High-precision map original data storage optimization method and system

Abstract

The application discloses a high-precision map original data storage optimization method and system, which associates trajectory data, point cloud data and picture data based on timestamp information; sets a trajectory point interval threshold value, deletes a next trajectory point data with an interval distance less than the trajectory point interval threshold value from a previous trajectory point; deletes point cloud data corresponding to the deleted trajectory point time, sets an intensity threshold value and a point cloud density threshold value, deletes road surface point cloud data with an intensity lower than the intensity threshold value and road outside point cloud data with a point cloud density less than the point cloud density threshold value; deletes picture data corresponding to the deleted trajectory point time, sets a gray average value threshold value, and deletes picture data with a gray average value less than the gray average value threshold value; thereby reducing data storage hardware consumption, reducing data storage cost and network bandwidth pressure in a data scheduling transmission process, and improving data processing and browsing efficiency and interactive operation response speed in a high-precision map production process.

Description

TECHNICAL FIELD

[0001] The present application relates to high-precision map making technology, and in particular to a high-precision map original data storage optimization method and system. BACKGROUND

[0002] High-precision map making generally adopts a vehicle-mounted mobile measurement system to collect three-dimensional data of a road network, including point clouds, images, trajectories, etc. The obtained sensor original data is very huge, and a large amount of storage resources need to be consumed. In particular, in the process of high-precision map making, process data of many production links need to be saved, resulting in that a PB-level storage system needs to be supported for high-precision map making. Therefore, reducing data storage cost is an important optimization item in high-precision map making engineering. SUMMARY

[0003] The present application aims to overcome the above technical deficiencies, and provides a high-precision map original data storage optimization method and system, which solves the problem that the existing high-precision map original data is huge and a large amount of storage resources need to be consumed.

[0004] To achieve the above technical purposes, the technical scheme of the present application provides a high-precision map original data storage optimization method in a first aspect, which includes the following steps:

[0005] Obtaining three-dimensional real scene data, associating trajectory data, point cloud data and picture data based on timestamp information;

[0006] Setting a trajectory point spacing threshold, and deleting a next trajectory point data whose interval distance from a previous trajectory point is less than the trajectory point spacing threshold;

[0007] Deleting point cloud data corresponding to the deleted trajectory point time, and setting an intensity threshold and a point cloud density threshold, deleting road surface point cloud data whose intensity is lower than the intensity threshold and road outside point cloud data whose point cloud density is less than the point cloud density threshold;

[0008] Deleting picture data corresponding to the deleted trajectory point time, and setting a gray average value threshold, deleting picture data whose gray average value is less than the gray average value threshold.

[0009] The present application provides a high-precision map original data storage optimization system in a second aspect, which includes the following functional modules:

[0010] A data association module for obtaining three-dimensional real scene data, associating trajectory data, point cloud data and picture data based on timestamp information;

[0011] A trajectory point reduction module for setting a trajectory point spacing threshold, and deleting a next trajectory point data whose interval distance from a previous trajectory point is less than the trajectory point spacing threshold;

[0012] The point cloud pruning module is used for deleting point cloud data corresponding to the deleted track point time, setting an intensity threshold and a point cloud density threshold, deleting road surface point cloud data with an intensity lower than the intensity threshold, and deleting road outside point cloud data with a point cloud density lower than the point cloud density threshold.

[0013] The picture pruning module is used for deleting picture data corresponding to the deleted track point time, and setting a gray average value threshold to delete picture data with a gray average value lower than the gray average value threshold.

[0014] The third aspect of the present application provides a server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned high-precision map original data storage optimization method.

[0015] The fourth aspect of the present application provides a computer readable storage medium storing a computer program, wherein the computer program is executed by a processor to implement the above-mentioned high-precision map original data storage optimization method.

[0016] Compared with the prior art, the high-precision map original data storage optimization method and system of the present application associates track data, point cloud data and picture data based on timestamp information, sets a track point spacing threshold to delete subsequent track point data with an interval distance from a previous track point less than the track point spacing threshold, deletes point cloud data corresponding to the deleted track point time, sets an intensity threshold and a point cloud density threshold, deletes road surface point cloud data with an intensity lower than the intensity threshold, and deletes road outside point cloud data with a point cloud density lower than the point cloud density threshold, deletes picture data corresponding to the deleted track point time, and sets a gray average value threshold to delete picture data with a gray average value lower than the gray average value threshold, thereby reducing the total amount of track data, point cloud data and image data in the high-precision map data, further reducing the storage hardware consumption of the data, reducing the data storage cost in the high-precision map data production process and the network bandwidth pressure in the data scheduling and transmission process, and improving the data processing efficiency in the high-precision map production process and the data browsing efficiency and interactive operation response speed. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a flow chart of the high-precision map original data storage optimization method according to the embodiments of the present application;

[0018] Figure 2 is a module block diagram of the high-precision map original data storage optimization system according to the embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0020] As shown in Figure 1 The embodiment of the present application provides a high-precision map original data storage optimization method, which comprises the following steps:

[0021] S1, acquiring three-dimensional real scene data, associating trajectory data, point cloud data and picture data based on timestamp information.

