Vehicle positioning method and device in indoor parking lot, equipment and storage medium

Abstract

The embodiment of the invention provides a vehicle positioning method and device in an indoor parking lot, electronic equipment and a computer readable storage medium, and relates to the technical field of navigation. The method comprises the following steps: acquiring navigation positioning data of a vehicle and parking lot image data; if the power-off memory pose cannot be read, calculating an initial pose of the vehicle based on the navigation positioning data; correcting the initial pose according to pre-acquired parking lot map data and parking lot image data to obtain a corrected pose of the vehicle; and optimizing the corrected pose to obtain a target pose of the vehicle. According to the embodiment of the invention, the pose correction of the parking lot map and the parking lot image acquired by the vehicle in real time is realized, and the accuracy and reliability of vehicle positioning can be effectively improved under the condition that the accuracy of the navigation positioning data acquired indoors is insufficient.

Description

Technical Field

[0001] The present application relates to the field of navigation technology, and in particular to a method, device, equipment, and storage medium for positioning a vehicle in an indoor parking lot. Background Art

[0002] With the acceleration of urbanization and the continuous increase in car ownership, the number of large parking lots is increasing. For example, indoor parking lots in places such as large shopping malls, airports, and office buildings can cover tens of thousands of square meters and have thousands of parking spaces.

[0003] However, in indoor environments, satellite signals can be severely attenuated or completely lost due to obstruction by buildings. Parking lots are often located inside or underground buildings, which have complex structures and significant signal shielding. This can cause satellite positioning systems like the Global Positioning System (GPS) to operate ineffectively, leading to inaccurate vehicle positioning in existing technologies. Summary of the Invention

[0004] The present application provides a method, apparatus, device and storage medium for positioning a vehicle in an indoor parking lot, which are used to solve the problem of inaccurate vehicle positioning in the prior art.

[0005] According to one aspect of the present application, a method for positioning a vehicle in an indoor parking lot is provided, the method comprising:

[0006] Collect vehicle navigation positioning data and parking lot image data;

[0007] If the power-off memory posture cannot be read, the initial posture of the vehicle is calculated based on the navigation positioning data;

[0008] Correcting the initial posture of the vehicle based on pre-acquired parking lot map data and parking lot image data to obtain a corrected posture of the vehicle;

[0009] The corrected posture is optimized to obtain the target posture of the vehicle.

[0010] In one possible implementation, the above-mentioned correcting the initial posture based on the pre-acquired parking lot map data and parking lot image data to obtain the corrected posture of the vehicle includes:

[0011] Extracting map semantic data of indoor parking lots based on parking lot map data, and extracting perception semantic data of indoor parking lots based on parking lot image data;

[0012] Performing multi-level rasterization processing on the semantic data to obtain image data; wherein, when the semantic data is map semantic data, the image data is a map; when the semantic data is perception semantic data, the image data is a perception image;

[0013] The initial posture is corrected based on the image data to obtain the corrected posture of the vehicle.

[0014] In one possible implementation, the map includes a first layer map and a second layer map; the first layer map includes a plurality of first map grids, and the second layer map includes a plurality of second map grids; the area of the second map grid is smaller than the area of the first map grid;

[0015] The above-mentioned correction of the initial posture based on the image data to obtain the corrected posture of the vehicle includes:

[0016] Projecting the initial pose onto the first layer map, and determining a plurality of first projection grids from the first map grid according to the projection result;

[0017] Based on pixel difference information between the first layer map and the first layer perception image, a first target grid is selected from the plurality of first projection grids;

[0018] Projecting the first target grid onto the second layer map, and determining a plurality of second projected grids from the second map grid;

[0019] Based on the pixel difference information between the second layer map and the second layer perception image, a second target grid is selected from the plurality of second projection grids;

[0020] A corrected pose of the vehicle is determined based on the second target grid.

[0021] In another possible implementation, the map further includes a third layer map; the third layer map includes a plurality of third map grids; the area of the second map grid is greater than the area of the third map grid;

[0022] The above-mentioned determination of the corrected posture of the vehicle based on the second target grid includes:

[0023] Projecting the second target grid onto a third layer map, and determining a plurality of third projection grids from the third map grid;

[0024] Based on the pixel difference information between the third layer map and the third layer perception image, a third target grid is screened out from a plurality of third projection grids;

[0025] Convert the third target grid to the corrected pose.

[0026] In another possible implementation, projecting the initial pose onto the first layer map and determining a plurality of first projection grids from the first map grid according to the projection result include:

[0027] Project the initial pose to the first layer of the map, and use the first map grid corresponding to the initial pose projection as the center grid;

[0028] Taking the central grid as the center and based on a preset expansion range, a plurality of first projection grids are selected from the first map grid.

