Data detection method and apparatus, electronic device, and storage medium

Abstract

The application discloses a data detection method and device, electronic equipment and storage medium. The method comprises the following steps: acquiring map rendering data, wherein the map rendering data is rendering data used when a map is generated; detecting the map rendering data through multiple detection modes to obtain multiple detection results, wherein the detection result is used at least to indicate the confidence of a road element in the map under the map rendering data; in the case that the target detection result corresponding to the data is inconsistent with the preset result, adjusting the adjustment parameter corresponding to the multiple detection modes in the detection process, wherein the target detection result is determined by at least one of the multiple detection results; and re-detecting the map rendering data according to the adjustment parameter. The application solves the technical problem that the existing technology checks the quality of the high-precision map manually, resulting in low efficiency and large workload.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of high-precision maps, in particular to a data detection method and device, electronic equipment and a storage medium. BACKGROUND

[0002] In recent years, with the deep integration of artificial intelligence and surveying and mapping and the automobile industry, automatic driving and high-precision map technology have gradually become the focus of the industry. In the process of making high-precision maps, the mapping results of high-precision maps need to be inspected. The existing technology uses a large amount of manpower for manual verification, resulting in low detection efficiency and heavy workload.

[0003] At present, there is no effective solution to the above problems. SUMMARY

[0004] The embodiments of the present application provide a data detection method and device, electronic equipment and a storage medium to at least solve the technical problem of low efficiency and heavy workload caused by manual quality verification of high-precision maps in the prior art.

[0005] According to an aspect of an embodiment of the present application, a data detection method is provided, comprising: obtaining map rendering data, wherein the map rendering data is rendering data used when generating a map; detecting the map rendering data by multiple detection methods respectively to obtain multiple detection results, wherein the detection result is used at least to indicate the confidence of the road element in the map under the map rendering data; adjusting the adjustment parameter corresponding to the multiple detection methods in the detection process in the case that the target detection result corresponding to the data is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results; and re-detecting the map rendering data according to the adjustment parameter.

[0006] Optionally, detecting the map rendering data by multiple detection methods respectively comprises: generating trajectory playback data according to the map rendering data; and performing path planning on the map according to the trajectory playback data to obtain at least one target road path.

[0007] Optionally, performing path planning on the map according to the trajectory playback data comprises: selecting a root node from the trajectory playback data, wherein the root node is a node in the trajectory playback data that is connected to only one road line; determining a second road line having a connection relationship with a first road line where the root node is located according to the root node; determining a road path according to at least the first road line and the second road line; and traversing a tree structure generated according to the road path and determining at least one target road path corresponding to the tree structure.

[0008] Optionally, in the case of the detection mode being the first detection mode, the map rendering data is detected by multiple detection modes respectively, including: loading the target road path and the map rendering data into a detection tool corresponding to the first detection mode; controlling the vehicle to drive in the map rendered by the map rendering data according to the target road path, to generate a first recorded video; splitting the first recorded video into multiple first pictures according to a splitting principle; inputting the multiple first pictures into a first detection model for identification, to obtain a first detection result, wherein the first detection model is used for detecting road defects, and the first detection result at least includes a confidence degree of a classification to which each road element in the map belongs.

[0009] Optionally, in the case of the detection mode being the second detection mode, the map rendering data is detected by multiple detection modes respectively, including: loading the target road path and the map rendering data into a detection tool corresponding to the second detection mode; simulating the vehicle to drive in the map rendered by the map rendering data according to the target road path, to generate a second recorded video; splitting the second recorded video into multiple second pictures according to a splitting principle; inputting the multiple second pictures into a second detection model for identification, to obtain a second detection result, wherein the second detection model is used for simulating the first detection model to detect road defects, and the second detection result includes a detection result of an internal logic of a road element, the internal logic of the road element being determined based on a confidence degree of a classification to which each road element in the map belongs and a driving state of the vehicle.

[0010] Optionally, the method further includes: obtaining a first detection mode and a second detection mode in the multiple detection modes; determining a first detection result corresponding to the first detection mode and a first weight value, and a second detection result corresponding to the second detection mode and a second weight value; in the case that an absolute value of a difference between the first weight value and the second weight value is less than a first preset threshold, comparing the first detection result and the second detection result; in the case that an absolute value of a difference between a confidence degree in the first detection result and a confidence degree in the second detection result is greater than a second preset threshold, determining that a difference between the first detection result and the second detection result is a difference caused by a compiling method used by the first detection mode and the second detection mode.

[0011] Optionally, in the case of the detection mode being the third detection mode, the map rendering data is detected by multiple detection modes respectively, including: obtaining first road information collected by the vehicle in a driving process; comparing the first road information and second road information, to obtain a comparison result, wherein the second road information is road information in the map rendered by the map rendering data.

[0012] Optionally, the comparing the first road information and the second road information comprises: obtaining a first coordinate in the first road information and a second coordinate in the second road information; and determining a comparison result according to the first coordinate and the second coordinate, wherein the confidence of the different classification to which any road element in the map belongs is determined according to a condition met by the comparison result.

[0013] Optionally, the determining the confidence of the different classification to which any road element in the map belongs comprises: when the comparison result is greater than or equal to a first parameter, determining the confidence as 0, wherein the first parameter is used to represent a matching degree of data in the first road information and data in the second road information; when the comparison result is 0, determining the confidence as 1; and when the comparison result is less than the first parameter, determining the confidence according to the comparison result and the first parameter.

[0014] Optionally, when the detection manner is the third detection manner, the detecting the map rendering data by the plurality of detection manners comprises: determining a confidence threshold corresponding to each road element in the map, wherein the confidence threshold is determined by at least a minimum value of the confidence of the classification to which a point constituting the road element belongs; and determining data when the confidence is less than the confidence threshold as the third detection result.