[0022] Since the trajectory, picture and point cloud data all have timestamp information, the trajectory data, point cloud data and picture data can be associated based on the timestamp information, and specifically, the trajectory data, point cloud data and picture data at the same time point are associated and set.

[0023] S2, setting a trajectory point spacing threshold, and deleting a next trajectory point data with a distance from a previous trajectory point less than the trajectory point spacing threshold.

[0024] That is, the trajectory point spacing threshold is set, the first trajectory point is taken as a reference point, the second trajectory point on the same trajectory is sequentially set, the spacing between the first trajectory point and the second trajectory point is calculated, if the spacing between the first trajectory point and the second trajectory point is less than the trajectory point spacing threshold, the second trajectory point is determined as an invalid trajectory point, otherwise, the second trajectory point is determined as a valid trajectory point, and the invalid trajectory point is deleted.

[0025] Specifically, the trajectory coordinates are converted into WGS84 coordinate system, the spatial distance is analyzed based on the three-dimensional coordinates of the trajectory points, the spatial distance D between the adjacent two trajectory points (Pn, Pn+1) is calculated. When D is less than the set threshold L, Pn+1 is marked as an invalid trajectory point and is deleted; the distance D between the trajectory points (Pn, Pn+2) is recalculated, whether the trajectory point (Pn+2) is an invalid trajectory point is re-judged, until the distance D between the trajectory points (Pn, Pn+i) is greater than the set threshold L, the point (Pn+i) is determined as a valid trajectory point and is retained.

[0026] S3, deleting the point cloud data corresponding to the time of the deleted trajectory point, setting an intensity threshold and a point cloud density threshold, deleting the road surface point cloud data with an intensity lower than the intensity threshold and the road outside point cloud data with a point cloud density less than the point cloud density threshold.

[0027] That is, based on the time information of the trajectory point marked as invalid, the associated point cloud frame data is queried, the related discrete points in the point cloud frame are marked as invalid data and deleted, and the total amount of discrete data is reduced.

[0028] Based on the pruned point cloud data, a seed growing method is used to identify the road surface point cloud data. Specifically, the pruned point cloud data is taken as a seed to grow to both sides, the growing direction is based on the vertical direction of the front and rear trajectory points, and the growing condition is set as: the Z value difference in the plane formed by the generated points does not exceed 5 cm, and the range of each growth does not exceed 0.5 m; the road surface point cloud data is identified according to the growth condition.

[0029] Since the road surface data has the characteristic that the intensity of the printed body is higher than that of other artificial objects, based on this feature, a watershed method is used to segment the road surface point cloud data based on the intensity threshold, to separate the road surface printing from the road surface background, and delete the point cloud data of the road surface background. However, the real scene is complex, and there may be problems such as wear and tear and obstruction of the printed matter, resulting in low intensity, which leads to incomplete separation of the printed matter, so the point cloud of the road surface printing is taken as a seed to grow around, if the intensity difference between the point cloud of the previous growth and the next growth is less than the threshold span range, then continue to iterate 3 times outward, so as to ensure the integrity of the printed matter while filtering the data as much as possible. Preferably, the threshold span range is 5% of the maximum intensity span of the current point cloud data.

[0030] The main object of the production of high-precision map data is the road and the artificial facilities on both sides of the road, and the trees and vegetation on both sides of the road can be classified as invalid data, and there are basically trees and vegetation on both sides of the road, accounting for a large proportion of the overall data. The point cloud of trees and vegetation is characterized by dispersion, not forming a plane, and being on both sides of the road. Based on point cloud analysis, the area on both sides of the road is identified, the point cloud in the area is identified and extracted, and the discrete points are filtered to reduce the vegetation data and reduce the total amount of point cloud data.

