Method and device for evaluating the accuracy of point cloud maps

Abstract

The present invention discloses a method and device for evaluating the accuracy of a point cloud map, and relates to the field of computer technology. A specific implementation of the method includes: S101, selecting a feature object on a point cloud map; S102, obtaining the coordinates of the feature points of the feature object in the real world to generate a feature point coordinate map; S103, selecting at least two real feature points on the feature point coordinate map, and determining the point cloud feature points corresponding to the at least two real feature points on the point cloud map; S104, calculating the real distance between the real feature points selected on the feature point coordinate map, and calculating the point cloud distance between the point cloud feature points selected on the point cloud map; S105, evaluating the accuracy of the point cloud map based on the distance difference between the real distance and the point cloud distance. This implementation reduces the manpower and time required to measure feature edges, and improves the reliability of the accuracy of the point cloud map.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for evaluating the accuracy of a point cloud map. Background Art

[0002] In recent years, with the development of technologies such as robots, drones, and autonomous driving, SLAM technology has also received more and more attention. SLAM (Simultaneous Localization and Mapping), that is, simultaneous positioning and map construction, refers to the process of fusing observation data on the basis of positioning. The result of SLAM data fusion is a point cloud map with a free space rectangular coordinate system. Unlike traditional point cloud maps constructed using GPS, inertial navigation systems, etc., SLAM point cloud maps are unconstrained and transformable, so they cannot be used like traditional point cloud maps. Figure 1 In this way, the accuracy of the point cloud map is evaluated directly by the coordinates of a single feature point on the point cloud map.

[0003] Currently, the commonly used method for evaluating the accuracy of SLAM point cloud maps is the direct length comparison method: directly measure the length of the characteristic edges of feature objects such as flower bed edges, speed bumps, ground markings, tree pits, etc. in the real world through steel rulers, meter rulers, laser rangefinders, etc., that is, the length of the geometric edges of the feature objects that can be identified by the naked eye; then compare the measured length of the characteristic edge with the length of the corresponding characteristic edge on the SLAM point cloud map and calculate the difference, thereby deriving the accuracy of the point cloud map.

[0004] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art: when measuring the characteristic edge lines of characteristic objects outdoors, at least two people are required at the same time, and only one sample can be obtained for each measurement. Therefore, when a large number of samples are needed, a lot of time and manpower are required; since the lengths of characteristic objects such as flower beds, tree pits, ground markings, etc. are all within 20 meters, the length of the measured samples is also within 20 meters, and it is impossible to evaluate the accuracy of the SLAM point cloud map of a larger scene; in addition, the multiple characteristic edge lines of the directly measured characteristic objects are independent of each other and lack correlation. Therefore, when the point cloud map is distorted, deformed, etc., the accuracy of the point cloud map cannot be well evaluated. Summary of the invention

[0005] In view of this, the present invention provides a method and device for evaluating the accuracy of point cloud maps, which can not only efficiently obtain a large number of samples for evaluating the accuracy of point cloud maps, but also evaluate point cloud maps of scenes of various scales, and can truly reflect the accuracy of SLAM point cloud maps when the point cloud maps are distorted or deformed.

[0006] To achieve the above-mentioned purpose, according to the first aspect of the present invention, a method for evaluating the accuracy of a point cloud map is provided. The method comprises: selecting a feature object on a point cloud map; obtaining the coordinates of the feature points of the feature object in the real world to generate a feature point coordinate map; selecting at least two real feature points on the feature point coordinate map, and determining the point cloud feature points corresponding to the at least two real feature points on the point cloud map; calculating the real distance between the real feature points selected on the feature point coordinate map, and calculating the point cloud distance between the point cloud feature points selected on the point cloud map to determine the distance difference between the real distance and the point cloud distance; evaluating the accuracy of the point cloud map according to the distance difference between the real distance and the point cloud distance.

[0007] Optionally, the coordinates of the feature points of the feature objects in the real world are obtained by real-time dynamic difference method to form a feature point coordinate map;

[0008] Optionally, it also includes: selecting a first point cloud feature point and a second point cloud feature point on the point cloud map, and determining a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map; translating the point cloud map so that the first point cloud feature point coincides with the first real feature point; rotating the point cloud map with the first point cloud feature point as a reference point so that the second point cloud feature point is close to the second real feature point; and determining other point cloud feature points corresponding to other real feature points on the point cloud map based on the principle of closest distance.

