Map element attribute information updating method and device, and storage medium

Abstract

The method, device and storage medium for updating map element attribute information provided by the present disclosure relate to map technology and crowd-sourcing updating technology, and include: obtaining sensor data, determining accuracy information of the sensor data according to the sensor data; obtaining attribute distribution information of a map element according to description information corresponding to the map element included in the sensor data in a preset high-definition map; and updating the attribute distribution information of the map element according to the accuracy information and an attribute recognition result included in the sensor data, to obtain current attribute distribution information of the map element, the attribute distribution information of the map element being used to determine a real attribute of the map element. The scheme of the present disclosure can update the attributes of various elements in the high-definition map in a timely manner without consuming a large amount of storage resources, and can improve the accuracy of the attribute distribution information, without the need to store a large amount of sensor data.

Description

Technical Field

[0001] This disclosure relates to map technology and crowdsourced update technology, and in particular to a method, device, and storage medium for updating map feature attribute information. Background Technology

[0002] High-precision maps refer to maps with high accuracy and detailed definition. In the field of driver assistance systems, driver assistance systems need to specify strategies based on high-precision maps. For example, driver assistance systems need to know the road conditions based on high-precision maps in order to determine the route.

[0003] In real-world driving environments, road conditions change frequently, such as road expansion and changes in solid and dashed lines. To ensure that the information in high-precision maps matches the actual driving environment, crowdsourced data is typically used to update the high-precision maps.

[0004] A crucial aspect of crowdsourced updates is determining element attributes. Currently, the primary method is voting. For example, data collected by all sensors over a period of time is statistically analyzed, and the attribute that appears most frequently is used as the attribute value for that element. However, this approach treats data from different sensors in the same way. When faced with data sources of poor quality, especially those with significant systematic errors, these sources can negatively influence the voting results, easily leading to incorrect cognitive outcomes. Summary of the Invention

[0005] This disclosure provides a method, device, and storage medium for updating map feature attribute information, in order to solve the problem that when conducting crowdsourced updates in the prior art, when faced with data sources of poor quality, especially those with obvious systematic errors, these data sources can negatively interfere with the voting results and easily lead to incorrect cognitive results.

[0006] The first aspect of this disclosure is to provide a method for updating map feature attribute information, including:

[0007] Acquire sensor data, and determine the accuracy information of the sensor data based on the sensor data;

[0008] Based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map, the attribute distribution information of the map elements is obtained; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute;

[0009] Based on the accuracy information and the attribute recognition results included in the sensor data, the attribute distribution information of the map element is updated to obtain the current attribute distribution information of the map element. The attribute distribution information of the map element is used to determine the true attributes of the map element.

[0010] Another aspect of this disclosure is to provide a device for updating map feature attribute information, including:

[0011] The data acquisition module is used to acquire sensor data and determine the accuracy information of the sensor data based on the sensor data.

[0012] The distribution acquisition module is used to acquire the attribute distribution information of the map elements based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute;

[0013] An update module is used to update the attribute distribution information of the map element based on the accuracy information and the attribute recognition results included in the sensor data, to obtain the current attribute distribution information of the map element, wherein the attribute distribution information of the map element is used to determine the true attribute of the map element. Another aspect of this disclosure is to provide a map update device, comprising:

[0014] Memory;

[0015] Processor; and

[0016] Computer programs;

[0017] The computer program is stored in the memory and configured to be executed by the processor to implement the method for updating map feature attribute information as described in the first aspect above.

[0018] Another aspect of this disclosure is to provide a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method for updating map feature attribute information as described in the first aspect above.

[0019] Another aspect of this disclosure is to provide a computer program, including program code, which, when a computer runs the computer program, performs the method for updating map feature attribute information as described in the first aspect above.

[0020] The method, apparatus, and storage medium for updating element attribute information disclosed herein include: acquiring sensor data; determining the accuracy information of the sensor data based on the sensor data; acquiring the attribute distribution information of map elements based on the description information corresponding to map elements included in the sensor data in a preset high-precision map; the attribute distribution information includes the distribution information of map elements belonging to each possible attribute; updating the attribute distribution information of map elements based on the accuracy information and the attribute identification results included in the sensor data, thereby obtaining the current attribute distribution information of map elements, which is used to determine the true attributes of map elements. In the solution provided by this application, the attribute distribution of elements is recorded in the form of attribute distribution information, and the attribute distribution information can be updated based on the received sensor data and its accuracy information, thereby making the attribute distribution information of map elements increasingly accurate. Based on this attribute distribution information, the attribute recognition result of the element can be directly determined. Therefore, the solution of this application does not require storing a large amount of sensor data, and can update the attributes of each element in the high-precision map in a timely manner without consuming a large amount of storage resources, and can improve the accuracy of the attribute distribution information. Attached Figure Description

[0021] Figure 1 This is an example of an application scenario.

[0022] Figure 2 This is a schematic flowchart illustrating a method for determining map feature attribute information, as shown in an exemplary embodiment of this application.