[0029] In another possible implementation, the step of selecting the first target grid from the plurality of first projection grids based on the pixel difference information between the first layer map and the first layer perception image includes:

[0030] For each first projection grid, determining a first pixel value of the first projection grid in the first layer map and a second pixel value of the first projection grid projected onto the first layer perceptual image; and determining a pixel difference between the first pixel value and the second pixel value;

[0031] A first target grid is selected from the plurality of first projection grids based on a pixel difference; wherein the pixel difference corresponding to the first target grid is less than a preset pixel difference threshold.

[0032] In another possible implementation, the above-mentioned optimization of the corrected posture to obtain the target posture of the vehicle includes:

[0033] Rotating the corrected posture a preset number of times based on a preset rotation angle to obtain multiple initial corrected postures;

[0034] For each initial corrected pose, pose optimization is performed based on the least squares method to obtain the optimized pose corresponding to the initial corrected pose and the residual value corresponding to the optimized pose; the residual value is used to represent the difference between the map pixel value and the perceived image pixel value determined by projecting the initial corrected pose onto the map and the perceived image respectively;

[0035] A target pose is selected from multiple optimized poses; wherein the residual value corresponding to the target pose is less than a preset residual threshold.

[0036] According to another aspect of an embodiment of the present application, a vehicle positioning device in an indoor parking lot is provided, the device comprising:

[0037] An acquisition module, used to collect vehicle navigation positioning data and parking lot image data;

[0038] A calculation module, configured to calculate the vehicle's initial posture based on navigation positioning data if the power-off memory posture cannot be read;

[0039] A correction module, configured to correct the initial posture of the vehicle based on pre-acquired parking lot map data and parking lot image data to obtain a corrected posture of the vehicle;

[0040] The optimization module is used to optimize the corrected posture to obtain the target posture of the vehicle.

[0041] According to another aspect of the present application, an electronic device is provided, which includes: a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to implement the steps of the method shown in the first aspect of the present application.

[0042] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method shown in the first aspect of the present application are implemented.

[0043] The beneficial effects of the technical solution provided by this application are:

[0044] The vehicle positioning method in an indoor parking lot provided by this application can calculate the vehicle's initial position using the collected navigation positioning data when the power-off memory position cannot be read; then, the initial position is corrected based on the parking lot map data and parking lot image data to obtain the vehicle's corrected position, and the corrected position is optimized to obtain the vehicle's final target position. This application implements position correction by integrating the parking lot map and the parking lot image collected in real time by the vehicle. When the accuracy of the navigation positioning data collected indoors is insufficient, it can effectively improve the accuracy and reliability of vehicle positioning, providing strong support for subsequent parking space search or automatic parking functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0046] Figure 1 A flow chart of a method for locating a vehicle in an indoor parking lot provided in an embodiment of the present application;

[0047] Figure 2 A schematic diagram of a map semantic data structure in a vehicle positioning method in an indoor parking lot provided in an embodiment of the present application;

[0048] Figure 3 A schematic diagram of the rasterized structure of map semantic data in a vehicle positioning method in an indoor parking lot provided by an embodiment of the present application;

[0049] Figure 4 A schematic diagram of projection positioning of a three-level grid in a vehicle positioning method in an indoor parking lot provided by an embodiment of the present application;

[0050] Figure 5A schematic structural diagram of a vehicle positioning device in an indoor parking lot provided by an embodiment of the present application;

[0051] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely illustrative and are not intended to limit the scope of the present disclosure. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0053] The accompanying drawings illustrate various schematic diagrams of structures according to embodiments of the present disclosure. These figures are not drawn to scale, and for the purpose of clarity, certain details are exaggerated and certain details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0054] In the context of the present disclosure, when a layer / element is referred to as being "on" another layer / element, it can be directly on the other layer / element or an intervening layer / element may be present therebetween. In addition, if a layer / element is "on" another layer / element in one orientation, it may be "below" the other layer / element when the orientation is reversed.

[0055] With the rapid development of the internet and mobile applications, car owners have increasingly sophisticated expectations for their parking experience. They expect to be able to quickly find parking spaces through mobile apps and easily locate their vehicles once parked. Indoor positioning technology provides the technical support for this intelligent parking experience. For example, some parking lots have launched smart parking apps that leverage indoor positioning technology to provide car owners with features such as parking reservations, parking navigation, and reverse vehicle tracking, significantly enhancing their parking experience.

[0056] For parking lot managers, indoor positioning technology not only improves parking efficiency but also enables intelligent parking lot management. By monitoring parking space usage in real time, managers can optimize parking allocation and improve utilization. Furthermore, combining indoor positioning data allows for traffic flow analysis within the parking lot, optimizing the layout and signage system, and ultimately improving overall parking lot operations and management.

[0057] The inventors discovered that while common satellite positioning systems, such as GPS, perform well in outdoor environments, they can experience significant attenuation or complete loss of reception indoors due to obstruction by buildings. Parking lots are often located inside buildings or underground, and their complex structures can significantly shield signals, rendering GPS and other satellite positioning systems ineffective. For example, in underground parking lots, metal ceilings, walls, and vehicles themselves can interfere with satellite signals, preventing GPS receivers from receiving sufficient satellite signals to determine location.

