] Among them, O xs Indicates the number of common view feature points corresponding to the submap MX, O ij represents the number of common view feature points between submap i and submap j, O 1s is the number of common view feature points corresponding to sub-map M1, O 2s is the number of common view feature points corresponding to sub-map M2, O 3s is the number of common view feature points corresponding to sub-map M3, ..., O ns is the number of common view feature points corresponding to the sub-map Mn.

[0070] The score of the number of common view space points of the first submap is determined based on the ratio of the number of common view space points corresponding to the first submap to the total number of common view space points. The total number of common view space points is the sum of the number of common view space points corresponding to multiple submaps. The score of the number of common view space points of submap MX can be expressed as follows:

[0071] S LX = L xs / L num(2)

[0072] Among them, L xs Indicates the number of common view space points corresponding to the submap MX, L ij represents the number of common view points between submap i and submap j, L 1s is the number of common view space points corresponding to the sub-map M1, L 2s is the number of common view space points corresponding to the sub-map M2, L 3s is the number of common view space points corresponding to submap M3, ..., L ns is the number of common view space points corresponding to the submap Mn.

[0073] The moving path length score of the first submap is determined based on the ratio of the moving path length corresponding to the first submap to the total moving path length. The total moving path length is the sum of the moving path lengths corresponding to multiple submaps. The moving path length score of submap MX can be expressed as follows:

[0074] S PX = PX / (P1+P2+…+Pn) (3)

[0075] Among them, PX is the moving path length corresponding to the submap MX, P1 is the moving path length corresponding to the submap M1, P2 is the moving path length corresponding to the submap M2, ..., Pn is the moving path length corresponding to the submap Mn.

[0076] The area score of the first submap is determined based on the ratio of the area corresponding to the first submap to the total area. The total area is the sum of the areas corresponding to the multiple submaps. The area score of the submap MX can be expressed as follows:

[0077] S AX = AX / (A1+A2+…+An) (4)

[0078] Among them, AX is the area corresponding to the submap MX, A1 is the area corresponding to the submap M1, A2 is the area corresponding to the submap M2, ..., An is the area corresponding to the submap Mn.

[0079] In practical applications, when determining the scene information quantity evaluation index corresponding to the first submap based on the score of the number of common view feature points, the score of the number of common view space points, the score of the movement path length, and the score of the area of ​​the first submap, a corresponding weight can optionally be pre-set for each score. In a specific implementation, a first product of the score of the number of common view feature points and a first set weight, a second product of the score of the number of common view space points and a second set weight, a third product of the score of the movement path length and a third set weight, and a fourth product of the score of the area and a fourth set weight are determined for the first submap; the sum of the first, second, third, and fourth products is used as the scene information quantity evaluation index for the first submap.

[0080] Still taking the submap MX as an example, its corresponding scene information evaluation index can be expressed as the following formula (5):

[0081] SX=W O *S OX +W L *S LX +W P *S PX +W A *S AX (5)

[0082] Among them, W O Indicates the weight corresponding to the number of common view feature points, W L Indicates the weight corresponding to the number of points in the common view space, W P Represents the weight corresponding to the moving path length score, W A Indicates the weight corresponding to the area score.

[0083] Among them, the weight W O , W L , W P and W A The value of is greater than 0. In practical applications, different values ​​can be set according to the contribution of the parameters corresponding to each weight to the evaluation of the sub-map scene information volume, such as: W O =0.5,W L =0.3,W P =0.1,W A =0.1.

[0084] 206. Determine a target submap from the multiple submaps as a reference map based on the scene information volume evaluation indicators corresponding to the multiple submaps, wherein the reference map is a starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene.

[0085] Specifically, after determining the scene information amount evaluation index corresponding to each submap according to formula (5), the submap corresponding to the scene information amount evaluation index with the largest value is determined from multiple submaps as the reference map.

[0086] In this embodiment, the amount of scene information corresponding to each submap is characterized based on four parameters: the number of common view feature points, the number of common view space points, and the length and area of ​​the moving path. This allows for determining an evaluation index for the amount of scene information corresponding to each submap. The submap containing the most scene information is selected from multiple submaps as the reference map, and multiple submaps are spliced ​​together to efficiently generate a map of the target scene.

