Method and system for measuring local visibility in real time based on high-definition map

Abstract

The application discloses a kind of based on high-precision map real-time measurement local visibility method and system, this method includes: using vehicle positioning information to query high-precision map, extract high-precision map element along the way of vehicle;Video data in front of vehicle is collected by vehicle-mounted camera, and ground feature element is identified;High-precision map element and ground feature element are matched on high-precision map, determine the shielding relationship of smoke / fog to ground feature element and high-precision map element;According to the shielding relationship of smoke / fog to ground feature element and high-precision map element, calculate local visibility.The present application utilizes the rich ground information contained in high-precision map and the video data collected by vehicle-mounted camera to identify and match ground feature element, analyze the shielding of smoke / fog to ground feature element and high-precision map element, and combine the distance between vehicle and each element to calculate the upper limit and lower limit of visibility, so as to quickly estimate visibility, and can adapt to the influence of various local, short-term meteorological changes on visibility change.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving technology, specifically relating to a method and system for real-time measurement of local visibility based on high-precision maps. Background Technology

[0002] Local visibility around a vehicle is crucial for safe driving. Whether manually or autonomously driven, visibility within a small area determines safe driving. For manual driving, although the driver can detect a sharp decrease in visibility in real time and take safety measures such as slowing down, this sudden braking can introduce other safety hazards. The most direct hazard is that following vehicles may not be able to respond quickly enough to the sudden deceleration of the vehicle in front, leading to a rear-end collision. Furthermore, human drivers cannot always react quickly and safely to sudden dense fog or smoke. While autonomous vehicles possess more timely perception and decision-making capabilities than human vision, allowing for appropriate evasive maneuvers, the risk of loss of control due to sudden braking and the risk of rear-end collisions still exist. Therefore, fog patches or scattered smoke from fires in autumn and winter, which reduce visibility, all pose safety hazards. Timely and rapid detection beyond visual range is key to providing as much information and reaction time as possible for safe driving.

[0003] Weather forecasts can predict visibility over a wide area, but they struggle to accurately predict visibility over a narrower area. This narrow visibility is not only related to weather conditions but also to non-weather phenomena, such as local geographic scenes and the burning of oranges. For human drivers, we possess two high-resolution and highly synchronized visual sensors: our eyes. Through the brain's stereoscopic visual processing, our eyes can measure distances and thus determine the visibility of the surrounding environment. The starting point of this patent is how to quickly and efficiently perceive this local weather phenomenon of visibility using high-precision maps and sensors in an automated simulation.

[0004] Currently, there are two main types of methods for local visibility measurement: one type uses non-vehicle-mounted cameras and radar to measure visibility on highways, but because these are fixed to the road, they are not suitable for the continuous visibility acquisition requirements of autonomous driving on continuous highways; the other type uses vehicle-mounted cameras and weather forecasts to measure haze, such as visibility measurement for regional environmental governance needs (e.g., detection of waste incineration) and highway visibility estimation for traffic safety. Visibility measurement for regional environmental governance needs can only identify single-target haze and is not suitable for autonomous driving to identify patchy fog (not always haze) on linear highways, as it may ignore dense fog. Highway visibility estimation for traffic safety combines machine vision and atmospheric visibility estimation, but it can also only identify single-target haze or patchy fog. For example, invention patent CN109214331A discloses a traffic fog visibility detection method based on image spectrum. This method establishes a fog image library based on scene-specific images of fog on highways and trains a deep neural network for highway fog visibility detection. However, this method requires a large amount of training data and is only applicable to visibility detection in specific foggy scenarios. For other untrained scenarios, such as dense fog or scattered smoke, the accuracy is significantly reduced. During autonomous driving, the surrounding environment changes constantly. Localized and short-term fog, fog, smoke, rain, and sandstorms can all cause visibility changes, and existing visibility detection methods cannot quickly and accurately identify these changes. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for real-time measurement of local visibility based on high-precision maps, which is used to solve the problem that existing visibility recognition methods cannot be applied to local and short-term visibility change scenarios.

[0006] The first invention discloses a method for real-time measurement of local visibility based on high-precision maps, the method comprising:

[0007] Use vehicle location information to query high-precision maps and extract high-precision map elements along the vehicle's route;

[0008] The vehicle uses an onboard camera to collect video data from the front of the vehicle and identify ground features.