[0023] Figure 3 A schematic flowchart illustrating a method for determining map feature attribute information, as shown in another exemplary embodiment of this application;

[0024] Figure 4 This is a schematic diagram illustrating the generation of initial attribute distribution information, as shown in an exemplary embodiment of this application.

[0025] Figure 5 This is a schematic diagram of the structure of a device for determining map feature attribute information, as shown in an exemplary embodiment of this application.

[0026] Figure 6 A structural diagram of a device for determining map feature attribute information, as shown in another exemplary embodiment of this application;

[0027] Figure 7 This is a structural diagram of a map updating device illustrated in an exemplary embodiment of this application. Detailed Implementation

[0028] To address the challenge of timely updates to high-definition maps, existing technologies employ crowdsourcing update methods. For example, a user's mobile device can be used as a basic sensing unit to acquire environmental data and report it to a backend server. The backend server can then update the high-definition map based on this data.

[0029] Sensors can be installed in mobile devices, such as vehicles, to collect data, which the devices can then report. Because this method can collect data from a wider range and ensure high data freshness, high-precision maps can be updated in a timely manner.

[0030] When updating high-precision maps based on crowdsourced data, the attributes of elements within the high-precision map can be updated. Elements in a high-precision map can be road elements, such as lane lines, as well as signs, guardrails, and roadside units. These elements have attributes; for example, lane lines can be categorized as dashed, solid, or special dashed lines. Similarly, signs can be categorized as speed limit signs or stop signs.

[0031] Figure 1 This is an example illustration of an application scenario.

[0032] like Figure 1 As shown, the sensors installed in each mobile device 11 can perceive the surrounding environment and obtain sensor data. The mobile device 11 can upload the sensor data to the server 12. The server 12 stores a large amount of sensor data 13. When it is necessary to determine the attribute of an element, the server 12 counts the sensor data corresponding to the element over a period of time, and selects the attribute result that appears most frequently from these sensor data, which is the target attribute of the element.

[0033] For example, if there are 100 sensor data points used to report the attributes of the same lane line, and 80% of these data points indicate that the lane line is a solid line, while 20% of these data points indicate that the lane line is a dashed line, then the attribute of this element in the high-precision map is directly determined to be a solid line.

[0034] In this implementation, element attributes need to be determined based on historical crowdsourced data, which requires a large amount of storage resources to store the crowdsourced data, resulting in wasted storage resources. At the same time, this voting method performs the same processing procedure on all sensor data without distinguishing the confidence level of each sensor data. As a result, data sources with poor quality, especially those with obvious systematic errors, can negatively interfere with the voting results, making it very easy to determine incorrect cognitive results.

[0035] To address the aforementioned technical problems, the method provided in this application utilizes attribute distribution information to record element attribute information. Upon receiving sensor data, the accuracy of the data can be determined, and the attribute distribution information can be updated based on the sensor data and its accuracy information. This method eliminates the need to store sensor data; storing only the attribute distribution information is sufficient to determine the element's attributes. Consequently, it allows for timely updates to element attributes without requiring substantial storage resources. Furthermore, this method considers the accuracy of the sensor data, resulting in more accurate results.

[0036] Figure 2 This is a schematic diagram illustrating a method for determining map feature attribute information, which is an exemplary embodiment of this application.

[0037] like Figure 2 As shown, the method for determining map feature attribute information provided in this application includes:

[0038] Step 201: Receive sensor data and determine the accuracy information of the sensor data based on the sensor data.

[0039] The method provided in this application can be executed by an electronic device with computing capabilities, such as a back-end server or an on-board unit in a vehicle.

[0040] Specifically, electronic devices are able to receive sensor data.

[0041] In one implementation, the sensor data can be sent by a mobile device, such as an onboard unit in a vehicle. Sensors mounted on the mobile device can perceive the surrounding environment, and the mobile device or sensors can obtain attribute recognition results for elements based on the perception results. The mobile device then sends these results to the electronic device executing this embodiment, for example, to a server or to other mobile devices.

[0042] In another implementation, the sensor data can be sent by a processing device that processes the raw sensed data. For example, sensors on a mobile device can sense the surrounding environment and collect raw sensed data, such as point cloud data or image data. The mobile device can send this raw sensed data to a processing device, which processes the raw sensed data to obtain sensor data. For example, the processing device can identify the raw sensed data and determine the attribute identification results of elements. The processing device can then send the sensor data to an electronic device for performing the method provided in this application.

[0043] Furthermore, electronic devices can determine their accuracy information based on sensor data. This accuracy information is used to characterize the accuracy of the sensor data.

[0044] In practical applications, some types of sensors acquire data that is more accurate, such as radar data. Therefore, in some implementations, the accuracy of sensor data can be determined based on the type of sensor acquiring the data. The accuracy of sensor data can also be determined based on other methods, such as the type of map elements included in the sensor data.

[0045] The sensor data includes attribute identification results for map elements, which are obtained based on the sensor's perception results. For example, the sensor might detect that a lane line is a solid line.

[0046] In practical applications, map elements include, but are not limited to, information such as lane lines, traffic signs, and fences in the vehicle's driving environment.