[0015] Optionally, the adjusting the adjustment parameters corresponding to the plurality of detection manners in the detection process comprises: performing normalization processing on the plurality of detection results to obtain a target detection result, wherein the target detection result is determined by at least one of the plurality of detection results and a second parameter, and the second parameter is used to represent a weight value corresponding to the plurality of detection results respectively; determining that data corresponding to the target detection result is incorrect when the target detection result is greater than a target threshold; comparing the target detection result with a preset result; and adjusting the first parameter and the second parameter when the target detection result is inconsistent with the preset result.

[0016] According to another aspect of the embodiments of the present application, a data detection device is further provided, comprising: an acquisition module configured to acquire map rendering data, wherein the map rendering data is rendering data used when a map is generated; a first detection module configured to detect the map rendering data by a plurality of detection manners respectively to obtain a plurality of detection results, wherein the detection result is used to indicate at least a confidence of a road element in the map under the map rendering data; an adjustment module configured to adjust adjustment parameters corresponding to the plurality of detection manners in a detection process when data corresponding to a target detection result is inconsistent with a preset result, wherein the target detection result is determined by at least one of the plurality of detection results; and a second detection module configured to re-detect the map rendering data according to the adjustment parameters.

[0017] According to a further aspect of the embodiments of the present application, an electronic device is also provided, comprising: a memory for storing program instructions; a processor connected with the memory, for executing the program instructions to realize the following functions: obtaining map rendering data, wherein the map rendering data is rendering data used when generating a map; detecting the map rendering data by multiple detection methods respectively to obtain multiple detection results, wherein the detection results are used at least to indicate the confidence of road elements in the map under the map rendering data; adjusting the adjustment parameters corresponding to the multiple detection methods in the detection process in the case that the data corresponding to the target detection result is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results; and re-detecting the map rendering data according to the adjustment parameters.

[0018] According to a further aspect of the embodiments of the present application, a non-volatile storage medium is also provided, comprising a stored computer program, wherein the device where the non-volatile storage medium is located executes the above-mentioned data detection method by running the computer program.

[0019] In the embodiments of the present application, the map rendering data is obtained, wherein the map rendering data is rendering data used when generating a map; the map rendering data is detected by multiple detection methods respectively to obtain multiple detection results, wherein the detection results are used at least to indicate the confidence of road elements in the map under the map rendering data; the adjustment parameters corresponding to the multiple detection methods in the detection process are adjusted in the case that the data corresponding to the target detection result is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results; and the map rendering data is re-detected according to the adjustment parameters, which achieves the purpose of detecting the map rendering data by multiple methods, thereby realizing the technical effect of improving the detection efficiency of the map rendering data, and further solving the technical problem that the existing technology checks the quality of the high-precision map manually, resulting in low efficiency and large workload. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which are included to provide a further understanding of the present application, constitute a part of the present application and illustrate the illustrative embodiments of the present application and their description serve to explain the present application, and do not constitute improper limitations on the present application. In the drawings:

[0021] Figure 1 is a hardware structure block diagram of a computer terminal (or electronic device) for realizing the data detection method according to the embodiments of the present application;

[0022] Figure 2 is a flowchart of a data detection method according to the embodiments of the present application;

[0023] Figure 3 is a schematic diagram of trajectory playback data according to the embodiments of the present application;

[0024] Figure 4 is a structural diagram of a data detection device according to an embodiment of the application;

[0025] Figure 5 is a structural diagram of a data detection system according to an embodiment of the application. DETAILED DESCRIPTION

[0026] In order to enable persons skilled in the art to better understand the scheme of the present application, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device that includes a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to the process, method, product, or device.

[0028] The data detection method embodiments provided by the embodiments of the present application can be executed in a mobile terminal, a computer terminal, or a similar computing device. Figure 1 A hardware structural block diagram of a computer terminal (or electronic device) for implementing the data detection method is shown. As shown in Figure 1 The computer terminal 10 (or electronic device 10) can include one or more processors (the processor can include but is not limited to a microprocessor MCU or a programmable logic device FPGA processing device), a memory 104 for storing data, and a transmission module 106 for communication functions. In addition, it can also include a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which can be included as one of the ports of the I / O interface), a network interface, a power supply and / or a camera. Those skilled in the art can understand that Figure 1 The structure shown is only schematic, and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 can also include moreFigure 1 more or less components than those shown, or configured differently. Figure 1

[0029] It should be noted that the one or more processors and / or other data processing circuitry described above can be referred to herein generically as "data processing circuitry". The data processing circuitry can be embodied as software, hardware, firmware, or any combination thereof, in whole or in part. Moreover, the data processing circuitry can be a single standalone processing module, or incorporated in whole or in part within any other element of the computer terminal 10 (or electronic device). As referred to in embodiments of the present application, the data processing circuitry functions as a processor to control, for example, the selection of the variable resistance terminal path in connection with the interface.

[0030] The memory 104 can be used to store software programs of application software and modules, such as program instructions / data storage means corresponding to the data detection method in embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, i.e. implements the data detection method described above. The memory 104 can include a high-speed random access memory, and can further include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory 104 can further include a memory disposed remotely with respect to the processor, which can be connected to the computer terminal 10 through a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0031] The transmission module 106 is configured to receive or send data via a network. Examples of the network include, but are not limited to, a wireless network provided by a communication service provider of the computer terminal 10. In one example, the transmission module 106 includes a network interface controller (NIC) that can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission module 106 can be a radio frequency (RF) module configured to communicate with the Internet in a wireless manner.

[0032] The display can be, for example, a touch screen type liquid crystal display (LCD) that enables a user to interact with the user interface of the computer terminal 10 (or electronic device).

[0033] It should be noted that in some alternative embodiments, the above-mentioned Figure 1 ​The illustrated computer device (or electronic device) can include hardware elements (including circuitry), software elements (including computer code stored on a computer readable medium), or a combination of both hardware and software elements. It should be noted that Figure 1 is merely one instance of a particular, concrete example, and is intended to show the types of components that can be present in the above-described computer device (or electronic device).

[0034] Under the above operating environment, the embodiment of the present application provides a data detection method embodiment. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown.