[0009] Match high-definition map features and ground features on a high-definition map to determine the occlusion relationship between smoke / fog and ground features and high-definition map features;

[0010] Local visibility is calculated based on the occlusion relationship between smoke / fog and ground features and high-precision map features.

[0011] Based on the above technical solutions, preferably, the high-precision map elements are traffic elements and their corresponding locations, including road information, streetlights, gas stations, traffic lights, and road signs;

[0012] Based on the above technical solutions, preferably, the feature is that the land features include traffic features, other features, and the relative positions between the features; other features include trees on both sides of the road.

[0013] Based on the above technical solutions, preferably, the step of matching high-precision map elements and geographic features on the high-precision map specifically includes:

[0014] Based on vehicle positioning, the relative positions of ground features are converted into absolute positions and overlaid onto a high-precision map to perform matching between high-precision map features and ground features.

[0015] Based on the above technical solutions, preferably, the occlusion relationship between the smoke / fog and the ground features and high-precision map features includes:

[0016] For elements that exist in both high-precision map features and ground feature features, the element farthest from the vehicle is not obscured.

[0017] For features that exist in the high-precision map but not in the ground features, the feature closest to the vehicle is occluded;

[0018] For features that exist in ground features but not in high-precision maps, the feature farthest from the vehicle is not obscured.

[0019] Based on the above technical solutions, preferably, the calculation of local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features specifically includes:

[0020] Set up a set of geographic features A and a set of high-precision map features B;

[0021] Calculate the intersection of set B and set A, the difference between set B and set A, and the difference between set A and set B respectively. Based on the distance between the vehicle and each element in the set and the occlusion relationship, calculate the upper and lower limits of local visibility to obtain the range of local visibility.

[0022] Based on the above technical solutions, preferably, the step of calculating the upper and lower limits of local visibility based on the distance between the vehicle and each element in the set and the occlusion relationship, to obtain the range of local visibility, specifically includes:

[0023] Calculate the intersection of set B and set A to obtain a set B∩A consisting of elements present in both high-definition map features and ground feature features. Calculate the distance between each element in set B∩A and the vehicle, and denote the maximum distance as D. max (B∩A);

[0024] Calculate the difference between set B and set A to obtain set BA, which consists of elements that exist in the high-precision map but not in the geographic features. Calculate the distance between each element in set BA and the vehicle, and denote the minimum distance as D. min (BA);

[0025] Calculate the difference between set A and set B to obtain set AB, which consists of elements that exist in the geographic features but not in the high-definition map. Calculate the distance between each element in set AB and the vehicle, and denote the maximum distance as D. max (AB);

[0026] The range of local visibility is:

[0027] max(D max (B∩A),D max (AB))≤Visibility≤D min (BA).

[0028] max(·) is the function to find the maximum value.

[0029] In a first aspect, the present invention discloses a system for real-time measurement of local visibility based on high-precision maps, the system comprising:

[0030] Map feature extraction module: Used to query high-precision maps using vehicle location information and extract high-precision map features along the vehicle's route;

[0031] Ground feature recognition module: used to collect video data in front of the vehicle through the vehicle-mounted camera and recognize ground features;

[0032] Feature matching module: used to match high-definition map features and ground feature features on a high-definition map, and to determine the occlusion relationship between ground feature features and high-definition map features;

[0033] Visibility Calculation Module: Used to calculate the range of local visibility based on the occlusion relationship between ground features and high-precision map features.

[0034] A third aspect of the present invention discloses an electronic device comprising: at least one processor, at least one memory, a communication interface, and a bus;

[0035] The processor, memory, and communication interface communicate with each other through the bus.

[0036] The memory stores program instructions that can be executed by the processor, which invokes the program instructions to implement the method as described in the first aspect of the present invention.

[0037] In a fourth aspect, the present invention discloses a computer-readable storage medium storing computer instructions that cause a computer to perform the method described in the first aspect of the present invention.

[0038] The present invention has the following advantages over the prior art:

[0039] 1) This invention utilizes the rich ground feature information contained in high-precision maps and video data collected by vehicle cameras to identify and match ground feature elements, analyze the occlusion relationship of fog on traffic elements and ground feature elements, and measure the visibility range of the vehicle's surrounding environment. It can be applied to the real-time continuous measurement of local visibility along highways in autonomous driving scenarios.