[0058] Based on the above technical issues, some embodiments of the present application can calculate the vehicle's initial posture through the collected navigation positioning data when the power-off memory posture cannot be read; then, the initial posture is corrected based on the parking lot map data and parking lot image data to obtain the vehicle's corrected posture, and the corrected posture is optimized to obtain the vehicle's final target posture. This application realizes posture correction by integrating the parking lot map and the parking lot image collected by the vehicle in real time. In the case of insufficient accuracy of navigation positioning data collected indoors, it can effectively improve the accuracy and reliability of vehicle positioning, providing strong support for subsequent parking space search or automatic parking functions.

[0059] The following describes in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0060] The embodiment of the present application provides a method for positioning a vehicle in an indoor parking lot. Figure 1 As shown, the method includes:

[0061] S101, collecting vehicle navigation positioning data and parking lot image data.

[0062] The above-mentioned navigation and positioning data may include inertial measurement data, vehicle speed, global navigation satellite data, etc.

[0063] Specifically, the navigation positioning data can be collected in real time by the navigation system, and the parking lot image data can be collected in real time by the vehicle body image collection device or the infrared sensor.

[0064] Optionally, navigation positioning data can be collected in real time through the navigation system, and the environmental image outside the vehicle can be collected through the image acquisition device; when the vehicle position is determined to be near the parking lot based on the above navigation positioning data, and the key components of the parking lot (such as the parking lot gate) are identified through the environmental image, the above environmental image is used as parking lot image data.

[0065] In some embodiments, the validity of the navigation positioning data and parking lot image data collected in real time can be judged. When it is judged that there is a disconnection or abnormality in the data, a fault warning can be issued to remind the car owner to slow down and pick up the car. When it is judged that there is no disconnection or abnormality in the data, that is, the data is valid, subsequent data processing will continue.

[0066] S102: If the power-off memory posture cannot be read, the initial posture of the vehicle is calculated based on the navigation positioning data.

[0067] Specifically, the vehicle's power-off memory posture can be read from a preset database. If the reading fails, it indicates that the vehicle entered the parking lot dynamically rather than being powered on while stationary in the parking lot. In this case, the vehicle's initial posture is calculated using an Error-State Kalman Filter (ESKF) based on the navigation positioning data collected in real time.

[0068] S103 , correcting the initial posture according to the pre-acquired parking lot map data and parking lot image data to obtain a corrected posture of the vehicle.

[0069] The parking lot map data may be obtained from a preset map library.

[0070] Specifically, the parking lot can be identified based on the vehicle's navigation positioning data and the parking lot image data to obtain a parking lot sign; then, based on the parking lot sign, parking lot map data corresponding to the parking lot sign can be obtained from a preset map library.

[0071] Optionally, the parking lot map data and the parking lot image data may be semantically converted to obtain semantic data, and then the initial posture may be corrected using the semantic data to obtain a corrected posture.

[0072] S104: Optimize the corrected posture to obtain the target posture of the vehicle.

[0073] Optionally, the corrected posture can be optimized based on the least squares method to obtain the target posture of the vehicle.

[0074] In some embodiments, if the power-off memory posture is read, it indicates that the vehicle is in a stationary and powered-on state in the parking lot, and the above-mentioned power-off memory posture is a relatively accurate vehicle posture; that is, there is no need to perform the above-mentioned correction operation on the power-off memory posture, and the above-mentioned power-off memory posture can be directly optimized to obtain the target posture of the vehicle.

[0075] In an embodiment of the present application, when the power-off memorized posture cannot be read, the vehicle's initial posture can be calculated using the collected navigation positioning data. The initial posture is then corrected based on the parking lot map data and parking lot image data to obtain the vehicle's corrected posture, and the corrected posture is optimized to obtain the vehicle's final target posture. This application implements posture correction by integrating the parking lot map and parking lot images collected in real time by the vehicle. In the case of insufficient accuracy of navigation positioning data collected indoors, it can effectively improve the accuracy and reliability of vehicle positioning, providing strong support for subsequent parking space search or automatic parking functions.

[0076] In an embodiment of the present application, a possible implementation method is provided, wherein the initial posture is corrected based on the pre-acquired parking lot map data and parking lot image data to obtain the corrected posture of the vehicle, including:

[0077] S201 , extracting map semantic data of an indoor parking lot based on parking lot map data, and extracting perception semantic data of the indoor parking lot based on parking lot image data.

[0078] The image semantic data is used to represent the geometric information of each element in the map, and the perception semantic data is used to represent the geometric information of each element in the parking lot image.

[0079] Optionally, elements in the parking lot may include pillars, parking spaces, and ground signs, etc., which are not specifically limited in the embodiments of the present application.