[0035] Figure 2 is a flowchart of a data detection method according to an embodiment of the present application, as shown in Figure 2 The method comprises the following steps:

[0036] Step S202, acquiring map rendering data, wherein the map rendering data is rendering data used when generating a map.

[0037] In the above step S202, two types of data are generated during the production of the map, namely process data such as a GeoJSON type file, and compiled map rendering data such as an OpenDRIVE type file. Since the computer can only recognize the compiled map rendering data, the embodiment of the present application detects the map rendering data by acquiring the map rendering data and detecting it.

[0038] Step S204, detecting the map rendering data by multiple detection methods to obtain multiple detection results, wherein the detection results are used to indicate at least the confidence of the road elements in the map under the map rendering data.

[0039] In the above step S204, when using multiple detection methods to detect the map rendering data, each detection method corresponds to one detection result, and multiple methods correspond to multiple detection results. The multiple detection methods in the embodiment of the present application can include but are not limited to three detection methods. In an optional embodiment, when the multiple detection methods are three detection methods, the multiple detection methods are briefly described and introduced as follows:

[0040] The first detection mode simulates vehicle driving by using lane-level navigation simulation, detects the high-definition map rendering data during driving using a road defect detection model, and detects mapping defects of the high-definition map. The second detection mode detects mapping defects of the high-definition map by simulating using a simulation tool and detecting simulation data using a simulation defect detection model. The third detection mode detects mapping defects of the high-definition map by recording data using a recording playback tool and comparing the recorded data with the high-definition map data. The specific detection processes of the above three detection modes will be described in detail below.

[0041] In step S206, if the target detection result corresponding to the data is inconsistent with the preset result, the adjustment parameters corresponding to the detection modes in the detection process are adjusted, wherein the target detection result is determined by at least one of the detection results.

[0042] In step S206, a single detection mode (such as the high-definition map navigation, simulation tool, and recording playback tool) cannot accurately determine whether the rendering defect is caused by defect data or the tool itself. Therefore, multiple detection modes are needed for identification and detection to prevent the tools used by the multiple detection modes from having abnormalities that cause abnormal detection results, thereby providing more accurate feedback on data quality issues. Therefore, the detection results of the multiple detection modes are normalized by normalization, and different weight values are set for different detection modes in the normalization process to obtain a target detection result, thereby further determining the mapping defects of the high-definition map and excluding detection results caused by the tools themselves. It should be noted that the weight values corresponding to different detection modes are the adjustment parameters described above.

[0043] In step S206, the preset result may be a detection result confirmed by a human being or a pre-set detection result. In an optional embodiment, the target detection result is confirmed by a human being, the data with errors is manually corrected again, the data with correct detection and correction is generated as a training set, and the model is iterated according to the adjustment parameters. The model refers to the detection model used in the multiple detection modes, such as the road defect detection model.

[0044] In step S208, the map rendering data is detected again according to the adjustment parameters.

[0045] In step S208, the accuracy of the detection model is improved by iterating the model used in the multiple detection modes according to the adjustment parameters, and better detection results can be obtained when the map rendering data is detected according to the adjusted adjustment parameters and the detection model.

[0046] In steps S202 to S208 above, a quantitative evaluation is performed based on the quality of the map rendering data. Multiple detection methods are used to detect the parts of the map rendering data that differ from reality, and finally, a quantitative evaluation result is generated.

[0047] In step S204 of the above data detection method, map rendering data is detected by multiple detection methods, specifically including the following steps: generating trajectory playback data based on map rendering data; performing path planning on the map based on trajectory playback data to obtain at least one target road path.

[0048] The path planning process described above is a preparatory step before inspecting the map rendering data. Specifically, the map rendering data produced by the map, such as OpenDRIVE, RMD, or OSM data, is first loaded into various inspection tools corresponding to different inspection methods. These tools can be high-precision map navigation, simulation platforms, and autonomous driving recording and playback modules, thereby generating trajectory playback data of the inspected data. Figure 3 The diagram shown illustrates the trajectory playback data. By analyzing the trajectory playback data, it is possible to observe whether nodes at different locations are connected, thereby enabling path planning on the map.

[0049] In the above steps, path planning on the map based on trajectory playback data specifically includes the following steps: selecting a root node from the trajectory playback data, wherein the root node is a node in the trajectory playback data that is connected to only one road line; determining a second road line that has a connection relationship with the first road line where the root node is located based on the root node, and determining a road path based on at least the first road line and the second road line; traversing the tree structure generated based on the road path, and determining at least one target road path corresponding to the tree structure.

[0050] In the embodiments of this application Figure 3 In the process, by parsing the GeoJSON data of the road lines, a node with only one road line connected to it is selected as the root node, for example... Figure 3 In the given list, A, H, G, and J are all root nodes. If multiple nodes meet the criteria for a root node, the first matching node is selected as the root node based on the first-in, first-out (FIFO) principle of storing content in the document. Figure 3 If node A is stored at the beginning of the document, then node A is loaded first according to the first-in-first-out principle, thus making A the root node.

[0051] If A is the root node, then in Figure 3The length of the road line AB adjacent to A is set to 1, AB (i.e., a) is the first road line where the root node is located, and any path connected with AB is the second road line, which can be a direct connection or an indirect connection. In the embodiment of the present application, the C node and the F node in the road line BC (i.e., b) and BF (i.e., f) have a length of 2 to the root node, and so on until no path adjacent to AB can be found, and finally a tree structure (T1) is generated.

[0052] From the road lines that have not been selected, the tree structure construction is performed again (T2,…,Tn) until all the road lines stored in the file are included in the tree T, and T includes (T1,T2,…,Tn). For example, in Figure 3 In the embodiment, if abcde has been selected in the road line, the next selection of the root will be preferentially selected from fghi, and after the road line f is selected, h is preferentially selected, and if there is no h, b and a will be selected. If A is used as the root node, the first generated path is A-B-C-D-E-H, and then the path G-F-B-I-J is generated. Figure 3 In the embodiment, if abcde has been selected in the road line, the next selection of the root will be preferentially selected from fghi, and after the road line f is selected, h is preferentially selected, and if there is no h, b and a will be selected. If A is used as the root node, the first generated path is A-B-C-D-E-H, and then the path G-F-B-I-J is generated.