[0047] A map element can include multiple attributes. For example, lane lines have attributes including but not limited to solid / dashed line type, color, and arrow direction. Traffic signs have attributes including but not limited to semantic type, shape, and color.

[0048] Step 202: Based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map, obtain the attribute distribution information of the map elements; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute.

[0049] After receiving sensor data, the electronic device can determine which map element the sensor data represents, for example, the recognition result of lane line L1. In one embodiment, the mobile device can combine its own location, the relative position of the sensor and itself, and a preset high-precision map to determine which map element the sensor data represents. It can then encapsulate the information of that map element within the sensor data, allowing the electronic device to read the map element information included in the sensor data and thus identify the map element contained within the sensor data.

[0050] Specifically, electronic devices can obtain the attribute distribution information of map elements based on the descriptive information corresponding to map elements included in sensor data within a preset high-precision map. For example, the attribute distribution information of each map element can be stored in a preset high-precision map, and the attribute distribution information matching the current map element can be obtained based on the descriptive information corresponding to the current map element in the preset high-precision map.

[0051] The acquired attribute distribution information is the current attribute distribution information of the map element at the moment after receiving sensor data.

[0052] For example, if the sensor data received by the electronic device includes the attribute recognition results of map element A, then the electronic device can obtain the attribute distribution information V1 of A.

[0053] Specifically, the attribute distribution information includes the distribution information of map elements belonging to each possible attribute. For example, the probability value p1 of a map element having the first attribute, the probability value p2 of a map element having the second attribute, and so on. Furthermore, the attribute distribution information can record the statistical results of sensor data, such as N1 sensor data points indicating that a map element has the first attribute, and N2 sensor data points indicating that a map element has the second attribute.

[0054] The method provided in this application can record the attribute recognition results of map elements through attribute distribution information, eliminating the need to store sensor data and thus freeing up storage resources.

[0055] Furthermore, the attribute distribution information records the distribution of each attribute of an element. When it is necessary to determine the attribute recognition result of an element, it can be determined directly based on this attribute distribution information. For example, for a type of attribute of a map element, the attribute with the highest probability value can be taken as the attribute of that map element.

[0056] In practical applications, attribute recognition results are determined by electronic devices based on the attribute distribution information of elements. They represent the current attributes of map elements determined by electronic devices based on existing information.

[0057] For example, if an electronic device receives sensor data, including map element attribute identification results such as "lane line link 1 is a solid line", the electronic device can obtain the attribute distribution information of link 1. The attribute distribution information of link 1 records the attribute distribution of link 1, for example, the probability of link 1 being a dashed line is 80%, and the probability of link 1 being a solid line is 20%.

[0058] Step 203: Based on the accuracy information and the attribute recognition results included in the sensor data, update the attribute distribution information of the map elements to obtain the current attribute distribution information of the map elements. The attribute distribution information of the map elements is used to determine the true attributes of the map elements.

[0059] Electronic devices can use the attribute identification results and accuracy information of the received sensor data to update the attribute distribution information of map elements, thereby obtaining the current attribute distribution information of the map element.

[0060] For example, if the attribute distribution information of a map element includes probability values ​​corresponding to multiple possible attributes, and the accuracy information of the currently received sensor data is 0.5, then the probability value of the map element belonging to each possible attribute can be updated based on this accuracy information and the recognition results in the sensor data.

[0061] For example, if the sensor data indicates that lane line Link1 is a dashed line, and the current attribute distribution information shows that the probability of lane line Link1 being a dashed line and a solid line is 50% each, and assuming the electronic device determines the accuracy of this sensor data to be 0.2, then the probability values ​​of each possible attribute of lane line Link1 can be adjusted based on this 0.2. For instance, the probability value of Link1 being a dashed line can be adjusted to 51%. If the accuracy of the sensor data is 0.9, then the probability value of Link1 being a dashed line can be adjusted to 55%.

[0062] In this implementation, the accuracy information of sensor data can be used to update the attribute distribution information of map elements. This allows the impact of a single sensor data reading on the attribute distribution information of map elements to be determined based on the accuracy difference. It avoids the significant impact of sensor data with low accuracy on the attribute distribution information and also increases the impact of sensor data with high accuracy on the attribute distribution information. As a result, the attribute distribution information of map elements becomes more and more accurate during the update process.

[0063] In another implementation, the attribute distribution information records the statistical results of the sensor data. For example, if N1 sensor data indicate that the element is the first attribute and N2 sensor data indicate that the element is the second attribute, the value of N1 or N2 can be updated based on the received sensor data and its accuracy information. For example, if the sensor data includes a map element as the first attribute and the electronic device has an accuracy of 0.2, then 0.2 can be added to N1.

[0064] For example, if the attribute distribution information includes the probability values ​​of the map element belonging to each different attribute, then the Bayesian reliability model can be used to update the obtained attribute distribution information based on the attribute identification results included in the sensor data and the accuracy information of the sensor data, so as to obtain the current attribute distribution information.