[0080] S202, performing multi-level rasterization processing on the semantic data to obtain image data.

[0081] When the semantic data is map semantic data, the image data is a map; when the semantic data is perception semantic data, the image data is a perception image.

[0082] Specifically, taking the three-level rasterization process as an example: the map semantic data can be rasterized using a square with a first side length to obtain multiple first map grids; then, the elements in each first map grid and the number of each element are counted to determine the state value of the first map grid; and the state value of each first map grid is normalized to serve as the pixel value of the first map grid; and the pixel values of all first map grids are used as the pixel information of the first layer map;

[0083] Optionally, the pixel information of the second layer map can be obtained based on the same rasterization processing steps as above with the second side length; the pixel information of the third layer map can be obtained based on the same rasterization processing steps as above with the third side length.

[0084] Optionally, the second side length may be half of the first side length, and the third side length may be half of the second side length.

[0085] In the embodiment of this application, Figure 2 As an example of the map semantic data shown in the figure, the first level rasterization process can be performed on a square with a side length of L to obtain Figure 3 ; Calculate each grid and count the number n of each element (including but not limited to parking spaces, columns, and ground signs) in the grid. When the semantics cross grids, the current semantics needs to be counted in each grid where it is located and record a coefficient c k , where c k The calculation is as follows:

[0086]

[0087] Among them, s i is the area of a single semantic in grid i, s all is the total area of a single semantics; c k Needs to be satisfied The coefficient without cross-grid is c0=1, and the state value S of each grid is obtained ij :

[0088] S ij =

[0089] {(n park ,(c0,c1,c2,...)),(n pole ,(c0,c1,c2,...)),(n marker ,(c0,c1,c2,...))...} (2)

[0091] Among them, S ij is the grid status value of row i and column j. Figure 2 The pixel information of the first layer map obtained by rasterization is shown in Table 1 below;

[0092]

[0093] Table 1

[0094] Further, traverse each S ij , add the weights of each element in a single grid to get a new value p ij (v1, v2, v3) (taking three elements as an example), normalize the three-dimensional v values to the range of 0 to 255, that is, convert the map into RGB pixel data I img .

[0095] S203: Correct the initial posture based on the image data to obtain a corrected posture of the vehicle.

[0096] Specifically, the initial pose can be projected onto the image data, and the initial pose can be corrected according to the pixel difference between the map and the perceived image in the projection range to obtain the corrected pose.

[0097] like Figure 4 As shown in the figure, there are three levels of rasterization processing. The corrected posture is determined by grid projection layer by layer. The higher the projection level, the smaller the area of a single grid and the more precise the vehicle positioning.

[0098] The embodiment of the present application obtains image data such as maps and perception images through multi-level rasterization processing. The above image data can represent the geometric information of each element in the indoor parking lot in a hierarchical manner; the initial posture can be corrected based on the difference between the map and the perception image, which effectively improves the correction efficiency and lays a good foundation for subsequent posture optimization.

[0099] It should be noted that the above embodiment takes the rasterization processing of three levels of resolution as an example, and can also be extended to raster processing of N levels of resolution, where N is a positive integer not less than 2, and is not specifically limited in the embodiments of this application.

[0100] In an embodiment of the present application, a possible implementation method is provided, wherein the map includes a first-layer map and a second-layer map; the first-layer map includes a plurality of first map grids, and the second-layer map includes a plurality of second map grids; the area of the second map grid is smaller than the area of the first map grid;

[0101] The above-mentioned correction of the initial posture based on the image data to obtain the corrected posture of the vehicle includes:

[0102] S301: Project the initial posture to the first layer map, and determine a plurality of first projection grids from the first map grid according to the projection result.

[0103] S302 : Screening out a first target grid from a plurality of first projection grids based on pixel difference information between the first layer map and the first layer perception image.

[0104] Specifically, the first layer perception image can be projected onto the first layer map with the first projection grid as the origin, and the pixel difference Δp between the first layer map and the first layer perception image in each first projection grid can be determined. i ; The first projection grid corresponding to the minimum pixel difference is used as the first target grid.

[0105] Δp i =p(v mapi1 -v peri1 ,v mapi2 -v peri2 ,v mapi3 -v peri3 )

[0106] Among them, (vmapi1 ,v mapi2 ,v mapi3 ) is the pixel value of the first projection grid in the first layer map, (v peri1 ,v peri2 ,v peri3 ) is the pixel value of the i-th first projection grid in the first layer item; i∈[1,k], k is the total number of first projection grids.

[0107] S303: Project the first target grid onto the second layer map, and determine a plurality of second projection grids from the second map grids.

[0108] Specifically, the map grid covered by the first target grid may be selected from the second map grid as the second projection grid.

[0109] S304 : Screening out a second target grid from the plurality of second projection grids based on pixel difference information between the second layer map and the second layer perception image.

[0110] Optionally, the process of determining the second target grid is similar to the screening step in S302 and will not be described in detail here.