[0053] The target road path is selected from the tree structure T, for example, all the paths in the tree T can be traversed through a depth-first traversal algorithm, and finally m target road paths (L1,L2,…,Lm) are generated.

[0054] In step S204 of the above data detection method, in the case of the first detection mode, the map rendering data is detected by multiple detection modes, specifically including the following steps: loading the target road path and the map rendering data into a detection tool corresponding to the first detection mode, wherein the first detection mode is a mode of simulating navigation using a high-precision map; controlling the vehicle to travel in the map rendered by the map rendering data according to the target road path, and generating a first recorded video; dividing the first recorded video into multiple first pictures according to the division principle; inputting the multiple first pictures into a first detection model for identification to obtain a first detection result, wherein the first detection model is used for detecting road defects, and the first detection result at least includes the confidence of the classification to which each road element in the map belongs. In the embodiment of the present application, the classification includes redundancy, missing, correct, geometric error, elevation error, etc.

[0055] In the embodiment of the present application, the first detection method is to import the map rendering data (such as OpenDRIVE data) into the high-definition map navigation application (i.e. the detection tool corresponding to the first detection method), and output the video of the simulation running after the high-definition map navigation path planning. Specifically, by taking the loaded target road path as the navigation path, the process of the vehicle driving in the map rendered by the map rendering data is simulated, and the entire driving process is recorded to obtain a first recorded video. After recording is completed, the first recorded video is cut into frames according to the cutting principle, and the result of cutting is input into the road defect detection model (i.e. the first detection model M1) for identification, so as to detect the effect of the navigation road defect detection model (M1) in navigation.

[0056] In the high-definition map navigation simulator, the simulated vehicle will navigate according to the starting point, passing point and end point set by the target road path. During the simulation process, video recording is performed to obtain a first recorded video. After recording is completed, the video is automatically divided into multiple frames of pictures, i.e. multiple first pictures. In the embodiment of the present application, the cutting principle is as follows: Wherein fps is the number of frames per second, s is the single-frame visible distance (meters), v is the vehicle speed (meters / second), and a is the scale. Assuming that the screen range is 20 cm and the scale is 1:250, and the actual vehicle speed is 25 m / s, then That is, cut a frame every two seconds.

[0057] The navigation road defect detection model is trained based on Yolov5. The training process of the model is based on a large amount of abnormal data caused by mapping defects. The training process is as follows:

[0058] First, the display problems caused by map data defects after rendering the high-definition map are found by artificial means and are classified and labeled. The specific labeling content is shown in the following table:

[0059] Roadway or aerial road element Label class Arrow C (Correct) / E (Error) / R (Redundant) / L (Missing) Lane C (Correct) / E (Error) / R (Redundant) / L (Missing) AD_LaneDivider C (Correct) / E (Error) / R (Redundant) / L (Missing) … …

[0060] The labeled file and the original picture are input into the YoloV5 model for training. Through the YoloV5 detection model, the high-definition map navigation rendering defects can be detected. The detected results (confidence, i.e. the first detection result), coordinates and classification are input into the result normalization processing module.

[0061] In step S204 of the above data detection method, when the detection mode is the second detection mode, the map rendering data is detected by multiple detection modes, specifically including the following steps: loading the target road path and the map rendering data into a detection tool corresponding to the second detection mode, wherein the second detection mode is a mode of simulating vehicle driving; simulating the vehicle driving in the map rendered by the map rendering data according to the target road path to generate a second recorded video; dividing the second recorded video into multiple second pictures according to a simulated segmentation principle; inputting the multiple second pictures into a second detection model for recognition to obtain a second detection result, wherein the second detection model is used to simulate the first detection model to detect road defects, and the second detection result includes a detection result of the internal logic of the road element, and the internal logic of the road element is determined based on the confidence of the classification to which each road element in the map belongs and the driving state of the vehicle.

[0062] In the embodiments of the present application, the map rendering data and the target road path are loaded into an automatic driving simulation simulation platform (i.e., the detection tool corresponding to the second detection mode described above) to obtain a video. Specifically, the vehicle is simulated to drive in the map through simulation, and the driving is recorded to obtain a second recorded video. Through the rules of the simulation platform and automatic detection of the recording process, a simulation defect detection model (i.e., the second detection model M2 described above) is used to detect the performance effect of the vehicle in the automatic driving simulation simulation platform through Yolov5 to obtain a second detection result.

[0063] In the above simulation platform, the vehicle driving state or running state is mainly checked by a rule module and a detection module. The rule module is mainly used to detect the vehicle state, such as whether the vehicle is on the road, whether the vehicle collides, whether the vehicle overturns, speed limit abnormalities, road traffic conditions, etc. The detection module is mainly used to detect abnormal behaviors of the vehicle, such as vehicle jolting, vehicle reverse behavior, road crack layer, lane line loss, lane line error, curb rendering error, road barrier error, etc.

[0064] In the above data detection method, the method further includes the following steps: obtaining a first detection mode and a second detection mode in multiple detection modes; determining a first detection result and a first weight value corresponding to the first detection mode, and a second detection result and a second weight value corresponding to the second detection mode; when the absolute value of the difference between the first weight value and the second weight value is less than a first preset threshold, comparing the first detection result and the second detection result; when the absolute value of the difference between the confidence in the first detection result and the confidence in the second detection result is greater than a second preset threshold, determining that the difference between the first detection result and the second detection result is caused by the difference in the coding method used by the first detection mode and the second detection mode.