[0087] In an optional embodiment, during the process of splicing other submaps in the multiple submaps with the base map, the base map may be updated to the currently spliced ​​map until no unspliced ​​submaps exist in the multiple submaps. For example, if the target scene is divided into three different areas, corresponding to three submaps, and submap M1 is determined to be the base map, if submap M2 is first spliced ​​with submap M1, then after the splicing of submaps M2 and M1 is completed to obtain map M12, map M12 can be used as the base map to splice submap M3 with map M12 to obtain map M123 of the target scene.

[0088] The following describes in detail one or more embodiments of the reference map acquisition device of the present invention. Those skilled in the art will appreciate that these devices can be constructed using commercially available hardware components and configured according to the steps taught in this solution.

[0089] Figure 4 A schematic diagram of a reference map acquisition device according to an embodiment of the present invention is shown in FIG. Figure 4 As shown, the device includes: an acquisition module 11 and a processing module 12.

[0090] The acquisition module 11 is configured to acquire multiple sub-maps generated for different areas in the target scene.

[0091] The processing module 12 is configured to determine, based on the map data corresponding to the multiple submaps, scene information volume evaluation indicators corresponding to the multiple submaps; and based on the scene information volume evaluation indicators, determine a target submap from the multiple submaps as a reference map, the reference map being the starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene.

[0092] Optionally, the map data includes: two-dimensional feature points corresponding to visual images collected for different areas of the target scene, three-dimensional spatial points corresponding to the two-dimensional feature points in the target scene, movement data when the visual images were collected, and size information of corresponding submaps, wherein the visual images are used to generate submaps. The processing module 12 is specifically configured to determine the number of common view feature points corresponding to the first submap based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to the second submap; the common view feature points are two-dimensional feature points contained in both the first submap and the second submap, the first submap being any one of the multiple submaps, and the second submap being any submap other than the first submap; the number of common view spatial points corresponding to the common view feature points corresponding to the first submap based on the three-dimensional spatial points corresponding to the common view feature points; the length of the movement path corresponding to the first submap based on the movement data corresponding to the first submap; the area corresponding to the first submap based on the size information corresponding to the first submap; and the scene information quantity evaluation index corresponding to each of the multiple submaps based on the number of common view feature points, the number of common view spatial points, the movement path length, and the area corresponding to each of the multiple submaps.

[0093] Optionally, there are multiple second submaps. The processing module 12 is further specifically configured to determine, based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to each second submap, the number of common view feature points between the first submap and each second submap; determine the sum of the number of common view feature points between the first submap and each second submap as the number of common view feature points corresponding to the first submap; determine the number of common view space points between the first submap and each second submap based on the three-dimensional spatial points corresponding to the common view feature points between the first submap and each second submap; and determine the sum of the number of common view space points between the first submap and each second submap as the number of common view space points corresponding to the first submap.

[0094] Optionally, the processing module 12 is further specifically configured to determine a score for the number of common view feature points of the first submap based on a ratio of the number of common view feature points corresponding to the first submap to the total number of common view feature points, where the total number of common view feature points is the sum of the numbers of common view feature points corresponding to the multiple submaps; determine a score for the number of common view space points of the first submap based on a ratio of the number of common view space points corresponding to the first submap to the total number of common view space points, where the total number of common view space points is the sum of the numbers of common view space points corresponding to the multiple submaps; and determine a score for the number of common view space points of the first submap based on a ratio of the number of common view feature points corresponding to the first submap to the total number of common view feature points. The moving path length score of the first submap is determined based on the ratio of the corresponding moving path length to the total moving path length, where the total moving path length is the sum of the moving path lengths corresponding to the multiple submaps respectively. The area score of the first submap is determined based on the ratio of the area corresponding to the first submap to the total area, where the total area is the sum of the areas corresponding to the multiple submaps respectively. The scene information quantity evaluation index corresponding to the first submap is determined based on the number score of common view feature points, the number score of common view space points, the moving path length score, and the area score of the first submap.