[0009] Optionally, the first point cloud feature point and the second point cloud feature point are the two feature points on the point cloud map that are the farthest apart; when there are multiple pairs of point cloud feature points with the farthest distance, or when there are multiple feature lines with the point cloud feature points as endpoints, the endpoints of the feature lines closest to the middle position of the point cloud map are selected as the first point cloud feature point and the second point cloud feature point.

[0010] Optionally, the number, position and spacing of the at least two feature points selected on the feature point coordinate map are determined according to the characteristics of the selected feature features; the characteristics of the feature features include: the number of feature features, the shape of feature features and the size of feature features.

[0011] Optionally, the accuracy of the point cloud map is evaluated based on the variance, standard deviation or maximum value of the distance difference between the real distance and the point cloud distance.

[0012] To achieve the above-mentioned purpose, according to a second aspect of the present invention, there is provided a device for evaluating the accuracy of a point cloud map, comprising: a feature selection module, a coordinate map generation module, a feature point selection module, and a distance calculation module; wherein the feature selection module is used to select feature features on a point cloud map; the coordinate map generation module is used to obtain the coordinates of feature points of the feature features in the real world to generate a feature point coordinate map; the feature point selection module is used to select at least two real feature points on the feature point coordinate map, and determine point cloud feature points corresponding to the at least two real feature points on the point cloud map; the distance calculation module is used to calculate the real distance between the real feature points selected on the feature point coordinate map, and calculate the point cloud distance between the point cloud feature points selected on the point cloud map, so as to determine the distance difference between the real distance and the point cloud distance; and the accuracy of the point cloud map is evaluated according to the distance difference between the real distance and the point cloud distance.

[0013] Optionally, the feature point selection module is also used to: select a first point cloud feature point and a second point cloud feature point on the point cloud map, and determine a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map; translate the point cloud map so that the first point cloud feature point coincides with the first real feature point; rotate the point cloud map with the first point cloud feature point as a reference point so that the second point cloud feature point is close to the second real feature point; and determine other point cloud feature points corresponding to other real feature points on the point cloud map based on the principle of closest distance.

[0014] To achieve the above-mentioned purpose, according to the third aspect of the present invention, a server for evaluating the accuracy of point cloud maps is provided, comprising: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement any of the methods for evaluating the accuracy of point cloud maps as described above.

[0015] In order to achieve the above-mentioned purpose, according to a fourth aspect of the present invention, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, any of the methods for evaluating the accuracy of a point cloud map as described above is implemented.

[0016] The above invention has the following advantages or beneficial effects: since the distance between feature points on the feature point coordinate diagram is used instead of the length of the feature edge line actually measured in the prior art for comparison with the distance between the corresponding feature points on the point cloud map, the technical problem that a large amount of manpower and material resources are required to obtain a large number of feature edge line lengths and the obtained feature edge line lengths are within 20 meters in the prior art is overcome, thereby greatly improving the efficiency of evaluating the accuracy of the point cloud map, and also expanding the scale of the scene of the point cloud map that can be evaluated; in addition, since the calculated distance between feature points includes not only the length corresponding to the feature edge line of the feature object, but also the length of the non-feature edge line or the length of the virtual feature edge line, when the point cloud map is distorted, deformed, etc., the length of the virtual feature edge line can well reflect the relative error of the point cloud map caused by distortion, deformation, etc., thereby improving the accuracy of the evaluated point cloud map to a certain extent.