[0031] That is, after processing the road surface point cloud data, further deleting the road outside point cloud data with a point cloud density less than a point cloud density threshold, including: widening the road to both sides by a predetermined distance range to obtain a road outside effective area, based on point cloud analysis, identifying the road outside effective area, and performing plane identification and discrete point filtering on the point cloud in the road outside effective area. Specifically, the following steps are included:

[0032] Based on the identified road surface point cloud data, a polygon outer frame of the road surface is found, and two edge lines are fitted in the driving direction;

[0033] The road line is extrapolated to both sides by a predetermined distance to form a road outside area, and the point cloud in the road outside area is collected as a point cloud data set of the road outside area;

[0034] Based on the neighboring point algorithm, the field point cloud density of each discrete point in the point cloud data set of the road outside area is analyzed, the point cloud data set with a point cloud density lower than a first predetermined density threshold is deleted, and isolated points are filtered;

[0035] For the remaining point cloud, project to the xoy plane, do field density analysis in the xoy plane, delete the point cloud data set whose point cloud density is lower than the second preset density threshold, and filter the isolated points;

[0036] For the remaining points, based on the plane recognition method provided by the PCL library, mark as non-planar data, since the road side plane data is basically a route guide sign, considering that the rod below the sign needs to be made, the points on the upper and lower edges of the sign can be used as seed points to grow downward to avoid filtering the rod point cloud.

[0037] The point cloud data has the characteristic that the closer the scanning distance, the greater the data density, especially on the road surface directly below the mobile measurement system, the scanning distance is less than 3 meters, the distance between adjacent points is less than 1 cm, and the density is much greater than the accuracy requirement of high-precision map making. Therefore, after deleting the point cloud data whose road surface intensity is lower than the intensity threshold and the point cloud data whose point cloud density is less than the point cloud density threshold, a point cloud data interval distance threshold is set, and one of the adjacent point clouds with an interval distance less than the point cloud data interval distance threshold is deleted.

[0038] Specifically, the distance threshold L of adjacent points can be set, taking the point cloud directly below the laser scanner as the reference, traversing the adjacent points to the left and right respectively, calculating the distance d1 of adjacent point clouds (Pt, Pt+1), when d<L, Pt+1 is marked as invalid point cloud and one of them is deleted, the distance d2 of point clouds (Pt, Pt+2) is calculated, until the distance of point clouds (Pt, Pt+i) is greater than L, the calculation is stopped, Pt+i is marked as valid point cloud and retained. According to the invalid marking of the point cloud, the discrete points meaningless for subsequent data making are removed, and the total amount of point cloud data is reduced.

[0039] S4, delete the picture data corresponding to the deleted trajectory point time, set a gray average value threshold, and delete the picture data whose gray average value is less than the gray average value threshold.

[0040] Based on the time information marked as invalid trajectory points, the associated image data is queried, the image data is marked as invalid data and deleted.

[0041] Since the mobile measurement system will pass through tunnels, underpasses and other sunlight shielding areas during the collection process, there are also cases where the collection direction is opposite to the sunlight direction, resulting in overexposure or overdarkness of the image data collected by the image sensor; Therefore, the collected GRB image is converted into a gray image, the gray average value and the average value deviation are calculated, whether the picture is overexposed or overdark is judged according to the set gray average value threshold, and the overexposed or overdark picture is deleted from the image data.

[0042] After deleting the picture data with the average gray value less than the average gray value threshold, in the effective area outside the road, the deep learning model is used to identify whether it is the sky and the road vegetation, and the pixels in the sky and the vegetation range are modified to black. That is, based on the road surface identified by the point cloud, the effective area is set to be 10 meters on both sides of the road, the effective area is projected to the image data, and the pixels in the effective area are marked as effective pixels. For the unmarked pixel set, the deep learning model is used to identify whether it is the sky and the road vegetation, and the pixels in the sky and the vegetation range are modified to black. The characteristics of the jpeg format are used to realize the compression of invalid data, so as to reduce the total amount of image data.

[0043] For filtering invalid trajectory data, point cloud data and image data, they are saved to the directory specified by the rule according to the use requirements of subsequent high-precision map making.