[0065] In this implementation, after collecting new sensor data, the attribute distribution information of map elements can be updated based on this sensor data and its accuracy information. Furthermore, the attribute distribution information of map elements can characterize the attributes of those map elements, thus achieving timely updates to map element attributes. In addition, this method does not require storing large amounts of sensor data, enabling timely updates to the attributes of each map element in a high-precision map without consuming significant storage resources.

[0066] Furthermore, this implementation can utilize the accuracy information of sensor data to update the attribute distribution information of map elements. This allows for the determination of the impact of a single sensor data session on the attribute distribution information of map elements based on the accuracy differences. It avoids the significant impact of sensor data with low accuracy on the attribute distribution information and also enhances the impact of sensor data with high accuracy on the attribute distribution information. As a result, the attribute distribution information of map elements becomes increasingly accurate during the update process.

[0067] The method provided in this embodiment is used to update the attribute information of map features. The method is executed by a device equipped with the method provided in this embodiment, which is typically implemented in hardware and / or software.

[0068] The method for updating map element attribute information provided in this application includes: acquiring sensor data; determining the accuracy information of the sensor data based on the sensor data; acquiring the attribute distribution information of the map elements based on the description information corresponding to the map elements included in the sensor data in a preset high-precision map; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute; updating the attribute distribution information of the map elements based on the accuracy information and the attribute recognition results included in the sensor data, thereby obtaining the current attribute distribution information of the map elements, which is used to determine the true attributes of the map elements. In the method provided in this application, the attribute distribution of elements is recorded in the form of attribute distribution information, and the attribute distribution information can be updated based on the received sensor data and its accuracy information, thereby making the attribute distribution information of map elements increasingly accurate. Based on this attribute distribution information, the attribute recognition result of the element can be directly determined. Therefore, the solution of this application does not require storing a large amount of sensor data, and can update the attributes of each element in the high-precision map in a timely manner without consuming a large amount of storage resources, and can improve the accuracy of the attribute distribution information.

[0069] Figure 3 This is a schematic diagram illustrating a method for determining map feature attribute information, which is another exemplary embodiment of this application.

[0070] like Figure 3 As shown, the method for determining map feature attribute information provided in this application includes:

[0071] Step 301: Acquire sensor data.

[0072] Step 301 is similar to step 201, and will not be described again.

[0073] Step 302: Determine the accuracy information of the sensor data based on any of the following information from the sensor data: determine the accuracy information of the sensor data based on the type of sensor used to acquire the sensor data; determine the accuracy information of the sensor data based on the attribute recognition results of the map elements included in the sensor data; determine the accuracy information of the sensor data based on the observation conditions when the sensor acquires the sensor data.

[0074] The accuracy information is used to characterize the probability that the map element is identified as the attribute recognition result in the sensor data when the true attribute of the map element is each possible attribute.

[0075] The method provided in this application can obtain accuracy information corresponding to the sensor, which is used to characterize the confidence level of the sensor data.

[0076] Specifically, there are no restrictions on the execution order of steps 302 and 303.

[0077] Furthermore, different sensor data can include attribute recognition results of the same map elements. For example, when vehicle 1 travels to location p, it collects information about lane line L and generates first sensor data, which includes lane line L as a solid line. When vehicle 2 travels to location p, it also collects information about lane line L and generates second sensor data, which includes lane line L as a dashed line.

[0078] In practical applications, the accuracy of data from different sensors varies. Treating all sensor data equally may lead to inaccurate results. Therefore, in one embodiment provided in this application, the accuracy information of sensor data can be determined, and then the attribute distribution information of map elements can be updated using the sensor data based on this accuracy information.

[0079] Different types of sensors can have varying degrees of accuracy when observing the same element; for example, the accuracy of radar observations may differ from that of camera observations. Therefore, the accuracy of sensor data can be determined based on the type of sensor used to acquire the data.

[0080] Specifically, the accuracy of sensors of the same type varies when identifying elements with different attributes, and the accuracy of sensor observations also differs under different observation conditions. Therefore, the accuracy of sensor data can be determined based on the attribute identification results of map elements included in the sensor data, or based on the observation conditions under which the sensor acquired the data.

[0081] The accuracy information can specifically include: the probability that an element will be identified as an attribute included in the sensor data when its true attribute is any possible attribute. For example, if the sensor data includes lane lines that are solid lines, then the probability that the sensor will identify it as a solid line when its true attribute is solid, the probability that the sensor will identify it as a solid line when its true attribute is dashed, and the probability that the sensor will identify it as a solid line when its true attribute is a special dashed line can be obtained.

[0082] Step 303A: Determine the type information of the map elements based on the description information of the map elements in the preset high-precision map; generate the initial attribute distribution information of the map elements based on their type.

[0083] Step 303B: In the data of the preset high-precision map, obtain the attribute distribution information corresponding to the map elements based on the description information of the map elements.

[0084] Step 303 is implemented in a similar way to step 202.

[0085] In one optional implementation, if the attribute identification result of the map element included in the received sensor data is the first attribute identification result of that element received by the electronic device, then the acquired attribute distribution information of the map element can be the initial attribute distribution information of the map element. The initial attribute distribution information can include the initial probability that the map element is each possible attribute. For example, the probability that the map element is the first attribute is an initial value p1, and the probability that the element is the first attribute is an initial value p2.