[0040] 2) This invention analyzes the occlusion of ground features and high-definition map features by the inclusion relationship between the sets of high-definition map features and ground features, and calculates the upper and lower limits of visibility by combining the distance between the vehicle and each feature. Visibility can be quickly estimated without a large amount of data or complex calculations.

[0041] 3) This invention can adapt to the impact of various local and short-term weather changes on visibility changes. When a small-scale fog or smoke is measured, visibility can be measured by the obstruction of the line of sight to ground objects by the smoke / fog, so as to estimate the visibility index required for autonomous driving decision-making without bias. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 Examples of the impact of smoke / fog on roads;

[0044] Figure 2 This is a schematic diagram of high-precision map elements extracted from a high-precision map.

[0045] Figure 3 A schematic diagram of ground features collected from video data;

[0046] Figure 4 This is a diagram illustrating the result of finding the intersection of set B and set A.

[0047] Figure 5 A high-precision map showing the result of finding the intersection of set B and set A.

[0048] Figure 6 This is a diagram illustrating the result of subtracting set A from set B.

[0049] Figure 7 A high-precision map showing the result of subtracting set A from set B.

[0050] Figure 8 This is a diagram illustrating the result of subtracting set B from set A.

[0051] Figure 9 A high-precision map showing the result of subtracting set B from set A.

[0052] Figure 10 This is a schematic diagram showing the distances between each element in set A∩B and the vehicle.

[0053] Figure 11 This is a diagram illustrating the distances between various elements in the Business Architecture (BA) and the vehicles.

[0054] Figure 12 This is a schematic diagram showing the distances between each element in set AB and the vehicle.

[0055] Figure 13 This is a schematic diagram illustrating the relationship between visibility and various sets. Detailed Implementation

[0056] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0057] In smoky / foggy scenarios, vehicles cannot perceive ground features within or in front of the smoke area, especially for autonomous vehicles relying solely on visual perception. Figure 1The image illustrates the impact of smoke / fog on roads. In autonomous driving, decisions and controls such as automatic cruise control, lane change safety checks, and lane change execution all rely on the vehicle's accurate judgment of the distance to road features and surrounding objects. Existing visibility measurement methods are generally geared towards non-driving scenarios, such as visibility measurement in fixed areas or regional-level fog and haze detection, and cannot be directly applied to continuous visibility measurements along highways. Furthermore, existing methods struggle to effectively assess small patches of fog or smoke appearing on roads. For small patches of smoke or fog, due to their small size, measurements can easily be made using surrounding visibility to mask the visibility along the vehicle's direction of travel, leading to overestimation of visibility and misjudgments in autonomous driving decisions, potentially causing traffic accidents. Traditional weather forecasts are insufficient for driving requirements at a microscale; this invention fully utilizes the vehicle's physical sensors and high-precision maps to capture real-time changes in visibility.

[0058] Please see Figure 1 This invention proposes a method for real-time measurement of local visibility based on high-precision maps, the method comprising:

[0059] S1. Use vehicle location information to query high-precision maps and extract high-precision map elements along the vehicle's route.

[0060] Based on vehicle positioning devices, such as GPS, the vehicle's location coordinates are determined. Then, using line-of-sight analysis, high-precision map elements along the vehicle's route are extracted from the high-precision map. These elements primarily include road information, streetlights, gas stations, traffic lights, and road signs, along with their corresponding locations. Road information includes road width, curvature, slope, and aspect. Figure 2 The image shown is a schematic diagram of high-precision map elements extracted from a high-precision map.

[0061] S2. Collect video data in front of the vehicle using the vehicle-mounted camera and identify ground features.

[0062] For video images captured by vehicle-mounted cameras, image recognition tools are used to perceive ground features and identify them. For example... Figure 3 The image shows a schematic diagram of ground features collected from video data. These features mainly include traffic elements and other elements. Traffic elements primarily include road information, streetlights, gas stations, traffic lights, and road signs. Other elements include trees along the roadside and other landmarks. The relative positions of these features are then extracted.

[0063] S3. Match high-definition map features and ground features on the high-definition map to determine the occlusion relationship between smoke / fog and ground features and high-definition map features.