[0111] S305: Determine a corrected posture of the vehicle based on the second target grid.

[0112] Specifically, the above correction posture determination process will be described in detail below.

[0113] A possible implementation method is provided in an embodiment of the present application, where the above-mentioned map also includes a third layer map; the third layer map includes multiple third map grids; and the area of the second map grid is larger than the area of the third map grid.

[0114] The above-mentioned determination of the corrected posture of the vehicle based on the second target grid includes:

[0115] S401, projecting the second target grid onto the third layer map, determining a plurality of third projection grids from the third map grids; and screening a third target grid from the plurality of third projection grids based on pixel difference information between the third layer map and the third layer perception image.

[0116] The above projection and screening steps are similar to S303 and S304 and will not be described in detail here.

[0117] S402: Convert the third target grid into a corrected posture.

[0118] Specifically, the corrected posture can be obtained by performing conversion based on the coordinate information of the third target grid in the image data.

[0119] In the embodiment of the present application, through three levels of rasterization processing, the initial posture can be grid-projected and matched at multiple levels. The higher the level, the more refined and accurate the geometric information in the grid, which effectively enhances the accuracy of the corrected posture.

[0120] In an embodiment of the present application, a possible implementation method is provided, in which the initial posture is projected onto the first layer map, and a plurality of first projection grids are determined from the first map grid according to the projection result, including:

[0121] S501: Project the initial posture onto the first layer map, and use the first map grid corresponding to the initial posture projection as the center grid.

[0122] S502 : Taking the central grid as the center and based on a preset extension range, select a plurality of first projection grids from the first map grid.

[0123] In the embodiment of the present application, n grids may be expanded in four directions with the central grid as the center to obtain multiple first projection grids; the value of n may generally be in the range of [3, 5].

[0124] In an embodiment of the present application, a possible implementation method is provided, wherein the first target grid is selected from a plurality of first projection grids based on the pixel difference information between the first layer map and the first layer perception image, including:

[0125] S601 , for each first projection grid, determining a first pixel value of the first projection grid in a first layer map and a second pixel value of the first projection grid projected onto a first layer perception image; and determining a pixel difference between the first pixel value and the second pixel value.

[0126] S602 : Filtering out a first target grid from a plurality of first projection grids based on pixel difference; wherein the pixel difference corresponding to the first target grid is less than a preset pixel difference threshold.

[0127] Specifically, the first layer perception image can be projected onto the first layer map with the first projection grid as the origin, and the pixel difference Δp between the first layer map and the first layer perception image in each first projection grid can be determined. i ; The first projection grid corresponding to the minimum pixel difference is used as the first target grid.

[0128] Δp i =p(v mapi1 -v peri1 ,v mapi2 -v peri2 ,v mapi3 -v peri3 )

[0129] Among them, (v mapi1 ,v mapi2 ,vmapi3 ) is the pixel value of the first projection grid in the first layer map, (v peri1 ,v peri2 ,v peri3 ) is the pixel value of the i-th first projection grid in the first layer item; i∈[1,k], k is the total number of first projection grids.

[0130] In an embodiment of the present application, a possible implementation method is provided, wherein the above-mentioned optimization of the correction posture to obtain the target posture of the vehicle includes:

[0131] S701: Rotate the corrected posture a preset number of times based on a preset rotation angle to obtain a plurality of initial corrected postures.

[0132] Specifically, a rotation of θ degrees can be preset, and x rotations can be performed with m degrees (recommended about 3 degrees) as a single preset rotation angle to obtain multiple initial correction postures, where x = θ / m.

[0133] S702 , performing posture optimization based on the least squares method for each initial correction posture, and obtaining an optimized posture corresponding to the initial correction posture and a residual value corresponding to the optimized posture.

[0134] Among them, the residual value is used to represent the difference between the map pixel value and the perception image pixel value determined by projecting the initial corrected pose onto the map and the perception image respectively.

[0135] S703: Filter out a target pose from multiple optimized poses.

[0136] Among them, the residual value corresponding to the target posture is less than the preset residual threshold.

[0137] Specifically, each initial correction pose can be traversed, and the initial correction pose can be associated with the semantic information of the relative correspondence between the perceived parking space, ground signs, poles and maps. The optimization problem can be constructed using point-to-line and point-to-point residuals to solve the optimal residual. The residuals corresponding to each initial correction pose can be compared, and the initial correction pose corresponding to the minimum residual can be obtained as the target pose of the vehicle.

[0138] The embodiment of the present application further obtains the final target posture of the vehicle through rotation expansion of the correction posture and posture optimization using the least squares method, so that the target posture is more closely matched with the map information elements, effectively improving the vehicle positioning accuracy and providing effective data support for subsequent automatic parking services, parking space search and other services of the vehicle.