[0065] In the embodiment of the present application, when the first detection method is the simulation navigation method using the high-precision map, the second detection method is the method of simulating vehicle driving through the simulation platform, the map rendering data used by the first detection method is OSM format data, and the map rendering data used by the second detection method is OpenDRIVE format data, the weight values and detection results corresponding to the two detection methods are first obtained, and the weight value is the second parameter λ for normalization processing in the following text i If the absolute value of the difference between the first weight value corresponding to the first detection method and the second weight value corresponding to the second detection method is less than the first preset threshold value, it is considered that the first weight value and the second weight value are similar at this time. In this case, the confidence in the first detection result corresponding to the first detection method and the confidence in the second detection result corresponding to the second detection method are obtained. If the absolute value of the difference between the confidence in the first detection result and the confidence in the second detection result is greater than the second preset threshold value, it is considered that the detection ability of the first detection method and the second detection method at this time is greatly different. Then, the difference between the first detection result and the second detection result is the difference caused by the compilation method used by the first detection method and the second detection method.

[0066] In step S204 of the above data detection method, when the detection method is the third detection method, the map rendering data is detected by a plurality of detection methods, specifically including the following steps: obtaining the first road information collected by the vehicle during driving; comparing the first road information and the second road information to obtain a comparison result, wherein the second road information is the road information in the map rendered by the map rendering data.

[0067] In the embodiment of the present application, the third detection method is to record and play back the road information data (i.e. the first road information) collected by the vehicle during driving, and to compare the road information (i.e. the second road information) of the rendered map with the recorded road information for consistency, thereby obtaining the consistency. Specifically, the data recorded by the vehicle-side perception device (i.e. the first road information) is sent into the world model for processing, and then compared with the second road information in the high-precision map data. The output result includes: ground elements or road elements (such as lane lines, arrows, etc.) related coordinates and attributes, lane speed limits and other related information. By comparing the output information with the process data for consistency, the map data with poor consistency is sent into the normalization module for normalization processing.

[0068] In the step, the first road information is compared with the second road information, specifically including the following steps: obtaining a first coordinate in the first road information and a second coordinate in the second road information; determining a comparison result according to the first coordinate and the second coordinate, wherein the confidence of any road element in the map belonging to different categories is determined according to a condition met by the comparison result.

[0069] In the embodiment of the application, the second coordinate is (x1, y1), that is, the position of the corresponding point in the high-definition map, the first coordinate is (x2, y2), that is, the position of the corresponding point in the perception map, and β is a tolerance threshold, that is, the first parameter. The tolerance threshold can be understood as a distance threshold between the first coordinate and the second coordinate. The similarity between the first coordinate and the second coordinate (that is, the comparison result) is calculated in the following manner:

[0070]

[0071] If |s|≥|β|, it is considered that no corresponding data is matched, and the confidence of matching f(x) = 1, otherwise, it is considered that corresponding data can be matched, and the next confidence calculation is performed.

[0072] In the step, the confidence of any road element in the map belonging to different categories is determined, specifically including the following cases: when the comparison result is greater than or equal to the first parameter, the confidence is determined to be 0, wherein the first parameter is used to represent the matching degree of the data in the first road information and the data in the second road information; when the comparison result is 0, the confidence is determined to be 1; and when the comparison result is less than the first parameter, the confidence is determined according to the comparison result and the first parameter.

[0073] In the embodiment of the application, the condition met by the comparison result can be expressed as:

[0074]

[0075] In the process of determining the confidence of any road element in the map belonging to different categories, f(x) is the confidence of any road element in the map in a certain category.

[0076] In the step, when the detection mode is the third detection mode, the map rendering data is detected by multiple detection modes, specifically including the following steps: determining a confidence threshold corresponding to each road element in the map, wherein the confidence threshold is determined by at least the minimum value of the confidence of the category to which the point constituting the road element belongs; and determining data with a confidence less than the confidence threshold as the third detection result.

[0077] In the embodiment of the application, the calculation formula of the confidence threshold is:

[0078] f(θ) = min(f(xn ))

[0079] Wherein, n is the number of points of a road element, f(θ) is the confidence threshold of the road element, and when the calculated confidence is less than the confidence threshold, it is considered that the consistency is poor, and the data under this condition is taken as the third detection result.

[0080] In step S206 of the above data detection method, the adjustment parameters corresponding to the multiple detection methods in the detection process are adjusted, specifically including the following steps: normalizing the multiple detection results to obtain a target detection result, wherein the target detection result is determined by at least one of the multiple detection results (i.e. the first detection result, the second detection result and the third detection result) and a second parameter, the second parameter being used to represent the weight values corresponding to the multiple detection results respectively; in the case that the target detection result is greater than a target threshold, it is determined that the data corresponding to the target detection result is incorrect; comparing the target detection result with a preset result; in the case that the target detection result and the preset result are inconsistent, adjusting the first parameter and the second parameter.

[0081] In the embodiments of the present application, when the data is detected by using three detection methods, the target detection result is obtained by performing consistency processing on the results corresponding to the above three detection methods, so as to realize quantitative evaluation of the quality of the rendered map data. The consistency processing process can be realized by a consistency module, for example. The consistency processing module receives the detection results corresponding to the above three detection methods, and performs problem classification processing according to the pre-configured weight (i.e. the second parameter) corresponding to each detection method. The specific formula is as follows:

[0082] C(θ)=max(∑λ i *f i (θ))

[0083] Wherein, θ is a road element in the map, such as an arrow, a road line, etc., i is a detection method, λ i is a weight value set by a person, which is a weight value corresponding to the detection result (or confidence) of the detection method, and can also be understood as a weight value corresponding to the detection method, i.e. the second parameter, f i is the confidence (or detection result) obtained by the detection method corresponding to the category, and C is the confidence after normalization, i.e. the target detection result.

[0084] If the threshold value of C(θ) is greater than the set threshold value C ′ (θ) (i.e. the target threshold value, which can be set according to the actual situation), it is considered that a problem occurs at this position. The specific calculation method is as follows:

[0085]

[0086] Wherein, p is the detection accuracy (or precision) of the road element in the map, and r is the detection recall rate of the road element.