[0064] Based on vehicle positioning, the relative positions of ground features are converted into absolute positions and overlaid onto a high-precision map to perform matching between high-precision map features and ground features.

[0065] Let A be the set of geographic features extracted from video data, with... Figure 3 Taking the corresponding land features as an example, the set A composed of these land features includes tree a, tree b, tree c, street lamp a, street lamp b, and road sign a.

[0066] Let B be a set of high-definition map elements extracted from the high-definition map, with... Figure 2 Taking the corresponding high-precision map elements as an example, the set B composed of these high-precision map elements includes street lamp 1, street lamp 2, road sign 1, and road sign 2.

[0067] Based on the matching results of high-definition map features and ground feature features, determine the occlusion relationship between smoke / fog and ground feature features and high-definition map features. Regarding the matching results, the following situations exist:

[0068] ① For elements that exist in both high-precision map elements and ground feature elements, the element farthest from the vehicle is not obscured.

[0069] Finding the intersection of set B and set A primarily aims to match the geographic features extracted from video images with features from high-definition maps, thereby obtaining the features present in both the high-definition maps and video images. The result is as follows: Figure 4 As shown, mapping the intersection result onto a high-precision map will result in the following: Figure 5 As shown. At this time, Figure 5 The fact that the furthest element from the vehicle (road sign 1) is not obscured indicates that the visibility is not less than the distance between the vehicle and the furthest element.

[0070] ② For features that exist in the high-precision map but not in the ground features, the feature closest to the vehicle will be occluded.

[0071] Subtracting set A from set B, we retain only the elements contained in set B and those not contained in set A, denoted as set BA. Set BA is actually distributed only outside the video ranging area. These elements primarily distinguish between the elements theoretically recognizable by the vehicle and those actually recognized, determining which elements are unrecognizable by the vehicle's camera. These unrecognizable elements represent areas potentially affected by reduced visibility. The difference between set B and set A is shown below. Figure 6 As shown, mapping the results of BA onto a high-precision map will result in the following: Figure 7 As shown. At this time, Figure 7 The fact that the nearest element to the vehicle (road sign 2) is obscured indicates that the visibility is no greater than the distance between the vehicle and the nearest element.

[0072] ③ For features that are not present in the high-precision map but exist in the ground features, the feature farthest from the vehicle is not obscured.

[0073] Subtracting set B from set A, retaining only elements contained in set A and subtracting those not contained in set B, results in set AB. This set of elements primarily distinguishes between features actually visible to vehicles and features in high-definition maps. Features present in the real environment, such as trees, will not appear in high-definition maps. These features, only present in video data, further supplement visibility recognition. The difference between set A and set B is shown below. Figure 8 As shown, mapping the results of BA onto a high-precision map will result in the following: Figure 9 As shown. At this time, Figure 9 The fact that the furthest element (tree c) from the vehicle is not obscured indicates that the visibility is not less than the distance between the vehicle and the furthest element.

[0074] S4. Calculate local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features.

[0075] This invention calculates the intersection of set B and set A, the difference between set B and set A, and the difference between set A and set B. Based on the distance between the vehicle and each element in the set and the occlusion relationship, it establishes the relationship between visibility and each set, calculates the upper and lower limits of local visibility, and then performs visibility measurement to obtain the range of local visibility.

[0076] S41. Calculate the intersection of set B and set A to obtain a set B∩A composed of elements that exist in both high-precision map elements and ground feature elements. Obtain the geographic coordinates of these elements, calculate the distance between each element in set B∩A and the vehicle, and calculate the lower limit of visibility.

[0077] The distance calculation formula is:

[0078]

[0079] Where, d n X represents the distance between the vehicle and various elements, and X and Y represent the geographical coordinates of the vehicle's location. n y n These are the coordinates of element n, where the subscript n represents the number of elements, resulting in a series of distance values ​​between elements and vehicles. Figure 10 This is a schematic diagram showing the distances d1, d2, and d3 between each element in set B∩A and the vehicle. Find the maximum value among these distances and denote it as D. max(B∩A). According to the definition of visibility, the concept of visibility can be understood as the farthest object that can be seen. Therefore, visibility should be no less than the nearest unseen object. So, for D... max In the case of (B∩A), the visibility should be greater than or equal to D. max (B∩A).