[0017] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0019] Figure 1 is a schematic diagram of the main process of a method for evaluating the accuracy of a point cloud map according to an embodiment of the present invention;

[0020] Figure 2a is a schematic diagram of characteristic edges of characteristic features according to an embodiment of the present invention;

[0021] Figure 2b is a schematic diagram of feature points according to an embodiment of the present invention;

[0022] Figure 2c Schematic diagram of characteristic edge lines composed of selected point cloud characteristic points according to an embodiment of the present invention

[0023] Figure 2d Schematic diagram of virtual feature edge lines composed of selected point cloud feature points according to an embodiment of the present invention

[0024] Figure 2e is a schematic diagram of a distorted point cloud map according to an embodiment of the present invention;

[0025] Figure 3 is a schematic diagram of the main process of a method for coarse registration of a point cloud map and a feature point coordinate map according to an embodiment of the present invention;

[0026] Figure 4a Schematic diagram of a point cloud map and a feature point coordinate map according to an embodiment of the present invention

[0027] Figure 4b Schematic diagram of a translated point cloud map and a feature point coordinate map according to an embodiment of the present invention

[0028] Figure 4c Schematic diagram of a rotated point cloud map and a feature point coordinate map according to an embodiment of the present invention

[0029] Figure 4d Schematic diagram of another rotated point cloud map and feature point coordinate map according to an embodiment of the present invention

[0030] Figure 5 is a schematic diagram of main modules of a device for evaluating the accuracy of a point cloud map according to an embodiment of the present invention;

[0031] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0032] Figure 7 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0034] See also Figure 1 , an embodiment of the present invention provides a method for evaluating the accuracy of a point cloud map, such as Figure 1 As shown, the specific steps may include the following:

[0035] Step S101, selecting feature objects on the point cloud map.

[0036] There are various features in the real world, such as roads, buildings, blocks, bridges, etc. Point cloud maps can be used to reflect the real world, but different features are presented differently in point cloud maps: some features are fuzzy, with unclear boundaries or with clear boundaries but unable to determine feature points, such as round manhole covers; some features are very clear, with obvious feature points, easy to identify, and are therefore called feature features, such as speed bumps, ground markings, tree pits, flower beds, etc. Therefore, in order to facilitate subsequent identification and measurement, feature features with obvious feature points are selected on the point cloud map, and then used to obtain the corresponding feature features in the real world.

[0037] Step S102, obtaining the coordinates of the feature points of the feature objects in the real world to generate a feature point coordinate map.

[0038] Feature edges refer to the geometric shape edges of characteristic features such as speed bumps, ground markings, tree pits, and flower beds that can be recognized by the naked eye. Figure 2a The line segments marked 1, 2, 3, 4, and 5 shown in the figure are the characteristic edge lines, among which line segment 4 is the characteristic edge line of the flower bed, and the length of the line segment is the length of the corresponding characteristic edge line.

[0039] In order to evaluate the accuracy of the point cloud map, it is necessary to obtain the length of the characteristic edge line of the characteristic feature in the point cloud map or the distance between the characteristic points, and compare it with the real length of the characteristic edge line of the corresponding characteristic feature in the real world or the distance between the characteristic points. Therefore, it is necessary to locate or measure the real position or relative position of the endpoint or characteristic point of the characteristic feature in the real world to obtain the coordinates of the characteristic point, and then form a characteristic point coordinate map, so that the characteristic point or characteristic feature in the characteristic point coordinate map has a corresponding characteristic point or characteristic feature on the point cloud map. The so-called correspondence refers to the different expressions of the same characteristic feature or the same feature point of the same characteristic feature on the point cloud map and the feature point coordinate map. For example, the top of the same arrow is used as a feature point. The feature point is marked as point cloud feature point A in the point cloud map, and marked as real feature point A' in the feature point coordinate map. Then the point cloud feature point A corresponds to the real feature point A'.

[0040] In a possible implementation, the coordinates of the feature points of the feature features in the real world are obtained by real-time kinematic difference method to form a feature point coordinate map. Real-time kinematic difference method (RTK) is a technology for real-time dynamic positioning based on carrier phase observation values. It is a commonly used GPS measurement method and can obtain centimeter-level positioning accuracy in real time outdoors. The working mode of the real-time dynamic difference method is to set up a reference station at a known point, and the receiver sends its observation value and coordinate information to the mobile station receiver with the help of a radio. The mobile station receiver receives data from the reference station through a radio or a data link, and also collects GPS observation data, forms a carrier phase differential observation equation in the system, and uses Kalman filtering technology to obtain the three-dimensional coordinates of the point to be measured through real-time processing, and the accuracy can reach the centimeter level. It can be understood that the three-dimensional coordinates with the measuring point obtained by the real-time dynamic difference method in this embodiment are the absolutely true geographic coordinates of the point to be measured, that is, they contain longitude data, dimension data, etc.