[0044] The present application associates the trajectory data, point cloud data and picture data based on the timestamp information; sets a trajectory point spacing threshold, deletes the next trajectory point data with the interval distance from the previous trajectory point less than the trajectory point spacing threshold; deletes the point cloud data corresponding to the deleted trajectory point time, and sets an intensity threshold and a point cloud density threshold, deletes the road surface point cloud data with the intensity lower than the intensity threshold and the road outside point cloud data with the point cloud density less than the point cloud density threshold; deletes the picture data corresponding to the deleted trajectory point time, and sets an average gray value threshold, deletes the picture data with the average gray value less than the average gray value threshold; thereby reducing the total amount of trajectory data, point cloud data and image data in the high-precision map data, further reducing the storage hardware consumption of data, reducing the data storage cost in the process of high-precision map data making and the network bandwidth pressure in the process of data scheduling and transmission, and improving the data processing efficiency, data browsing efficiency and interactive operation response speed in the process of high-precision map making.

[0045] As shown in Figure 2 The present application also discloses a high-precision map original data storage optimization system, which comprises the following functional modules:

[0046] The data association module 10 is used for acquiring three-dimensional real scene data, and associating the trajectory data, point cloud data and picture data based on the timestamp information;

[0047] The trajectory point deletion module 20 is used for setting a trajectory point spacing threshold, and deleting the next trajectory point data with the interval distance from the previous trajectory point less than the trajectory point spacing threshold;

[0048] The point cloud deletion module 30 is used for deleting the point cloud data corresponding to the deleted trajectory point time, and setting an intensity threshold and a point cloud density threshold, deleting the point cloud data with the road surface intensity lower than the intensity threshold and the point cloud data with the point cloud density less than the point cloud density threshold;

[0049] The picture pruning module 40 is configured to delete picture data corresponding to the deleted track point time, and set a gray average value threshold, and delete picture data with a gray average value less than the gray average value threshold.

[0050] The execution mode of the high-precision map original data storage optimization system in this embodiment is basically the same as the high-precision map original data storage optimization method described above, and thus will not be described in detail.

[0051] The server in this embodiment is a device for providing computing services, and generally refers to a computer with high computing power, which is provided to multiple consumers through a network. The server in this embodiment includes a memory, a processor, and a system bus. The memory includes a runnable program stored thereon, and those skilled in the art can understand that the structure of the terminal device in this embodiment does not constitute a limitation on the terminal device, and can include more or fewer components than the illustration, or combine certain components, or different component arrangements.

[0052] The memory can be used to store software programs and modules, and the processor can execute various functional applications and data processing of the terminal by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required by at least one function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the terminal (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0053] The memory contains a runnable program of a high-precision map original data storage optimization method, which can be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete information acquisition and implementation processes. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the server. For example, the computer program can be divided into a data association module 10, a track point pruning module 20, a point cloud pruning module 30, and a picture pruning module 40.

[0054] The processor is the control center of the server, connects each part of the entire terminal device by using various interfaces and lines, executes various functions of the terminal and processes data by running or executing software programs and / or modules stored in the memory and calling data stored in the memory, thereby overall monitoring the terminal. Optionally, the processor can include one or more processing units; preferably, the processor can integrate an application processor and a modem processor, wherein the application processor mainly processes operating systems, application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor.

[0055] The system bus is used to connect the internal functional components of the computer and can transmit data information, address information and control information, and its types can be, for example, a PCI bus, an ISA bus, a VESA bus, etc. The instructions of the processor are transmitted to the memory through the bus, and the memory feeds back data to the processor, and the system bus is responsible for the interaction of data and instructions between the processor and the memory. Of course, the system bus can also access other devices, such as a network interface, a display device, etc.

[0056] The server should at least include a CPU, a chipset, a memory, a disk system, etc., and other constituent components are not described here.

[0057] In the embodiment of the application, the executable program executed by the processor included in the terminal is specifically a high-precision map original data storage optimization method, which includes the following steps:

[0058] Obtain three-dimensional real scene data, and associate trajectory data, point cloud data and picture data based on timestamp information;

[0059] Set a trajectory point interval threshold, and delete a subsequent trajectory point data whose interval distance from a previous trajectory point is less than the trajectory point interval threshold;

[0060] Delete point cloud data corresponding to the deleted trajectory point time, set an intensity threshold and a point cloud density threshold, delete road surface point cloud data whose intensity is lower than the intensity threshold, and delete road outside point cloud data whose point cloud density is less than the point cloud density threshold;

[0061] Delete picture data corresponding to the deleted trajectory point time, and set a gray average value threshold, and delete picture data whose gray average value is less than the gray average value threshold.

[0062] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-mentioned system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0063] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in a certain embodiment can be referred to the relevant description of other embodiments.