[0087] The above data detection method can be applied to a data detection system, such as a quality inspection system. If the target detection result is greater than the target threshold, it is considered that a problem occurs at the position, and the problematic data is compared with the preset result (such as being sent to a manual quality inspection link for confirmation). If the preset result (such as the manual confirmation result) is the same as the target detection result obtained by the system detection, the result is fed back to the manual maintenance node for re-maintenance. If the results are inconsistent, a label record is made, and the next training of the M1 and M2 models and the adjustment of the first parameter β and the second parameter λ are performed. i The data is used as a training set to train the model to improve the detection effect.

[0088] The data detection method provided by the embodiments of the present application acquires map rendering data, wherein the map rendering data is rendering data used when a map is generated. The map rendering data is detected by multiple detection methods respectively to obtain multiple detection results, wherein the detection result is at least used to indicate the confidence of a road element in a map under the map rendering data. In the case that the data corresponding to the target detection result is inconsistent with the preset result, the adjustment parameter corresponding to the multiple detection methods in the detection process is adjusted, wherein the target detection result is determined by at least one of the multiple detection results. The map rendering data is re-detected according to the adjustment parameter, so as to achieve the purpose of detecting the map rendering data by multiple methods, thereby realizing the technical effect of improving the detection efficiency of the map rendering data, and further solving the technical problem that the quality of the high-precision map is checked by manual work in the prior art, resulting in low efficiency and large workload.

[0089] Figure 4 is a structural diagram of a data detection device according to an embodiment of the present application, as shown in Figure 4 The device comprises:

[0090] The acquisition module 402 is configured to acquire map rendering data, wherein the map rendering data is rendering data used when a map is generated.

[0091] The first detection module 404 is configured to detect the map rendering data by multiple detection methods respectively to obtain multiple detection results, wherein the detection result is at least used to indicate the confidence of a road element in a map under the map rendering data.

[0092] The adjustment module 406 is configured to adjust the adjustment parameter corresponding to the multiple detection methods in the detection process in the case that the data corresponding to the target detection result is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results.

[0093] The second detection module 408 is configured to re-detect the map rendering data according to the adjustment parameter.

[0094] In the first detection module of the data detection device, the map rendering data is detected by multiple detection manners, and the detection manner includes the following process: trajectory playback data is generated according to the map rendering data; and path planning is performed on the map according to the trajectory playback data to obtain at least one target road path.

[0095] In the first detection module of the data detection device, the path planning is performed on the map according to the trajectory playback data, and the path planning includes the following process: a root node is selected from the trajectory playback data, wherein the root node is a node connected to only one road line in the trajectory playback data; a second road line having a connection relationship with a first road line where the root node is located is determined according to the root node; and a road path is determined according to at least the first road line and the second road line; and a tree structure generated according to the road path is traversed, and at least one target road path corresponding to the tree structure is determined.

[0096] In the first detection module of the data detection device, in the case that the detection manner is the first detection manner, the map rendering data is detected by multiple detection manners, and the detection manner includes the following process: the target road path and the map rendering data are loaded into a detection tool corresponding to the first detection manner, wherein the first detection manner is a manner of simulating navigation using a high-definition map; a vehicle is controlled to travel in a map rendered by the map rendering data according to the target road path to generate a first recorded video; the first recorded video is divided into multiple first pictures according to a division principle; and the multiple first pictures are input into a first detection model for identification to obtain a first detection result, wherein the first detection model is used for detecting road defects, and the first detection result at least includes a confidence degree of a classification to which each road element in the map belongs.

[0097] In the first detection module of the data detection device, in the case that the detection manner is the second detection manner, the map rendering data is detected by multiple detection manners, and the detection manner includes the following process: the target road path and the map rendering data are loaded into a detection tool corresponding to the second detection manner, wherein the second detection manner is a manner of simulating vehicle travel; a simulated vehicle is controlled to travel in a map rendered by the map rendering data according to the target road path to generate a second recorded video; the second recorded video is divided into multiple second pictures according to a simulated division principle; and the multiple second pictures are input into a second detection model for identification to obtain a second detection result, wherein the second detection model is used for simulating the first detection model to detect road defects, and the second detection result includes a detection result of an internal logic of a road element, and the internal logic of the road element is determined based on a confidence degree of a classification to which each road element in the map belongs and a driving state of the vehicle.

[0098] In the first detection module of the data detection device, the module is further configured to obtain a first detection manner and a second detection manner in a plurality of detection manners; determine a first detection result and a first weight value corresponding to the first detection manner, and a second detection result and a second weight value corresponding to the second detection manner; in a case where an absolute value of a difference between the first weight value and the second weight value is less than a first preset threshold, compare the first detection result and the second detection result; in a case where an absolute value of a difference between a confidence in the first detection result and a confidence in the second detection result is greater than a second preset threshold, determine that a difference between the first detection result and the second detection result is a difference caused by a compiling method used by the first detection manner and the second detection manner.

[0099] In the first detection module of the data detection device, in a case where the detection manner is a third detection manner, the map rendering data is detected by the plurality of detection manners respectively, and specifically includes the following process: obtaining first road information collected by the vehicle in a driving process; obtaining a comparison result by comparing the first road information and second road information, wherein the second road information is road information in a map rendered by the map rendering data.

[0100] In the first detection module of the data detection device, the comparison between the first road information and the second road information specifically includes the following process: obtaining a first coordinate in the first road information and a second coordinate in the second road information; determining the comparison result according to the first coordinate and the second coordinate, wherein the confidence of any road element in the map belonging to different categories is determined according to a condition satisfied by the comparison result.

[0101] In the first detection module of the data detection device, the confidence of any road element in the map belonging to different categories is determined, and specifically includes the following process: when the comparison result is greater than or equal to a first parameter, the confidence is determined to be 0, wherein the first parameter is used to represent a matching degree of data in the first road information and data in the second road information; when the comparison result is 0, the confidence is determined to be 1; when the comparison result is less than the first parameter, the confidence is determined according to the comparison result and the first parameter.