[0080] S42. Calculate the difference between set B and set A to obtain set BA, which consists of elements that exist in the high-precision map but not in the ground features. Calculate the distance between each element in set BA and the vehicle, and calculate the upper limit of visibility.

[0081] Figure 11 The diagram shows the distance d4 between elements in set BA and vehicles. The minimum distance is denoted as D. min (BA). According to the definition of visibility, the concept of visibility can be understood as the farthest object that can be seen. Therefore, visibility should not be greater than the nearest unseen object. So for D... min For (BA), the visibility should be ≤ D. min (BA).

[0082] S43. Calculate the difference between set A and set B to obtain set AB, which consists of elements that do not exist in the high-precision map but exist in the ground features. Calculate the distance between each element in set AB and the vehicle.

[0083] Figure 12 This is a schematic diagram showing the distances d5, d6, and d7 between each element in set AB and the vehicle. The maximum distance is denoted as D. max (AB), According to the definition of visibility, visibility can be understood as the farthest object that can be seen. Therefore, visibility should be no less than the nearest unseen object. So, for D... max For (AB), Visibility ≥ D should be true. max (AB).

[0084] Based on the above analysis, we can conclude that... Figure 13 The diagram shows the relationship between visibility and various sets. When D is obtained... max (B∩A), D max (AB) and D min After (BA), D needs to be addressed. max (B∩A), D max (A and B are further compared to obtain the maximum value of the two, max(D)) max (B∩A),D max(AB)), where max(·) is the function for finding the maximum value.

[0085] The final output is the range of visibility in the vehicle's environment, i.e., max(D). max (B∩A),D max (AB))≤Visibility≤D min (BA).

[0086] This invention addresses the shortcomings of existing visibility data in the field of autonomous driving, which lacks locality and short-term relevance. It constructs a visibility measurement method based on high-precision maps, matching road condition information acquired by cameras with road condition information in high-precision maps, thereby providing support for the localized and short-term relevance high-precision visibility required for autonomous driving. The high-precision map technology employed in this invention provides rich information on ground features and their high-precision locations. On one hand, when small-scale fog or smoke is measured, visibility can be accurately measured by the obstruction of the view of ground features by the fog / smoke, thus enabling unbiased estimation of the visibility indicators required for autonomous driving decisions. On the other hand, the prior ground feature information contained in the high-precision map improves the accuracy of ground feature identification from video data, thereby also improving the accuracy of visibility measurement based on ground feature identification.

[0087] This invention establishes a correlation between high-precision map data, vehicle camera data, and visibility, enabling visibility measurement based on high-precision maps. This provides the necessary local and short-term visibility data for autonomous vehicle control and decision-making. Furthermore, the method described in this invention can be extended to visibility measurement in various weather conditions, such as short-term heavy rainfall, sandstorms, and snowfall, demonstrating significant application potential. Currently, domestic standards only support visibility measurement methods based on vehicle-mounted cameras; no methods based on high-precision maps have been found. This invention fills this gap and better adapts to the actual needs of autonomous driving.

[0088] Corresponding to the above method embodiments, the present invention also proposes a system for real-time measurement of local visibility based on high-precision maps, the system comprising:

[0089] Map feature extraction module: Used to query high-precision maps using vehicle location information and extract high-precision map features along the vehicle's route;

[0090] Ground feature recognition module: used to collect video data in front of the vehicle through the vehicle-mounted camera and recognize ground features;

[0091] Feature matching module: used to match high-definition map features and ground features on a high-definition map, and to determine the occlusion relationship between smoke / fog and ground features and high-definition map features;

[0092] Visibility calculation module: used to calculate local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features.

[0093] The above system embodiments and method embodiments are one-to-one correspondences. For a brief description of the system embodiments, please refer to the method embodiments.

[0094] The present invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface, and a bus; wherein the processor, memory, and communication interface communicate with each other through the bus; the memory stores program instructions executable by the processor, and the processor calls the program instructions to implement the aforementioned method of the present invention.

[0095] The present invention also discloses a computer-readable storage medium that stores computer instructions, which cause the computer to implement all or part of the steps of the method described in the embodiments of the present invention. The storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0096] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, meaning they can be distributed across multiple network units. Those skilled in the art can select some or all of the modules to achieve the purpose of this embodiment without any inventive effort, based on actual needs.