[0095] Optionally, the processing module 12 is further specifically configured to determine a first product of the number of common view feature points of the first sub-map and a first set weight, a second product of the number of common view space points and a second set weight, a third product of the moving path length score and a third set weight, and a fourth product of the area score and a fourth set weight; and use the sum of the first product, the second product, the third product, and the fourth product as an evaluation index of the scene information quantity of the first sub-map.

[0096] Optionally, the processing module 12 is further configured to determine, based on the scene information volume evaluation index, a sub-map corresponding to a scene information volume evaluation index having a maximum value from among the multiple sub-maps as a reference map.

[0097] Optionally, the processing module 12 is further configured to update the reference map to the currently spliced ​​map during the process of splicing other submaps in the plurality of submaps with the reference map, until no unspliced ​​submap exists in the plurality of submaps.

[0098] Figure 4 The device shown can execute the steps in the aforementioned embodiments. For detailed execution process and technical effects, please refer to the description in the aforementioned embodiments and will not be repeated here.

[0099] In one possible design, the above Figure 5 The structure of the reference map acquisition device shown can be implemented as an electronic device. Figure 5As shown, the electronic device may include: a memory 21, a processor 22, and a communication interface 23. The memory 21 stores executable code, and when the executable code is executed by the processor 22, the processor 22 can at least implement the reference map acquisition method provided in the above embodiments.

[0100] In addition, an embodiment of the present invention provides a non-transitory machine-readable storage medium, on which executable code is stored. When the executable code is executed by a processor of an electronic device, the processor can at least implement the baseline map acquisition method provided in the aforementioned embodiment.

[0101] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Those skilled in the art can understand and implement the present invention without inventive effort.

[0102] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by adding a necessary general hardware platform, and of course can also be implemented by a combination of hardware and software. Based on this understanding, the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a computer product. The present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for obtaining a reference map, characterized in that: The method comprises: Obtain multiple submaps generated for different areas in the target scene; Determining scene information quantity evaluation indicators corresponding to the plurality of submaps respectively based on the map data corresponding to the plurality of submaps respectively; determining a target submap from the plurality of submaps as a reference map according to the scene information evaluation index, the reference map being a starting submap when the plurality of submaps are spliced ​​together to generate a map of the target scene; The map data includes: two-dimensional feature points corresponding to visual images collected for different areas of the target scene, three-dimensional spatial points corresponding to the two-dimensional feature points in the target scene, movement data when the visual images are collected, and size information of corresponding submaps, wherein the visual images are used to generate the submaps; The determining, based on the map data corresponding to the plurality of sub-maps, scene information quantity evaluation indicators corresponding to the plurality of sub-maps respectively, includes: determining, based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to the second submap, a number of common-view feature points corresponding to the first submap; the common-view feature points being two-dimensional feature points contained in both the first submap and the second submap, the first submap being any one of the plurality of submaps, and the second submap being any submap of the plurality of submaps other than the first submap; Determining the number of common view space points corresponding to the first submap based on the three-dimensional space points corresponding to the common view feature points corresponding to the first submap; determining a length of a movement path corresponding to the first sub-map according to the movement data corresponding to the first sub-map; determining an area corresponding to the first submap according to the size information corresponding to the first submap; Determining a score for the number of common view feature points of the first submap based on a ratio of the number of common view feature points corresponding to the first submap to the total number of common view feature points, where the total number of common view feature points is the sum of the numbers of common view feature points corresponding to the multiple submaps; Determining a score for the number of common view space points of the first submap based on a ratio of the number of common view space points corresponding to the first submap to the total number of common view space points, where the total number of common view space points is the sum of the numbers of common view space points corresponding to the multiple submaps; Determining a moving path length score for the first submap based on a ratio of a moving path length corresponding to the first submap to a total moving path length, where the total moving path length is the sum of the moving path lengths corresponding to the multiple submaps; determining an area score for the first submap based on a ratio of an area corresponding to the first submap to a total area, where the total area is the sum of the areas corresponding to the plurality of submaps; A scene information quantity evaluation index corresponding to the first submap is determined according to the common view feature point quantity score, the common view space point quantity score, the moving path length score, and the area score of the first submap.