[0041] Step S103, selecting at least two real feature points on the feature point coordinate map, and determining point cloud feature points corresponding to the at least two real feature points on the point cloud map.

[0042] In the real world or on the point cloud map that reflects the real world, each area has one or more characteristic objects, and each characteristic object has multiple measurable points. However, due to the limitation of time and energy, it is not appropriate to measure all the characteristic points of all characteristic objects. Therefore, it is necessary to formulate certain measurement rules in a targeted manner, including the number of characteristic points selected for measurement, the interval between characteristic points, the position of characteristic points, etc.

[0043] In a possible implementation, the number, position, and spacing of the at least two feature points selected on the feature point coordinate map are determined based on the characteristics of the selected feature features; the characteristics of the feature features include: the number of feature features, the shape of the feature features, and the size of the feature features.

[0044] Different feature points can be selected differently for different feature objects. For example, for arrow-shaped ground markings, you can select the feature point at the top of the arrow for measurement; for speed bumps, you can select the feature points at both ends of the speed bump for measurement; for crosswalks, you can select the feature points at the end points of the outermost markings of the crosswalk for measurement. For details, see Figure 2b , feature points A, B1, B2, C1, and C2 in the feature point coordinate map can be selected for measurement. On the other hand, when the number of feature objects is large or the distribution is dense, several or more representative feature points of feature objects can be selected for measurement. For example, when there are many tree pits of feature objects, a feature object measurement density can be set, that is, the distance interval between the measured feature objects. Taking a value of 30 meters as an example, only tree pits with a distance interval greater than or equal to 30 meters are selected for measurement, that is, one tree pit is measured every 30 meters on average, and tree pits within a range of less than 30 meters between two tree pits are ignored. In addition, a feature point measurement density can be set according to the size of the feature object or the number of feature points, that is, selective measurement of feature points. For example, taking a feature point measurement density of 20 meters as an example, one feature point is selected every 20 meters on average for measurement, and feature points within a range of 20 meters between two feature points are ignored. At the same time, the number of feature points to be measured can also be determined according to actual needs, such as selecting 20 or 30 per kilometer.

[0045] Step S104, calculating the real distance between the real feature points selected on the feature point coordinate map, and calculating the point cloud distance between the point cloud feature points selected on the point cloud map, so as to determine the distance difference between the real distance and the point cloud distance.

[0046] For the selected real feature points, the real distance between any two real feature points is calculated; similarly, for the selected point cloud feature points, the point cloud distance between any two point cloud feature points is calculated; then, based on the corresponding relationship between the feature points, the distance difference between the point cloud distance and the real distance is calculated. It can be understood that when the number of selected point cloud feature points or real feature points is large, the distance between feature points can be selectively calculated according to the actual situation. Specifically, see Figure 2c and Figure 2d , when the real feature points selected on the feature point coordinate map are marked as 1', 2', 3', 4', 5', 6', 7', 8', 9', 10' respectively, the corresponding point cloud feature points selected on the point cloud map are marked as 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 respectively (such as Figure 2c or 2d); the real distance between any two real feature points, such as 1'2', 1'3', 1'5', 2'3', etc., can be calculated, and the real distance between some real feature points can also be selectively calculated. In this embodiment, the calculation of the real distances 1'2', 3'4', 5'6', 7'8', 9'10' and 1'3', 1'8', 1'9', 2'8', 2'9' is used as an example for explanation; the point cloud feature points are calculated accordingly. The point cloud distances between the landmarks are 12, 34, 56, 78, 910 and 13, 18, 19, 28, and 29; the distance difference between the true distance 1'2' and the point cloud distance 12 is recorded as d12, the distance difference between the true distance 3'4' and the point cloud distance 34 is recorded as d34, and so on until all the distance differences d12, d34, d56, d78, d910, d13, d18, d19, d28, and d29 are obtained.

[0047] Step S105: evaluating the accuracy of the point cloud map according to the distance difference between the real distance and the point cloud distance.