[0097] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for real-time measurement of local visibility based on high-precision maps, characterized in that, The method includes: Use vehicle location information to query high-precision maps and extract high-precision map elements along the vehicle's route; The vehicle uses an onboard camera to collect video data from the front of the vehicle and identify ground features. Match high-definition map features and ground features on a high-definition map to determine the occlusion relationship between smoke / fog and ground features and high-definition map features; Local visibility is calculated based on the occlusion relationship between smoke / fog and ground features and high-precision map features; The calculation of local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features specifically includes: Set up a set of geographic features A and a set of high-precision map features B; Calculate the intersection of set B and set A, the difference between set B and set A, and the difference between set A and set B respectively. Based on the distance between the vehicle and each element in the set and the occlusion relationship, calculate the upper and lower limits of local visibility to obtain the range of local visibility.

2. The method for real-time measurement of local visibility based on high-precision maps according to claim 1, characterized in that, The high-precision map elements are traffic elements and their corresponding locations. Traffic elements include road information, streetlights, gas stations, traffic lights, and road signs.

3. The method for real-time measurement of local visibility based on high-precision maps according to claim 2, characterized in that, The geographical features include traffic features, other features, and the relative positions between features; other features include trees along the roads.

4. The method for real-time measurement of local visibility based on high-precision maps according to claim 1, characterized in that, The matching of high-precision map features and geographic features on a high-precision map specifically includes: Based on vehicle positioning, the relative positions of ground features are converted into absolute positions and overlaid onto a high-precision map to perform matching between high-precision map features and ground features.

5. The method for real-time measurement of local visibility based on high-precision maps according to claim 3, characterized in that, The occlusion relationships of the smoke / fog on ground features and high-precision map features include: For elements that exist in both high-precision map features and ground feature features, the element farthest from the vehicle is not obscured. For features that exist in the high-precision map but not in the ground features, the feature closest to the vehicle is occluded; For features that exist in ground features but not in high-precision maps, the feature farthest from the vehicle is not obscured.

6. The method for real-time measurement of local visibility based on high-precision maps according to claim 1, characterized in that, The calculation of the upper and lower limits of local visibility based on the distance between the vehicle and each element in the set, as well as the occlusion relationship, to obtain the range of local visibility specifically includes: Calculate the intersection of set B and set A to obtain a set B∩A consisting of elements present in both high-definition map features and ground feature features. Calculate the distance between each element in set B∩A and the vehicle, and denote the maximum distance as D. max (B∩A); Calculate the difference between set B and set A to obtain set BA, which consists of elements that exist in the high-precision map but not in the geographic features. Calculate the distance between each element in set BA and the vehicle, and denote the minimum distance as D. min (BA); Calculate the difference between set A and set B to obtain set AB, which consists of elements that exist in the geographic features but not in the high-definition map. Calculate the distance between each element in set AB and the vehicle, and denote the maximum distance as D. max (AB); The range of local visibility is: max(D max (B∩A),D max (A-B))≤Visibility≤D min (B-A); max(·) is the function to find the maximum value.

7. A system for real-time measurement of local visibility based on high-precision maps, characterized in that, The system includes: Map feature extraction module: Used to query high-precision maps using vehicle location information and extract high-precision map features along the vehicle's route; Ground feature recognition module: used to collect video data in front of the vehicle through the vehicle-mounted camera and recognize ground features; Feature matching module: used to match high-definition map features and ground features on a high-definition map, and to determine the occlusion relationship between smoke / fog and ground features and high-definition map features; Visibility calculation module: used to calculate local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features; The calculation of local visibility based on the occlusion relationship between smoke / fog and ground features and high-precision map features specifically includes: Set up a set of geographic features A and a set of high-precision map features B; Calculate the intersection of set B and set A, the difference between set B and set A, and the difference between set A and set B respectively. Based on the distance between the vehicle and each element in the set and the occlusion relationship, calculate the upper and lower limits of local visibility to obtain the range of local visibility.

8. An electronic device, characterized in that, include: At least one processor, at least one memory, a communication interface, and a bus; The processor, memory, and communication interface communicate with each other through the bus. The memory stores program instructions that can be executed by the processor, which invokes the program instructions to implement the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause the computer to perform the method as described in any one of claims 1 to 6.

Images