2. The method according to claim 1, characterized in that There are multiple second submaps, and determining the number of common view feature points corresponding to the first submap based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to the second submap includes: determining the number of common-view feature points between the first submap and each second submap based on the two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to each second submap; Determine the sum of the numbers of common view feature points between the first submap and each second submap as the number of common view feature points corresponding to the first submap; The determining the number of common view space points corresponding to the first submap according to the three-dimensional space points corresponding to the common view feature points corresponding to the first submap includes: determining the number of common view space points between the first submap and each second submap based on three-dimensional space points corresponding to common view feature points between the first submap and each second submap; The sum of the numbers of common view space points between the first submap and each second submap is determined to be the number of common view space points corresponding to the first submap.

3. The method according to claim 1, characterized in that The determining of the scene information quantity evaluation index corresponding to the first submap according to the common view feature point quantity score, the common view space point quantity score, the moving path length score, and the area score of the first submap includes: Determining a first product of the common view feature point quantity score and a first set weight, a second product of the common view space point quantity score and a second set weight, a third product of the movement path length score and a third set weight, and a fourth product of the area score and a fourth set weight for the first submap; The sum of the first product, the second product, the third product and the fourth product is used as an evaluation index of the scene information amount of the first sub-map.

4. The method according to claim 3, characterized in that The step of determining a target submap as a reference map from the plurality of submaps according to the scene information amount evaluation index includes: According to the scene information volume evaluation index, a submap corresponding to a scene information volume evaluation index having a maximum value is determined from the multiple submaps as a reference map.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: In the process of splicing other submaps among the plurality of submaps with the reference map, the reference map is updated to the currently spliced ​​map until no unspliced ​​submap exists among the plurality of submaps.

6. A reference map acquisition device, characterized in that: The device comprises: An acquisition module, used to acquire multiple sub-maps generated for different areas in the target scene; a processing module, configured to determine a number of common-view feature points corresponding to the first submap based on two-dimensional feature points corresponding to the first submap and the two-dimensional feature points corresponding to the second submap; the common-view feature points being two-dimensional feature points contained in both the first submap and the second submap, the first submap being any one of the plurality of submaps, and the second submap being any submap of the plurality of submaps other than the first submap; Determining the number of common view space points corresponding to the first submap based on the three-dimensional space points corresponding to the common view feature points corresponding to the first submap; determining a length of a movement path corresponding to the first submap based on movement data corresponding to the first submap; determining an area corresponding to the first submap according to the size information corresponding to the first submap; Determining a score for the number of common view feature points of the first submap based on a ratio of the number of common view feature points corresponding to the first submap to the total number of common view feature points, where the total number of common view feature points is the sum of the numbers of common view feature points corresponding to the multiple submaps; Determining a score for the number of common view space points of the first submap based on a ratio of the number of common view space points corresponding to the first submap to the total number of common view space points, where the total number of common view space points is the sum of the numbers of common view space points corresponding to the multiple submaps; Determining a moving path length score for the first submap based on a ratio of a moving path length corresponding to the first submap to a total moving path length, where the total moving path length is the sum of the moving path lengths corresponding to the multiple submaps; determining an area score for the first submap based on a ratio of an area corresponding to the first submap to a total area, where the total area is the sum of the areas corresponding to the plurality of submaps; A scene information quantity evaluation index corresponding to the first submap is determined based on the score of the number of common view feature points, the score of the number of common view spatial points, the score of the length of the movement path, and the score of the area of ​​the first submap; the map data includes: two-dimensional feature points corresponding to visual images collected for different areas of the target scene, three-dimensional spatial points corresponding to the two-dimensional feature points in the target scene, movement data when the visual images are collected, and size information of the corresponding submap, wherein the visual images are used to generate the submap; based on the scene information quantity evaluation index, a target submap is determined from the multiple submaps as a baseline map, and the baseline map is the starting submap when the multiple submaps are spliced ​​together to generate a map of the target scene.

7. An electronic device, characterized in that: include: A memory, a processor, and a communication interface; wherein the memory stores executable code, and when the executable code is executed by the processor, the processor executes the reference map acquisition method according to any one of claims 1 to 5.

8. A non-transitory machine-readable storage medium, characterized in that The non-transitory machine-readable storage medium stores executable code, and when the executable code is executed by a processor of an electronic device, the processor is caused to execute the reference map acquisition method according to any one of claims 1 to 5.

Images