[0048] Based on the above-mentioned difference between the real distance and the point cloud distance, we can judge whether the accuracy of the point cloud map meets the requirements according to the actual needs of the point cloud map application, such as judging the accuracy of the point cloud map according to whether the maximum value, average value, variance, standard deviation, etc. in the distance difference exceeds a certain accuracy threshold. The standard deviation and average value can be calculated according to the following formula:

[0049]

[0050]

[0051] Where σ is the standard deviation, d i is the ith difference, and d0 is the average difference.

[0052] Specifically, see Figure 2c and Figure 2d , still taking the distance differences obtained above as d12, d34, d56, d78, d910, d13, d18, d19, d28, and d29 as an example for explanation: if the evaluation is based on the maximum value in the distance difference, the maximum value among d12, d34, d56, d78, d910, d13, d18, d19, d28, and d29 is selected by comparison and screening, and whether the accuracy of the point cloud map meets the requirements is determined based on whether the maximum value reaches the accuracy threshold. If the maximum value in the distance difference is greater than or equal to the accuracy threshold, it is determined that the accuracy of the point cloud map does not meet the requirements and cannot be used; if the maximum value in the distance difference is lower than the accuracy threshold, it is determined that the accuracy of the point cloud map meets the requirements and can be used.

[0053] It is worth noting that see Figure 2d , Figure 2e , when evaluating the accuracy of the point cloud map, the above embodiment not only calculates the real distances 1'2', 3'4', 5'6', 7'8', 9'10' representing the length of the real feature edge line of the feature object, but also calculates the length of the feature edge line of the non-feature object or the length of the virtual feature edge line 1'3', 1'8', 1'9', 2'8', 2'9', and also calculates the corresponding point cloud distances 12, 34, 56, 78, 910 and 13, 18, 19, 28, 29 in the point cloud map. Therefore, even if the point cloud map has such Figure 2e When the distortion and deformation are shown, the combination of the length of the real feature edge line and the length of the virtual feature edge line can also well reflect the accuracy of the point cloud map and has higher reliability.

[0054] See also Figure 3 On the basis of the above embodiment, with respect to step S103 in the above embodiment, the embodiment of the present invention further provides a method for determining point cloud feature points corresponding to the at least two real feature points on the point cloud map based on rough registration of the point cloud map and the feature point coordinate map, which may specifically include the following steps:

[0055] Step S1031, selecting a first point cloud feature point and a second point cloud feature point on the point cloud map, and determining a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map.

[0056] The first point cloud feature point and the second point cloud feature point selected here are mainly used to achieve the rough registration of the point cloud map and the feature point coordinate map, so as to facilitate the selection of the corresponding point cloud feature points and the real feature points. Therefore, it is possible to select a pair of characteristic or easily identifiable feature points on the point cloud map as the first point cloud feature point and the second point cloud feature point by manual selection or image recognition, such as the feature points at the outermost part of the point cloud map and the feature points at the center of the point cloud map.

[0057] In a possible implementation, the first point cloud feature point and the second point cloud feature point are the two feature points with the farthest distance on the point cloud map; when there are multiple pairs of point cloud feature points with the farthest distance, or when there are multiple feature lines with the point cloud feature points as endpoints, the endpoints of the feature line closest to the middle position of the point cloud map are selected as the first point cloud feature point and the second point cloud feature point. Figure 4a , select the two feature points A and B with the farthest distance on the point cloud map as the first point cloud feature point and the second point cloud feature point respectively; then the real feature points a and b corresponding to the first point cloud feature point A and the second point cloud feature point B on the feature point coordinate map are the first real feature point a and the second real feature point b.

[0058] Step S1032: translate the point cloud map so that the first point cloud feature point coincides with the first real feature point.

[0059] Specifically, see Figure 4a and Figure 4b , taking the first real feature point a as reference, along Figure 4a The point cloud map is translated in the direction of the arrow shown in , until the first point cloud feature point A coincides with the first real feature point a. At this time, the positional relationship between the point cloud map and the feature point coordinate map is as follows: Figure 4b shown.

[0060] Step S1033: Taking the first point cloud feature point as a reference point, the point cloud map is rotated to bring the second point cloud feature point closer to the second real feature point.