[0086] In one possible scenario, when generating the initial attribute distribution information of a map element, this information can be generated based on the element's type information within the high-definition map. The type of the map element can be obtained from an existing high-definition map, and its initial attribute distribution information can then be generated accordingly. For example, the initial attribute distribution information can be generated based on the map element's type and prior knowledge. Alternatively, historical attribute values ​​of the map element can be obtained from historical high-definition maps, and the initial attribute distribution information can be generated based on these historical values.

[0087] Figure 4 This is a schematic diagram illustrating the generation of initial attribute distribution information, which is an exemplary embodiment of this application.

[0088] like Figure 4 As shown, for example, if the received sensor data includes the attribute recognition result of map element A, then the historical attribute value of map element A can be obtained from the historical high-precision map 41. For example, the obtained result includes the attribute of A being a solid line.

[0089] The electronic device can generate initial attribute distribution information 42 for map element A based on its historical attribute values. In this initial attribute distribution information 42, the historical attribute values ​​have the highest distribution probability. Since the attribute values ​​of map element A in historical high-precision maps are historical values, and the external environment remains unchanged, the attribute of map element A is highly likely to remain the same historical attribute value. Therefore, it is possible to generate initial attribute distribution information with the highest probability of historical attribute value distribution.

[0090] The initial attribute distribution information 42 includes multiple possible attributes of element A. These possible attributes belong to the same type of attribute, such as all belonging to the line type attribute or all belonging to the color type attribute.

[0091] The types of elements can include many kinds, such as lane line elements or traffic sign elements. In another implementation, the initial attribute distribution information corresponding to different element types can be preset. When needed, the initial attribute distribution information of the element can be obtained directly according to the element type.

[0092] For example, experts can determine the initial attribute distribution information of each type of element based on experience, or they can determine the initial attribute distribution information of each type of element based on prior knowledge.

[0093] In another implementation, it can be determined empirically which types of elements have easily changeable attributes. If the currently processed element is of an easily changeable attribute type, then an evenly distributed initial attribute distribution information can be generated. In this initial attribute distribution information, the probability values ​​of each attribute are evenly distributed.

[0094] In another possible scenario, if the attribute identification results of map elements included in the received sensor data are not the first attribute identification results of that element received by the electronic device, then the electronic device contains attribute distribution information for that map element. This attribute distribution information is obtained by the electronic device processing previously received sensor data. For example, if the current time is t, then the electronic device received attribute identification results for that element before time t and updated the attribute distribution information of that element based on the received attribute identification results.

[0095] For example, if t1 is a time before t, and at time t1 the electronic device first receives the attribute identification result Q of map element A, then the electronic device can generate initial attribute distribution information for map element A and update the initial attribute distribution information of A using the received attribute identification result Q to obtain the first attribute distribution information. At time t, after the electronic device receives the attribute identification result of element A again, it can obtain the first attribute distribution information of that element and update it.

[0096] Specifically, attribute distribution information corresponding to map elements can be obtained from the data of a preset high-precision map based on the descriptive information of the map elements. For example, the corresponding attribute distribution information can be obtained from the data of a preset high-precision map based on the descriptive information such as the identifier of the map element.

[0097] Step 304: Determine the accuracy corresponding to each possible attribute based on the first probability value of each possible attribute in the attribute distribution information and the second probability value of each possible attribute in the accuracy information of the sensor data.

[0098] The attribute distribution information includes a first probability value for each possible attribute of the map element; the accuracy information of the sensor data includes a second probability value for the map element being identified as an attribute recognition result in the sensor data when the true attribute of the map element is each possible attribute. Specifically, the accuracy corresponding to each possible attribute can be determined based on the probability values ​​corresponding to each possible attribute in the attribute distribution information of the element and the probability values ​​corresponding to each possible attribute included in the accuracy information. For example, in the attribute distribution information of the map element, the probability value corresponding to the first possible attribute is p. 1 1, meaning the probability that this map element has the first possible attribute is p. 1 1. The probability value corresponding to the first possible attribute in the accuracy information is p. 1 2, meaning that when the element is actually the first possible attribute, the probability of it being identified as the attribute recognition result in the sensor data is p. 1 2. p can be 1 1 and p 1 The product of 2, p 1 The accuracy of this first possible attribute.

[0099] Step 305: Determine the overall accuracy of the sensor data based on the accuracy of each possible attribute.

[0100] Furthermore, the overall accuracy can be determined based on the accuracy corresponding to each possible attribute. For example, the overall accuracy P is the sum of the accuracies of each possible attribute. For instance, if there are three possible attributes, p... 1 p 2 p 3 Let the accuracy of these three possible attributes be p, then the overall accuracy P is p 1 p 2 p 3 The sum of.

[0101] Step 306: Update the ratio of the accuracy of each possible attribute to the total accuracy to the probability value of the map element being each possible attribute.