[0139] To better understand the above-mentioned method for locating a vehicle in an indoor parking lot, an example of the method for locating a vehicle in an indoor parking lot of the present application is described in detail below. The method is applied to a smart car and includes the following steps:

[0140] S801, real-time collection of vehicle navigation positioning data and parking lot image data.

[0141] The above-mentioned navigation and positioning data may include inertial measurement data, vehicle speed, global navigation satellite data, etc.

[0142] S802: The validity of the navigation positioning data and parking lot image data is judged. If there is no interruption or abnormality in the data, that is, the data is valid, an attempt is made to read the vehicle's power-off memory posture.

[0143] S803: If the power-off memory posture cannot be read, the initial posture of the vehicle is calculated based on the navigation positioning data using an error state Kalman filter.

[0144] S804 , extracting map semantic data of the indoor parking lot based on the parking lot map data, and extracting perception semantic data of the indoor parking lot based on the parking lot image data.

[0145] S805: Perform multi-level rasterization processing on the semantic data to obtain image data.

[0146] When the semantic data is map semantic data, the image data is a map; when the semantic data is perception semantic data, the image data is a perception image.

[0147] Among them, the above-mentioned map includes a first-layer map, a second-layer map and a third-layer map; the first-layer map includes multiple first map grids, the second-layer map includes multiple second map grids; the third-layer map includes multiple third map grids; the area of the second map grid is smaller than the area of the first map grid; the area of the second map grid is larger than the area of the third map grid.

[0148] S806: Project the initial position onto the first layer map, determine a plurality of first projection grids from the first map grids based on the projection result, and select a first target grid from the plurality of first projection grids based on pixel difference information between the first layer map and the first layer perception image.

[0149] S807: Project the first target grid onto the second layer map, determine a plurality of second projection grids from the second map grids, and select a second target grid from the plurality of second projection grids based on pixel difference information between the second layer map and the second layer perception image.

[0150] S808: Project the second target grid onto the third layer map, and determine a plurality of third projection grids from the third map grids; select a third target grid from the plurality of third projection grids based on pixel difference information between the third layer map and the third layer perception image; and convert the third target grid into a corrected posture.

[0151] S809, rotate the corrected posture a preset number of times based on a preset rotation angle to obtain multiple initial corrected postures; for each initial corrected posture, perform posture optimization based on the least squares method to obtain the optimized posture corresponding to the initial corrected posture and the residual value corresponding to the optimized posture; then select the corresponding residual value from the multiple optimized postures to make a small target posture.

[0152] The embodiment of the present application provides a vehicle positioning device in an indoor parking lot, such as Figure 5 As shown, the vehicle positioning device 50 in the indoor parking lot may include: a collection module 501, a calculation module 502, a correction module 503 and an optimization module 504;

[0153] The acquisition module 501 is used to acquire the vehicle's navigation positioning data and parking lot image data;

[0154] The calculation module 502 is configured to calculate the initial posture of the vehicle based on the navigation positioning data if the power-off memory posture cannot be read;

[0155] A correction module 503 is used to correct the initial posture of the vehicle based on the pre-acquired parking lot map data and parking lot image data to obtain a corrected posture of the vehicle;

[0156] The optimization module 504 is used to optimize the corrected posture to obtain the target posture of the vehicle.

[0157] In one embodiment of the present application, a possible implementation is provided. When the correction module 503 corrects the initial posture based on the pre-acquired parking lot map data and parking lot image data to obtain the corrected posture of the vehicle, it is configured to:

[0158] Extracting map semantic data of indoor parking lots based on parking lot map data, and extracting perception semantic data of indoor parking lots based on parking lot image data;

[0159] Performing multi-level rasterization processing on the semantic data to obtain image data; wherein, when the semantic data is map semantic data, the image data is a map; when the semantic data is perception semantic data, the image data is a perception image;

[0160] The initial posture is corrected based on the image data to obtain the corrected posture of the vehicle.

[0161] In an embodiment of the present application, a possible implementation method is provided, wherein the map includes a first-layer map and a second-layer map; the first-layer map includes a plurality of first map grids, and the second-layer map includes a plurality of second map grids; the area of the second map grid is smaller than the area of the first map grid;

[0162] When correcting the initial posture based on the image data to obtain the corrected posture of the vehicle, the correction module 503 is used to:

[0163] Projecting the initial pose onto the first layer map, and determining a plurality of first projection grids from the first map grid according to the projection result;

[0164] Based on pixel difference information between the first layer map and the first layer perception image, a first target grid is selected from the plurality of first projection grids;

[0165] Projecting the first target grid onto the second layer map, and determining a plurality of second projected grids from the second map grid;

[0166] Based on the pixel difference information between the second layer map and the second layer perception image, a second target grid is selected from the plurality of second projection grids;

[0167] A corrected pose of the vehicle is determined based on the second target grid.