[0061] It is understandable that, since there is a certain error between the constructed point cloud map and the real world, when the first point cloud feature point A and the first real feature point a have already overlapped, and the first point cloud feature point A is used as the reference point, and the point cloud map is rotated so that the second point cloud feature point B is close to the second real feature point b, there may be a situation where the second point cloud feature point B and the second real feature point b overlap (such as Figure 4c As shown), there may also be a situation where the second point cloud feature point B and the second real feature point b cannot coincide with each other (as shown in Figure 4dAs shown in the figure, if the second point cloud feature point B and the second real feature point b cannot overlap and the distance between them is large, it means that the point cloud map and the feature point coordinate map are very different or the accuracy of the point cloud map is low, and you should consider giving up the current point cloud map. Therefore, based on the proximity between the second point cloud feature point B and the second real feature point b, that is, whether the minimum distance between the second point cloud feature point B and the second real feature point b exceeds the threshold distance, you can pre-evaluate the accuracy of the point cloud map and further consider whether to continue to evaluate the accuracy of the current point cloud map or give up the current point cloud map.

[0062] Step S1034, based on the principle of closest distance, determining other point cloud feature points corresponding to other real feature points on the point cloud map.

[0063] Specifically, see Figure 4c and Figure 4d When the second point cloud feature point B completely coincides with the second real feature point b or the minimum distance between them is less than the threshold distance, based on the principle of closest distance, the point cloud feature point corresponding to the real feature point is selected on the point cloud map to measure the point cloud distance and the real distance.

[0064] It is understandable that although the point cloud map is three-dimensional (using the coordinate axes X, Y, and Z as an example), when constructing the point cloud map using methods such as SLAM, the Z axis is vertically upward by default, so the point cloud map obtained is parallel to the real world or the feature point coordinate map in the height direction. Therefore, in the process of manipulating the point cloud map to approach or roughly align with the feature point coordinate map, there is no need to consider the pitch and flip problems, and only rotation and translation are required in the plane formed by X and Y. At the same time, since the height of the point cloud is controlled within 10-20 meters, in such a small range, the accuracy in the elevation direction is generally within 10cm and will not accumulate. At the same time, since the tolerance for relative elevation accuracy is large during the use of the point cloud map, the relative elevation accuracy fully meets the usual application requirements. There is no need to consider the elevation factor when evaluating the accuracy, and only the relative plane accuracy is concerned. Therefore, it is only necessary to rotate and translate in the plane formed by X and Y to evaluate the relative accuracy in the X and Y directions.

[0065] See also Figure 5 The embodiment of the present invention provides a device 500 for evaluating the accuracy of a point cloud map, comprising: a feature selection module 501, a coordinate map generation module 502, a feature point selection module 503, and a distance calculation module 504; wherein,

[0066] The feature selection module 501 is used to select feature features on the point cloud map;

[0067] The coordinate map generating module 502 is used to obtain the coordinates of the feature points of the feature objects in the real world to generate a feature point coordinate map;

[0068] The feature point selection module 503 is used to select at least two real feature points on the feature point coordinate map, and determine point cloud feature points corresponding to the at least two real feature points on the point cloud map;

[0069] The distance calculation module 504 is used to calculate the real distance between the real feature points selected on the feature point coordinate map, and calculate the point cloud distance between the point cloud feature points selected on the point cloud map, so as to determine the distance difference between the real distance and the point cloud distance; and evaluate the accuracy of the point cloud map based on the distance difference between the real distance and the point cloud distance.

[0070] In an optional embodiment, the feature point selection module 503 is also used to: select a first point cloud feature point and a second point cloud feature point on the point cloud map, and determine a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map; translate the point cloud map so that the first point cloud feature point coincides with the first real feature point; rotate the point cloud map with the first point cloud feature point as a reference point so that the second point cloud feature point is close to the second real feature point; and determine other point cloud feature points corresponding to other real feature points on the point cloud map based on the principle of closest distance.