[0102] In practical applications, the ratio of the accuracy of possible attributes to the overall accuracy can be updated to the probability value of a map element being a possible attribute. For example, the probability value of a map element being a possible attribute is updated to p. 1 The ratio to P.

[0103] In this implementation, the accuracy information of sensor data can be used to update the attribute distribution information of map elements. This allows the impact of a single sensor data reading on the attribute distribution information of elements to be determined based on the accuracy difference. This avoids the significant impact of low-accuracy sensor data on the attribute distribution information and also improves the impact of high-accuracy sensors on the attribute distribution information. As a result, the attribute distribution information of map elements becomes more and more accurate during the update process.

[0104] Step 307: If the number of sensor data obtained, including the attribute recognition results of map elements, is greater than a preset value, then the true attributes of the map elements are determined based on the current attribute distribution information of the map elements.

[0105] When the number of sensor data received by the electronic device for identifying the same map element is greater than a preset value, it can be assumed that the true attributes of the map element can be accurately determined based on the currently received sensor data. Therefore, the true attributes of the map element can be determined based on the current attribute distribution information of the map element.

[0106] Specifically, in the current attribute distribution information of map elements, possible attributes with a first probability value greater than a threshold can be identified as the true attributes of map elements; wherein, the current attribute distribution information of map elements includes the first probability value of each possible attribute for map elements.

[0107] For example, if the attribute distribution information includes three possible attributes with probability values ​​of p1, p2, and p3, then the attribute with the highest probability value can be taken as the attribute recognition result of the element.

[0108] Furthermore, a threshold can be set. When a possible attribute with a probability value greater than the threshold exists in the attribute distribution information, that possible attribute is used as the attribute recognition result for the element. This implementation can avoid the problem of inaccurate results in the attribute distribution information due to insufficient sensor data or other reasons.

[0109] Step 308: Update the attributes of map elements in the preset high-precision map according to the actual attributes of the map elements.

[0110] When it's necessary to update the attributes of map elements in a high-precision map, or to understand the attributes of a map element, the true attributes of the map element can be determined directly based on its current attribute distribution information. Then, the attributes of the map elements in the preset high-precision map can be updated based on these true attributes. This implementation eliminates the need to store large amounts of historical sensor data to update the attributes of map elements. This implementation also enables incremental updates to the high-precision map.

[0111] The following detailed embodiment illustrates the solution provided in this application. In this embodiment, the electronic device receives sensor data, which includes the attribute identification result of a lane line. It is assumed that this is the first time the electronic device has received the attribute identification result of this lane line, and the result is a solid line. The electronic device can generate initial attribute distribution information for this lane line.

[0112] In this embodiment, the initial attribute distribution information of the lane line can be generated based on the information of the lane line in a preset high-precision map. Assuming the lane line is marked as a dashed line in the preset high-precision map, dashed lines have a higher initial probability of 0.8, while solid lines and special dashed lines have lower initial probabilities of 0.1 each. Therefore, the initial attribute distribution information of the lane line is as follows:

[0113]

[0114] Electronic devices can also determine their accuracy information based on sensor data, which may include: the probability P(R=Solid|T=Solid) that the sensor identifies as a solid line when the true attribute is a solid line, the probability P(R=Solid|T=Dash) that the sensor identifies as a solid line when the true attribute is a dashed line, and the probability P(R=Solid|T=Special) that the sensor identifies as a solid line when the true attribute is a special dashed line.

[0115] Assuming the sensor is relatively accurate in recognizing solid lines (i.e., the probability of the sensor recognizing a solid line as a solid line is relatively high, at 0.8), and less accurate in recognizing dashed lines (i.e., the probability of the sensor recognizing a dashed line as a solid line is also relatively high, at 0.4), the accuracy of the queried data is represented as follows:

[0116]

[0117] The attribute distribution information of the lane line can be updated based on the accuracy of the acquired data. Since the sensor identified it as a solid line, the updated probability distribution is the posterior probability P(T=Solid|R=Solid), P(T=Dash|R=Solid), P(T=Dash|R=Solid), which represents the probability that the actual attribute of the lane line is solid, dashed, or a special dashed line when the sensor identifies it as solid. The specific calculation formula is as follows:

[0118]

[0119]

[0120]

[0121] in,

[0122] P(R=Solid)=

[0123] P(T=Solid)P(R=Solid|T=Solid)+P(T=Dash)P(R=Solid|T=Dash)+P(T=Special)P(R=Solid|T=Special)

[0124] If this is the first time the attribute recognition result for the element has been received, then P(T=Solid), P(T=Dash), and P(T=Special) in the formula are the probability values ​​in the initial attribute distribution information. Otherwise, they are the probability values ​​in the attribute distribution information updated based on the sensor data last time. That is, the probabilities P(T=Solid|R=Solid), P(T=Dash|R=Solid), and P(T=Dash|R=Solid) obtained from the last update are the P(T=Solid), P(T=Dash), and P(T=Special) used when updating the distribution this time.