[0168] In an embodiment of the present application, a possible implementation method is provided, wherein the map further includes a third layer map; the third layer map includes a plurality of third map grids; the area of the second map grid is larger than the area of the third map grid;

[0169] When determining the corrected posture of the vehicle based on the second target grid, the correction module 503 is used to:

[0170] Projecting the second target grid onto a third layer map, and determining a plurality of third projection grids from the third map grid;

[0171] Based on the pixel difference information between the third layer map and the third layer perception image, a third target grid is screened out from a plurality of third projection grids;

[0172] Convert the third target grid to the corrected pose.

[0173] In an embodiment of the present application, a possible implementation is provided. When the correction module 503 projects the initial posture onto the first layer map and determines a plurality of first projection grids from the first map grid according to the projection result, the correction module 503 is configured to:

[0174] Project the initial pose to the first layer of the map, and use the first map grid corresponding to the initial pose projection as the center grid;

[0175] Taking the central grid as the center and based on a preset expansion range, a plurality of first projection grids are selected from the first map grid.

[0176] In an embodiment of the present application, a possible implementation is provided. When the correction module 503 selects the first target grid from the plurality of first projection grids based on the pixel difference information between the first layer map and the first layer perception image, the correction module 503 is configured to:

[0177] For each first projection grid, determining a first pixel value of the first projection grid in the first layer map and a second pixel value of the first projection grid projected onto the first layer perceptual image; and determining a pixel difference between the first pixel value and the second pixel value;

[0178] A first target grid is selected from the plurality of first projection grids based on a pixel difference; wherein the pixel difference corresponding to the first target grid is less than a preset pixel difference threshold.

[0179] In an embodiment of the present application, a possible implementation is provided. When optimizing the corrected posture to obtain the target posture of the vehicle, the optimization module 504 is configured to:

[0180] Rotating the corrected posture a preset number of times based on a preset rotation angle to obtain multiple initial corrected postures;

[0181] For each initial corrected pose, pose optimization is performed based on the least squares method to obtain the optimized pose corresponding to the initial corrected pose and the residual value corresponding to the optimized pose; the residual value is used to represent the difference between the map pixel value and the perceived image pixel value determined by projecting the initial corrected pose onto the map and the perceived image respectively;

[0182] A target pose is selected from multiple optimized poses; wherein the residual value corresponding to the target pose is less than a preset residual threshold.

[0183] The device of the embodiment of the present application can execute the method provided by the embodiment of the present application, and its implementation principle is similar. The actions performed by each module in the device of each embodiment of the present application correspond to the steps in the method of each embodiment of the present application. For the detailed functional description of each module of the device, please refer to the description in the corresponding method shown in the previous text, and will not be repeated here.

[0184] In an embodiment of the present application, when the power-off memorized posture cannot be read, the vehicle's initial posture can be calculated using the collected navigation positioning data. The initial posture is then corrected based on the parking lot map data and parking lot image data to obtain the vehicle's corrected posture, and the corrected posture is optimized to obtain the vehicle's final target posture. This application implements posture correction by integrating the parking lot map and parking lot images collected in real time by the vehicle. In the case of insufficient accuracy of navigation positioning data collected indoors, it can effectively improve the accuracy and reliability of vehicle positioning, providing strong support for subsequent parking space search or automatic parking functions.

[0185] In an embodiment of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory. The processor executes the above-mentioned computer program to implement the steps of a method for positioning a vehicle in an indoor parking lot. Compared with the related art, the following can be achieved: in the case where the power-off memory posture cannot be read, the embodiment of the present application can calculate the initial posture of the vehicle through the collected navigation positioning data; then, based on the parking lot map data and the parking lot image data, the initial posture is corrected to obtain the corrected posture of the vehicle, and the corrected posture is optimized to obtain the final target posture of the vehicle. The present application realizes posture correction by integrating the parking lot map and the parking lot image collected in real time by the vehicle. In the case where the accuracy of the navigation positioning data collected indoors is insufficient, it can effectively improve the accuracy and reliability of vehicle positioning, and provide strong support for subsequent parking space search or automatic parking functions.

[0186] In an alternative embodiment, an electronic device is provided, such as Figure 6 As shown, Figure 6 The electronic device 60 shown includes: a processor 601 and a memory 603. The processor 601 and the memory 603 are connected, for example, via a bus 602. Optionally, the electronic device 60 may further include a transceiver 604, which can be used for data exchange between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the number of transceivers 604 is not limited to one, and the structure of the electronic device 60 does not constitute a limitation on the embodiments of the present application.

[0187] The processor 601 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor 601 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0188] The bus 602 may include a path for transmitting information between the above components. The bus 602 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 602 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0189] The memory 603 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store computer programs and can be read by a computer, without limitation here.

[0190] The memory 603 is used to store the computer program for executing the embodiment of the present application, and the execution is controlled by the processor 601. The processor 601 is used to execute the computer program stored in the memory 603 to implement the steps shown in the above method embodiment.