[0064] Those skilled in the art can appreciate that the modules, units and / or method steps of each embodiment described in combination with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0065] The above-described embodiments are merely used to illustrate the technical solutions of the present application, but not to limit the present application; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced by equivalent replacements; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A high-precision map raw data storage optimization method, characterized in that, The method comprises the following steps: acquiring three-dimensional real scene data, associating trajectory data, point cloud data and picture data based on timestamp information; setting a trajectory point spacing threshold, and deleting a subsequent trajectory point data with an interval distance from a previous trajectory point less than the trajectory point spacing threshold; deleting point cloud data corresponding to the deleted trajectory point time, setting an intensity threshold and a point cloud density threshold, deleting road surface point cloud data with an intensity less than the intensity threshold, and deleting road outside point cloud data with a point cloud density less than the point cloud density threshold; deleting picture data corresponding to the deleted trajectory point time, and setting a gray average value threshold to delete picture data with a gray average value less than the gray average value threshold; wherein the road surface point cloud data with an intensity less than the intensity threshold is deleted by: based on the pruned point cloud data, seed growing method is used to identify the road surface point cloud data; based on the intensity threshold, the road surface background is separated from the road surface background by the watershed method, and the point cloud data of the road surface background is deleted.

2. The high-precision map raw data storage optimization method according to claim 1, characterized in that, The setting of the trajectory point spacing threshold and the deletion of the subsequent trajectory point data with the interval distance from the previous trajectory point less than the trajectory point spacing threshold comprises: setting the trajectory point spacing threshold, taking the first trajectory point as the reference point, sequentially taking the second trajectory point on the same trajectory, calculating the spacing between the first trajectory point and the second trajectory point, if the spacing between the second trajectory point and the first trajectory point is less than the trajectory point spacing threshold, the second trajectory point is determined as an invalid trajectory point, otherwise, it is an effective trajectory point, and the invalid trajectory point is deleted. 3.The high-precision map raw data storage optimization method according to claim 1, characterized in that, After the road surface printing is separated from the road surface background by the watershed method, the separated road surface printing point cloud is used as a seed to grow around, if the point cloud intensity difference between the previous growth and the subsequent growth is less than the threshold span range, the iteration is continued for 3 times.

4. The high-precision map raw data storage optimization method according to claim 1, characterized in that, The deletion of the road outside point cloud data with the point cloud density less than the point cloud density threshold comprises: widening the road to both sides by a preset distance range to obtain a road outside effective area, identifying the road outside effective area based on the point cloud analysis, performing plane identification and extraction on the point cloud in the road outside effective area, and filtering the discrete points based on the point cloud density threshold. 5.The high-precision map raw data storage optimization method according to claim 1, characterized in that, After the road surface point cloud data with the intensity less than the intensity threshold and the road outside point cloud data with the point cloud density less than the point cloud density threshold are deleted, a point cloud spacing threshold is set, and one of the adjacent point cloud data with an interval distance less than the point cloud spacing threshold is deleted.

6. The high-precision map raw data storage optimization method according to claim 4, characterized in that, The deletion of the picture data with the gray average value less than the gray average value threshold further comprises: in the road outside effective area, a deep learning model is used to identify whether it is sky and road vegetation, and the pixels in the sky and vegetation range are modified to black.

7. A high-precision map raw data storage optimization system, characterized by, The method comprises the following functional modules: a data association module, configured to acquire three-dimensional real scene data, and associate trajectory data, point cloud data and picture data based on timestamp information; a trajectory point pruning module, configured to set a trajectory point spacing threshold, and delete a subsequent trajectory point data with an interval distance from a previous trajectory point less than the trajectory point spacing threshold; a trajectory point pruning module, configured to set a trajectory point spacing threshold, and delete a subsequent trajectory point data with an interval distance from a previous trajectory point less than the trajectory point spacing threshold; The point cloud pruning module is configured to delete point cloud data corresponding to the deleted track point time, set an intensity threshold and a point cloud density threshold, delete road surface point cloud data with an intensity lower than the intensity threshold, and delete road outside point cloud data with a point cloud density lower than the point cloud density threshold. The road surface point cloud data with an intensity lower than the intensity threshold is deleted by using a seed growing method to identify road surface point cloud data based on the pruned point cloud data, and by using a watershed method to separate road surface printing from a road surface background based on the intensity threshold and to delete point cloud data of the road surface background. The picture pruning module is configured to delete picture data corresponding to the deleted track point time, and set a gray average value threshold to delete picture data with a gray average value lower than the gray average value threshold.

8. A server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the high-precision map original data storage optimization method of any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the high-precision map original data storage optimization method of any one of claims 1 to 6.

Images