[0071] In summary, the method or device for evaluating the accuracy of a point cloud map provided by an embodiment of the present invention uses the real distance between real feature points on the feature point coordinate diagram instead of the actual measured length of the feature edge line in the prior art for comparison with the point cloud distance between the corresponding point cloud feature points on the point cloud map, thereby overcoming the technical problem that a large amount of manpower and material resources are required to obtain a large number of feature edge line lengths and the acquired feature edge line lengths are within 20 meters in the prior art, thereby greatly improving the efficiency of evaluating the accuracy of the point cloud map, and also expanding the scale of scenes of the point cloud map that can be evaluated; in addition, since the calculated distance between feature points includes not only the length corresponding to the feature edge line of the feature object, but also the length of the non-feature edge line or the length of the virtual feature edge line, when the point cloud map is distorted, deformed, etc., the length of the virtual feature edge line can well reflect the relative error of the point cloud map caused by distortion, deformation, etc., thereby improving the accuracy of the evaluated point cloud map to a certain extent.

[0072] Figure 6 An exemplary system architecture 600 to which the method or apparatus for evaluating the accuracy of a point cloud map according to an embodiment of the present invention may be applied is shown.

[0073] like Figure 6As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0074] The user can use the terminal devices 601, 602, 603 to interact with the server 605 through the network 604 to receive or send messages, etc. Various communication client applications can be installed on the terminal devices 601, 602, 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0075] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0076] The server 605 may be a server that provides various services, such as a backend management server that provides support for shopping websites browsed by users using the terminal devices 601, 602, and 603. The backend management server may analyze and process the received product information query request and other data, and feed back the processing results such as the calculated real distance, point cloud distance, and distance difference to the terminal device.

[0077] It should be noted that the method for evaluating the accuracy of a point cloud map provided in the embodiment of the present invention is generally executed by the server 605 , and accordingly, the device for evaluating the accuracy of a point cloud map is generally arranged in the server 605 .

[0078] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0079] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 of a terminal device suitable for implementing an embodiment of the present invention. Figure 7 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0080] like Figure 7As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0081] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0082] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0083] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0084] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0085] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor includes a feature selection module, a coordinate map generation module, a feature point selection module, and a distance calculation module. The names of these modules do not constitute a limitation on the modules themselves under certain circumstances. For example, the feature selection module may also be described as a "module for selecting feature features".

[0086] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: selecting feature objects on a point cloud map; obtaining the coordinates of feature points of the feature objects in the real world to generate a feature point coordinate map; selecting at least two real feature points on the feature point coordinate map, and determining point cloud feature points corresponding to the at least two real feature points on the point cloud map; calculating the real distance between the real feature points selected on the feature point coordinate map, and calculating the point cloud distance between the point cloud feature points selected on the point cloud map to determine the distance difference between the real distance and the point cloud distance; evaluating the accuracy of the point cloud map according to the distance difference between the real distance and the point cloud distance.

[0087] According to the technical solution of the embodiment of the present invention, the real distance between the real feature points on the feature point coordinate diagram is used instead of the length of the feature edge line actually measured in the prior art to be compared with the point cloud distance between the corresponding point cloud feature points on the point cloud map, thereby overcoming the technical problems in the prior art that a large amount of manpower and material resources are required to obtain a large number of feature edge line lengths and the obtained feature edge line lengths are within 20 meters, thereby greatly improving the efficiency of evaluating the accuracy of the point cloud map, and also expanding the scale of the scene of the point cloud map that can be evaluated; in addition, since the calculated distance between the feature points includes not only the length corresponding to the feature edge line of the feature object, but also the length of the non-feature edge line or the length of the virtual feature edge line, when the point cloud map is distorted, deformed, etc., the length of the virtual feature edge line can well reflect the relative error of the point cloud map caused by distortion, deformation, etc., thereby improving the accuracy of the evaluated point cloud map to a certain extent.

[0088] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for evaluating the accuracy of a point cloud map, characterized in that: include: Select feature objects on the point cloud map; Obtaining coordinates of feature points of the feature objects in the real world to generate a feature point coordinate map; Selecting at least two real feature points on the feature point coordinate map, and determining point cloud feature points corresponding to the at least two real feature points on the point cloud map; Calculate the real distance between the real feature points selected on the feature point coordinate map, and calculate the point cloud distance between the point cloud feature points selected on the point cloud map, and determine the distance difference between the real distance and the point cloud distance according to the corresponding relationship between the feature points, wherein for the selected real feature points, respectively calculate the real distance between any two real feature points; and for the selected point cloud feature points, respectively calculate the point cloud distance between any two point cloud feature points, wherein the calculated distance between the feature points includes not only the length corresponding to the real feature edge line of the feature feature, but also the length of the non-feature edge line or the length of the virtual feature edge line, wherein the feature feature refers to a feature with clear boundaries and clear feature points presented in the point cloud map; The accuracy of the point cloud map is evaluated according to the distance difference between the real distance and the point cloud distance.