[0125] Based on the above formulas, we can determine P(T=Solid|R=Solid), P(T=Dash|R=Solid), and P(T=Dash|R=Solid), and update the attribute distribution information based on these results, specifically:

[0126]

[0127] In this implementation, although the lane line is identified as a solid line in the sensor data, the accuracy of the data itself indicates that dashed lines are highly likely to be identified as solid lines, so the probability of a dashed line decreasing is not significant. If the identification of dashed lines is also accurate, the probability of the sensor identifying a dashed line as a solid line will decrease dramatically. It can be seen that throughout the update process, by using accurate data to evaluate and correct data errors, we can better utilize even low-quality crowdsourced data.

[0128] Continuing with the example above, suppose the system receives another set of sensor data, which includes the information that the lane lines are solid. The accuracy of the data retrieved in this case would be as follows:

[0129]

[0130] At this point, the dashed lines in the attribute distribution information are identified quite accurately, and the probability of them being reported as solid lines is very small. The attribute distribution information is the result after the last update:

[0131]

[0132] Then update the attribute distribution information, specifically as follows:

[0133]

[0134] Figure 5 This is a schematic diagram of the structure of a device for determining map element attribute information, which is an exemplary embodiment of this application.

[0135] like Figure 5 As shown, the map feature attribute information determination device 500 provided in this application includes:

[0136] The data acquisition module 510 is used to acquire sensor data and determine the accuracy information of the sensor data based on the sensor data.

[0137] The distribution acquisition module 520 is used to acquire the attribute distribution information of the map elements based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute;

[0138] The distribution update module 530 is used to update the attribute distribution information of the map element according to the accuracy information and the attribute recognition results included in the sensor data, so as to obtain the current attribute distribution information of the map element. The attribute distribution information of the map element is used to determine the true attribute of the map element.

[0139] The specific principle, implementation method, and effect of the map feature attribute information determination device provided in this embodiment are all the same as those in the previous embodiment. Figure 2 The embodiments shown are similar and will not be described again here.

[0140] Figure 6 This is a structural diagram of a device for determining map feature attribute information, as shown in another exemplary embodiment of this application.

[0141] like Figure 6 As shown, based on the above embodiments, the map feature attribute information determination device 600 provided in this embodiment,

[0142] Data acquisition module 510 includes: accurate information determination unit 511, used for:

[0143] The accuracy of the sensor data is determined based on any of the following information:

[0144] The accuracy information of the sensor data is determined based on the type of sensor used to acquire the sensor data, the attribute recognition results of the map elements included in the sensor data, and the observation conditions when the sensor acquires the sensor data.

[0145] The accuracy information is used to characterize the probability that the map element is identified as the attribute recognition result in the sensor data when the true attribute of the map element is each possible attribute.

[0146] In one optional implementation, the attribute distribution information includes a first probability value for the map element as each possible attribute; the accuracy information of the sensor data includes a second probability value for the map element being identified as the attribute identification result in the sensor data when the true attribute of the map element is each possible attribute.

[0147] The distributed update module 530 includes:

[0148] The attribute accuracy determination unit 531 is used to determine the accuracy corresponding to each possible attribute based on the first probability value of each possible attribute in the attribute distribution information and the second probability value of each possible attribute in the accuracy information of the sensor data.

[0149] The total accuracy determination unit 532 is used to determine the total accuracy of the sensor data based on the accuracy of each possible attribute.

[0150] The update unit 533 is used to update the ratio of the accuracy of each possible attribute to the total accuracy to the probability value of the map element being each possible attribute.

[0151] The attribute distribution information of the map elements includes initial attribute distribution information;

[0152] The distribution acquisition module 520 includes:

[0153] The type acquisition unit 521 is used to determine the type information of the map element based on the description information of the map element in the preset high-precision map;

[0154] The initialization unit 522 is used to generate initial attribute distribution information of the map element according to the type of the map element.

[0155] The initialization unit 522 is specifically used for:

[0156] Based on the type of the map element, determine whether the map element belongs to the easily changeable type;

[0157] If the map element is of a type that is easily changed, then the initial attribute distribution information of the map element is generated; wherein, in the initial attribute distribution information, the probability value of the map element belonging to each possible attribute is equally distributed.

[0158] The device further includes an attribute update module 540, configured to: if the number of sensor data points including the attribute recognition results of the map elements is greater than a preset value, then:

[0159] Based on the current attribute distribution information of the map element, determine the true attributes of the map element;

[0160] Update the attributes of the map elements in the preset high-precision map according to the actual attributes of the map elements.

[0161] The attribute update module 540 is specifically used for:

[0162] In the current attribute distribution information of the map element, the possible attributes with a first probability value greater than a threshold are determined as the true attributes of the map element; wherein, the current attribute distribution information of the map element includes the first probability value of the map element for each possible attribute.

[0163] The specific principle and implementation method of the map feature attribute information updating device provided in this embodiment are the same as those of the map feature attribute information updating device provided in this embodiment. Figure 3 The embodiments shown are similar and will not be described again here.

[0164] Figure 7 This is a structural diagram of a map updating device illustrated in an exemplary embodiment of this application.