[0191] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, PADs, and fixed terminals such as vehicle-mounted terminals.

[0192] In an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method shown in the first aspect of the embodiment of the present application are implemented.

[0193] While the above description does not provide detailed technical details regarding the patterning of each layer, those skilled in the art will appreciate that various technical means can be employed to form layers, regions, and the like in desired shapes. Furthermore, those skilled in the art may devise methods that differ from those described above to achieve the same structure. Furthermore, while each embodiment has been described separately, this does not mean that the measures in each embodiment cannot be advantageously combined.

[0194] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0195] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A vehicle positioning method in an indoor parking lot, characterized in that: The method comprises: Collect vehicle navigation positioning data and parking lot image data; If the power-off memory posture cannot be read, the initial posture of the vehicle is calculated based on the navigation positioning data; Correcting the initial posture according to pre-acquired parking lot map data and the parking lot image data to obtain a corrected posture of the vehicle; The corrected posture is optimized to obtain a target posture of the vehicle.

2. The method according to claim 1, characterized in that The correcting the initial posture according to the pre-acquired parking lot map data and the parking lot image data to obtain the corrected posture of the vehicle includes: extracting map semantic data of the indoor parking lot based on the parking lot map data, and extracting perception semantic data of the indoor parking lot based on the parking lot image data; Performing multi-level rasterization processing on the semantic data to obtain image data; wherein, when the semantic data is map semantic data, the image data is a map; when the semantic data is perceptual semantic data, the image data is a perceptual image; The initial posture is corrected based on the image data to obtain a corrected posture of the vehicle.

3. The method according to claim 2, characterized in that The map includes a first layer map and a second layer map; the first layer map includes a plurality of first map grids, and the second layer map includes a plurality of second map grids; the area of the second map grid is smaller than the area of the first map grid; The correcting the initial posture based on the image data to obtain a corrected posture of the vehicle includes: Projecting the initial posture onto a first layer map, and determining a plurality of first projection grids from the first map grids according to the projection result; Screening out a first target grid from a plurality of first projection grids based on pixel difference information between the first layer map and the first layer perception image; Projecting the first target grid onto a second layer map, and determining a plurality of second projected grids from the second map grid; Screening out a second target grid from a plurality of second projection grids based on pixel difference information between the second layer map and the second layer perception image; A corrected posture of the vehicle is determined based on the second target grid.

4. The method according to claim 3, characterized in that The map further includes a third layer map; the third layer map includes a plurality of third map grids; the area of the second map grid is larger than the area of the third map grid; The determining the corrected posture of the vehicle based on the second target grid includes: Projecting the second target grid onto a third layer map, and determining a plurality of third projected grids from the third map grid; Screening out a third target grid from the plurality of third projection grids based on pixel difference information between the third layer map and the third layer perception image; The third target grid is converted into a corrected pose.

5. The method according to claim 3, characterized in that The projecting the initial posture to the first layer map, and determining a plurality of first projection grids from the first map grid according to the projection result, comprises: Projecting the initial posture onto the first layer map, and taking the first map grid corresponding to the initial posture projection as the center grid; Taking the central grid as the center and based on a preset expansion range, a plurality of first projection grids are selected from the first map grids.

6. The method according to claim 3, characterized in that The step of selecting a first target grid from a plurality of first projection grids based on pixel difference information between the first layer map and the first layer perception image includes: For each of the first projected grids, determining a first pixel value of the first projected grid in the first layer map and a second pixel value of the first projected grid projected onto the first layer perceptual image; and determining a pixel difference between the first pixel value and the second pixel value; A first target grid is selected from the plurality of first projection grids based on the pixel difference; wherein the pixel difference corresponding to the first target grid is less than a preset pixel difference threshold.

7. The method according to claim 1, characterized in that Optimizing the corrected posture to obtain the target posture of the vehicle includes: Rotating the corrected posture a preset number of times based on a preset rotation angle to obtain a plurality of initial corrected postures; For each initial corrected pose, perform pose optimization based on the least squares method to obtain an optimized pose corresponding to the initial corrected pose and a residual value corresponding to the optimized pose; wherein the residual value is used to represent the difference between the map pixel value and the perceived image pixel value determined when the initial corrected pose is projected onto the map and the perceived image respectively; A target posture is selected from the plurality of optimized postures; wherein the residual value corresponding to the target posture is less than a preset residual threshold.

8. A vehicle positioning device in an indoor parking lot, characterized in that: The device comprises: An acquisition module, used to collect vehicle navigation positioning data and parking lot image data; A calculation module, configured to calculate an initial posture of the vehicle based on the navigation positioning data if the power-off memory posture cannot be read; a correction module, configured to correct the initial posture according to pre-acquired parking lot map data and the parking lot image data to obtain a corrected posture of the vehicle; An optimization module is used to optimize the corrected posture to obtain a target posture of the vehicle.

9. An electronic device, characterized in that: The method is characterized in that it comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images