[0102] In the first detection module of the data detection device, in a case where the detection manner is a third detection manner, the map rendering data is detected by the plurality of detection manners respectively, and specifically includes the following process: determining a confidence threshold corresponding to each road element in the map, wherein the confidence threshold is determined at least by a minimum value of the confidence of a category to which a point constituting the road element belongs; determining data when the confidence is less than the confidence threshold as the third detection result.

[0103] In the adjustment module in the data detection device, the adjustment parameters corresponding to the plurality of detection modes in the detection process are adjusted, and the adjustment process specifically includes the following processes: normalizing the plurality of detection results to obtain a target detection result, wherein the target detection result is determined by at least one of the plurality of detection results and a second parameter, and the second parameter is used to represent the weight values corresponding to the plurality of detection results respectively; in the case that the target detection result is greater than a target threshold, it is determined that the data corresponding to the target detection result is incorrect; comparing the target detection result with a preset result; in the case that the target detection result is inconsistent with the preset result, adjusting the first parameter and the second parameter.

[0104] It should be noted that, Figure 4 The data detection device shown is used to execute Figure 2 The data detection method shown, so the related explanations in the above data detection method also apply to the data detection device, which will not be repeated here.

[0105] Figure 5 The structure diagram of a data detection system according to an embodiment of the present application, the data detection system 500 includes a first detection unit 501, a second detection unit 502, a third detection unit 503, a processing unit 504 and a feedback unit 505, the first detection unit uses a first detection method to detect data, the second detection unit uses a second detection method to detect data, the third detection unit uses a third detection method to detect data, the detection results obtained by the first detection method, the second detection method and the third detection method are normalized by the processing unit, and the problem data is determined therefrom, and the feedback unit feeds back the problem data to the user.

[0106] It should be noted that, Figure 5 The data detection system shown is used to execute Figure 2 The data detection method shown, so the related explanations in the above data detection method also apply to the data detection system, which will not be repeated here.

[0107] The non-volatile storage medium provided by the embodiment of the present application includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the following data detection method by running the computer program: obtaining map rendering data, wherein the map rendering data is rendering data used when generating a map; detecting the map rendering data by a plurality of detection methods respectively to obtain a plurality of detection results, wherein the detection result is used to indicate at least the confidence of the road element in the map under the map rendering data; in the case that the data corresponding to the target detection result is inconsistent with the preset result, adjusting the adjustment parameters corresponding to the plurality of detection methods in the detection process, wherein the target detection result is determined by at least one of the plurality of detection results; and re-detecting the map rendering data according to the adjustment parameters.

[0108] The above-mentioned sequence numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments.

[0109] In the above-mentioned embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0110] In the several embodiments provided by the present application, it should be understood that the disclosed technology can be implemented in other ways. Of course, the unit embodiment described above is only illustrative, and for example, the division of units can be a logical function division, and there can be another division manner in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, unit or module, and can be electrical or other forms.

[0111] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0112] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The above integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0113] The integrated unit, if realized in the form of software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and various program codes that can be stored in the medium.

[0114] The above merely describes the preferred embodiments of the present application, and it should be pointed out that, for those skilled in the art, some improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A data detection method, characterized in that, include: Obtain map rendering data, wherein the map rendering data is the rendering data used when generating the map; The map rendering data is detected by multiple detection methods to obtain multiple detection results, wherein the detection results are used at least to indicate the confidence level of road features in the map under the map rendering data; If the data corresponding to the target detection result is inconsistent with the preset result, the adjustment parameters corresponding to the multiple detection methods during the detection process are adjusted, wherein the target detection result is determined by at least one of the multiple detection results; The map rendering data was re-detected based on the adjusted parameters; The map rendering data is detected using multiple detection methods, including: generating trajectory playback data based on the map rendering data; and performing path planning on the map based on the trajectory playback data to obtain at least one target road path. When the detection method is the second detection method, the map rendering data is detected by multiple detection methods, including: loading the target road path and the map rendering data into the detection tool corresponding to the second detection method; simulating a vehicle driving on the map rendered by the map rendering data according to the target road path to generate a second recorded video; simulating a segmentation principle to segment the second recorded video into multiple frames of second images; inputting the multiple frames of second images into a second detection model for recognition to obtain a second detection result, wherein the second detection model is used to simulate the first detection model to detect road defects, and the second detection result includes the detection result of the internal logic of the road elements, wherein the internal logic of the road elements is determined based on the confidence level of the category to which each road element belongs in the map and the driving state of the vehicle; When the detection method is the first detection method, the map rendering data is detected by multiple detection methods, including: loading the target road path and the map rendering data into the detection tool corresponding to the first detection method; controlling the vehicle to drive on the map rendered by the map rendering data according to the target road path to generate a first recorded video; dividing the first recorded video into multiple first images according to the segmentation principle; inputting the multiple first images into a first detection model for recognition to obtain a first detection result, wherein the first detection model is used to detect road defects, and the first detection result includes at least the confidence level of the category to which each road element in the map belongs.

2. The method according to claim 1, characterized in that, Based on the trajectory playback data, path planning is performed on the map, including: Select a root node from the trajectory playback data, wherein the root node is a node in the trajectory playback data that is connected to only one road line; Based on the root node, determine a second road line that is connected to the first road line where the root node is located, and determine the road path based at least on the first road line and the second road line; Traverse the tree structure generated based on the road path and determine at least one target road path corresponding to the tree structure.

3. The method according to claim 1, characterized in that, The method further includes: Obtain the first detection method and the second detection method from the multiple detection methods; Determine the first detection result and the first weight value corresponding to the first detection method, and the second detection result and the second weight value corresponding to the second detection method; If the absolute value of the difference between the first weight value and the second weight value is less than the first preset threshold, compare the first detection result and the second detection result. If the absolute value of the difference between the confidence level in the first detection result and the confidence level in the second detection result is greater than the second preset threshold, the difference between the first detection result and the second detection result is determined to be a difference caused by the compilation method used by the first detection method and the second detection method.