2. The method for evaluating the accuracy of a point cloud map according to claim 1, characterized in that: The coordinates of the feature points of the feature objects in the real world are obtained by real-time dynamic difference method to form a feature point coordinate map.

3. The method for evaluating the accuracy of a point cloud map according to claim 1, characterized in that: Also includes: Selecting a first point cloud feature point and a second point cloud feature point on the point cloud map, and determining a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map; Translating the point cloud map so that the first point cloud feature point coincides with the first real feature point; Taking the first point cloud feature point as a reference point, rotating the point cloud map so that the second point cloud feature point is close to the second real feature point; Based on the principle of closest distance, other point cloud feature points corresponding to other real feature points are determined on the point cloud map.

4. The method for evaluating the accuracy of a point cloud map according to claim 3, characterized in that: The first point cloud feature point and the second point cloud feature point are two feature points with the farthest distance on the point cloud map; When there are multiple pairs of point cloud feature points with the farthest distance, or when there are multiple feature lines with the point cloud feature points as endpoints, the endpoints of the feature lines closest to the middle position of the point cloud map are selected as the first point cloud feature point and the second point cloud feature point.

5. The method for evaluating the accuracy of a point cloud map according to claim 1, characterized in that: According to the characteristics of the selected feature features, the number, position and spacing of the at least two feature points selected on the feature point coordinate map are determined; the characteristics of the feature features include: the number of feature features, the shape of feature features and the size of feature features.

6. The method for evaluating the accuracy of a point cloud map according to claim 1, characterized in that: The accuracy of the point cloud map is evaluated according to the variance, standard deviation or maximum value of the distance difference between the real distance and the point cloud distance.

7. A device for evaluating the accuracy of a point cloud map, characterized in that: include: Feature selection module, coordinate map generation module, feature point selection module, distance calculation module; among them, The object selection module is used to select characteristic objects on the point cloud map; The coordinate map generation module is used to obtain the coordinates of the feature points of the feature objects in the real world to generate a feature point coordinate map; The feature point selection module is used to select at least two real feature points on the feature point coordinate map, and determine point cloud feature points corresponding to the at least two real feature points on the point cloud map; The distance calculation module is used to calculate the real distance between the real feature points selected on the feature point coordinate map, and calculate the point cloud distance between the point cloud feature points selected on the point cloud map, and determine the distance difference between the real distance and the point cloud distance according to the corresponding relationship between the feature points, wherein, for the selected real feature points, the real distance between any two real feature points is calculated respectively; and for the selected point cloud feature points, the point cloud distance between any two point cloud feature points is calculated respectively, wherein the calculated distance between the feature points includes not only the length corresponding to the real feature edge line of the feature feature, but also the length of the non-feature edge line or the length of the virtual feature edge line; and according to the distance difference between the real distance and the point cloud distance, the accuracy of the point cloud map is evaluated, wherein the feature feature refers to the feature with clear boundaries and clear feature points presented in the point cloud map.

8. The device for evaluating the accuracy of a point cloud map according to claim 7, characterized in that: The feature point selection module is also used for: Selecting a first point cloud feature point and a second point cloud feature point on the point cloud map, and determining a first real feature point and a second real feature point corresponding to the first point cloud feature point and the second point cloud feature point on the feature point coordinate map; Translating the point cloud map so that the first point cloud feature point coincides with the first real feature point; Taking the first point cloud feature point as a reference point, rotating the point cloud map so that the second point cloud feature point is close to the second real feature point; Based on the principle of closest distance, other point cloud feature points corresponding to other real feature points are determined on the point cloud map.

9. An electronic device for evaluating the accuracy of a point cloud map, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images