[0165] like Figure 7 As shown, the map update device provided in this embodiment includes:

[0166] Memory 71;

[0167] Processor 72; and

[0168] Computer programs;

[0169] The computer program is stored in the memory 71 and configured to be executed by the processor 72 to implement any of the map feature attribute information update methods described above.

[0170] This embodiment also provides a computer-readable storage medium on which a computer program is stored.

[0171] The computer program is executed by a processor to implement any of the map feature attribute information update methods described above.

[0172] This embodiment also provides a computer program, including program code, which, when the computer runs the computer program, executes any of the map feature attribute information update methods described above.

[0173] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for updating map feature attribute information, characterized in that, include: Acquire sensor data, and determine the accuracy information of the sensor data based on the sensor data; Based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map, the attribute distribution information of the map elements is obtained; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute; Based on the accuracy information and the attribute recognition results included in the sensor data, the attribute distribution information of the map element is updated to obtain the current attribute distribution information of the map element. The attribute distribution information of the map element is used to determine the true attribute of the map element. The attribute distribution information includes a first probability value for each possible attribute of the map element; the accuracy information of the sensor data includes a second probability value for the map element being identified as the attribute identification result in the sensor data when the true attribute of the map element is each possible attribute. The step of updating the attribute distribution information of the map element based on the accuracy information and the attribute recognition results included in the sensor data to obtain the current attribute distribution information of the map element includes: The product of the first probability value of each possible attribute in the attribute distribution information and the second probability value of each possible attribute in the accuracy information of the sensor data is used as the accuracy corresponding to each possible attribute. The total accuracy of the sensor data is determined by summing the accuracies of each possible attribute. The ratio of the accuracy of each possible attribute to the total accuracy is updated to the probability value of the map element being each possible attribute; the method further includes: If the amount of sensor data acquired, including the attribute recognition results of map elements, is greater than a preset value, then the true attributes of the map elements are determined based on the current attribute distribution information of the map elements. Update the attributes of map elements in the preset high-precision map based on the actual attributes of the map elements.

2. The method according to claim 1, characterized in that, Determining the accuracy information of the sensor data based on the sensor data includes: The accuracy of the sensor data is determined based on any of the following information: The accuracy information of the sensor data is determined based on the type of sensor used to acquire the sensor data, the attribute recognition results of the map elements included in the sensor data, and the observation conditions when the sensor acquires the sensor data. The accuracy information is used to characterize the probability that the map element is identified as the attribute recognition result in the sensor data when the true attribute of the map element is each possible attribute.

3. The method according to claim 1 or 2, characterized in that, The attribute distribution information of the map elements includes initial attribute distribution information; The step of obtaining the attribute distribution information of the map elements based on the description information of the map elements in the preset high-precision map includes: The type information of the map element is determined based on the description information of the map element in the preset high-precision map; Based on the type of the map element, generate the initial attribute distribution information of the map element.

4. The method according to claim 3, characterized in that, The step of generating initial attribute distribution information for the map elements based on their type includes: Based on the type of the map element, determine whether the map element belongs to the easily changeable type; If the map element is of a type that is easily changed, then the initial attribute distribution information of the map element is generated; wherein, in the initial attribute distribution information, the probability value of the map element belonging to each possible attribute is equally distributed.

5. The method according to claim 1, characterized in that, Determining the true attributes of the map elements based on their current attribute distribution information includes: In the current attribute distribution information of the map element, the possible attributes with a first probability value greater than a threshold are determined as the true attributes of the map element; wherein, the current attribute distribution information of the map element includes the first probability value of the map element for each possible attribute.

6. A device for updating map feature attribute information, characterized in that, include: The data acquisition module is used to acquire sensor data and determine the accuracy information of the sensor data based on the sensor data. The distribution acquisition module is used to acquire the attribute distribution information of the map elements based on the description information corresponding to the map elements included in the sensor data in the preset high-precision map; the attribute distribution information includes the distribution information of the map elements belonging to each possible attribute; The update module is used to update the attribute distribution information of the map element based on the accuracy information and the attribute recognition results included in the sensor data, so as to obtain the current attribute distribution information of the map element. The attribute distribution information of the map element is used to determine the true attribute of the map element. The attribute distribution information includes a first probability value for each possible attribute of the map element; the accuracy information of the sensor data includes a second probability value for the map element being identified as the attribute identification result in the sensor data when the true attribute of the map element is each possible attribute. The update module is specifically used to multiply the first probability value of each possible attribute in the attribute distribution information by the second probability value of each possible attribute in the accuracy information of the sensor data, and use this as the accuracy corresponding to each possible attribute; determine the total accuracy of the sensor data based on the sum of the accuracies of each possible attribute; and update the ratio of the accuracy of each possible attribute to the total accuracy to the probability value of the map element being each possible attribute. The attribute update module is used to: if the number of sensor data acquisitions including attribute recognition results of map elements is greater than a preset value, determine the true attribute of the map element based on the current attribute distribution information of the map element; and update the attributes of the map elements in the preset high-precision map based on the true attributes of the map element.

7. A map updating device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, It contains computer programs. The computer program is executed by a processor to implement the method as described in any one of claims 1-5.

Images