4. The method according to claim 1, characterized in that, When the detection method is the third detection method, the map rendering data is detected using multiple detection methods, including: Acquire the first road information collected during the vehicle's journey; By comparing the first road information and the second road information, a comparison result is obtained, wherein the second road information is the road information in the map rendered by the map rendering data.

5. The method according to claim 4, characterized in that, The comparison between the first road information and the second road information includes: Obtain the first coordinate from the first road information and the second coordinate from the second road information; Based on the first coordinate and the second coordinate, the comparison result is determined, wherein, based on the conditions satisfied by the comparison result, the confidence level of any road element in the map to which it belongs is determined.

6. The method according to claim 5, characterized in that, Determine the confidence level of different categories to which any road feature in the map belongs, including: When the comparison result is greater than or equal to the first parameter, the confidence level is determined to be 0, wherein the first parameter is used to represent the degree of matching between the data in the first road information and the data in the second road information; When the comparison result is 0, the confidence level is determined to be 1; When the comparison result is less than the first parameter, the confidence level is determined based on the comparison result and the first parameter.

7. The method according to claim 6, characterized in that, When the detection method is the third detection method, the map rendering data is detected using multiple detection methods, including: Determine the confidence threshold for each road feature in the map, wherein the confidence threshold is determined at least by the minimum confidence value of the category to which the points constituting the road feature belong; Data with a confidence level less than the confidence threshold are identified as the third detection result.

8. The method according to claim 6, characterized in that, Adjusting the adjustment parameters corresponding to the various detection methods during the detection process, including: The multiple detection results are normalized to obtain the target detection result, wherein the target detection result is determined by at least one of the multiple detection results and a second parameter, the second parameter being used to represent the weight value corresponding to each of the multiple detection results; If the target detection result is greater than the target threshold, it is determined that the data corresponding to the target detection result is incorrect; Compare the target detection result with the preset result; If the target detection result and the preset result are inconsistent, adjust the first parameter and the second parameter.

9. A data detection device, characterized in that, include: The acquisition module is used to acquire map rendering data, wherein the map rendering data is the rendering data used when generating the map; The first detection module is used to detect the map rendering data using multiple detection methods to obtain multiple detection results, wherein the detection results are at least used to indicate the confidence level of road features in the map under the map rendering data. Detecting the map rendering data using multiple detection methods includes: generating trajectory playback data based on the map rendering data; performing path planning on the map based on the trajectory playback data to obtain at least one target road path; when the detection method is the second detection method, detecting the map rendering data using multiple detection methods includes: loading the target road path and the map rendering data into a detection tool corresponding to the second detection method; simulating a vehicle driving on the map rendered by the map rendering data according to the target road path to generate a second recorded video; simulating a segmentation principle to segment the second recorded video into multiple frames of second images; inputting the multiple frames of second images into a second detection model for recognition to obtain a second detection result. The second detection model is used to simulate the first detection model in detecting road defects. The second detection result includes the detection result of the internal logic of the road element. The internal logic of the road element is determined based on the confidence level of the category to which each road element belongs in the map and the driving state of the vehicle. When the detection method is the first detection method, the map rendering data is detected by multiple detection methods, including: loading the target road path and the map rendering data into the detection tool corresponding to the first detection method; controlling the vehicle to drive in the map rendered by the map rendering data according to the target road path to generate a first recorded video; dividing the first recorded video into multiple first images according to the segmentation principle; inputting the multiple first images into the first detection model for recognition to obtain a first detection result. The first detection model is used to detect road defects, and the first detection result includes at least the confidence level of the category to which each road element belongs in the map. An adjustment module is used to adjust the adjustment parameters corresponding to the multiple detection methods during the detection process when the data corresponding to the target detection result is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results; The second detection module is used to re-detect the map rendering data based on the adjusted parameters.

10. An electronic device, characterized in that, include: Memory, used to store program instructions; A processor, connected to the memory, is configured to execute program instructions to perform the following functions: acquiring map rendering data, wherein the map rendering data is the rendering data used when generating the map; detecting the map rendering data using multiple detection methods to obtain multiple detection results, wherein the detection results are at least used to indicate the confidence level of road features in the map under the map rendering data; adjusting adjustment parameters corresponding to the multiple detection methods during the detection process when the data corresponding to the target detection result is inconsistent with the preset result, wherein the target detection result is determined by at least one of the multiple detection results; re-detecting the map rendering data according to the adjustment parameters; detecting the map rendering data using multiple detection methods, including: generating trajectory playback data based on the map rendering data; performing path planning on the map based on the trajectory playback data to obtain at least one target road path; when the detection method is a second detection method, detecting the map rendering data using multiple detection methods, including: loading the target road path and the map rendering data into a detection tool corresponding to the second detection method; simulating a vehicle traveling along the target road path through the map rendering data. The vehicle drives on the rendered map to generate a second recorded video. A simulated segmentation principle is used to segment the second recorded video into multiple frames of second images. These multiple frames of second images are then input into a second detection model for identification, yielding a second detection result. The second detection model simulates the first detection model's detection of road defects. The second detection result includes the detection result of the internal logic of the road elements, which is determined based on the confidence level of the category to which each road element belongs in the map and the vehicle's driving state. When the detection method is the first detection method, the map rendering data is detected using multiple detection methods, including: loading the target road path and the map rendering data into the detection tool corresponding to the first detection method; controlling the vehicle to drive on the map rendered by the map rendering data according to the target road path to generate a first recorded video; segmenting the first recorded video into multiple frames of first images according to a segmentation principle; and inputting these multiple frames of first images into a first detection model for identification, yielding a first detection result. The first detection model is used to detect road defects, and the first detection result includes at least the confidence level of the category to which each road element belongs in the map.

11. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored computer program, wherein the device containing the non-volatile storage medium executes the data detection method according to any one of claims 1 to 8 by running the computer program.

Images