A method and device for determining a region of interest

Abstract

The embodiment of the present invention discloses a method and device for determining an area of ​​interest. After locating the target road section where a vehicle is currently located, the area of ​​interest of the vehicle can be screened out according to the connection relationship between the sections and the distance between the boundary points of the sections and the vehicle. That is, some sections can be screened out as the area of ​​interest among all the sections within the perception range, so that when performing environmental perception, only the perception information of the area of ​​interest can be calculated. Compared with calculating all the perception information, the complexity of the calculation can be reduced, that is, the difficulty of environmental perception can be reduced, and the accuracy of the environmental perception results can be improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method and device for determining a region of interest. Background Art

[0002] In the closed scene autonomous driving technology, the existing environmental perception methods mainly use cameras and lidar as perception devices. Cameras can provide rich visual information, and then use the principle of binocular stereo vision to judge distance, depth and other information. They can also be used for 3D reconstruction, object recognition, etc. Lidar can provide rich point cloud data with advantages such as high accuracy and good robustness. Among them, the above-mentioned closed scenes can include ports, mines, airports, logistics parks and other scenes with limited areas.

[0003] Specifically, known camera- and radar-based environmental perception methods generally use all information acquired within the perception range for autonomous driving decision-making and navigation.

[0004] However, the above method has the following problems: cameras and radars can provide a wide range of perception and rich environmental information, but not all the information within their perception range can be used for autonomous driving decisions and navigation. In other words, there is some redundant information that is irrelevant to autonomous driving, such as road information far away from the vehicle and information outside the road. If all this redundant information is used, it will increase the complexity of calculation and the difficulty of environmental perception, which may cause large errors in environmental perception results. Summary of the invention

[0005] The present invention provides a method and device for determining a region of interest to reduce the difficulty of environmental perception and improve the accuracy of environmental perception results. The specific technical solution is as follows.

[0006] In a first aspect, an embodiment of the present invention provides a method for determining a region of interest, the method comprising:

[0007] Determine the current position of the vehicle in the map coordinate system;

[0008] Loading a high-precision map, wherein the high-precision map includes at least the location information of each road section and the connection relationship between the road sections;

[0009] Determine the target road section where the vehicle is currently located according to the current position of the vehicle and the position information of each road section included in the high-precision map;

[0010] According to the connection relationship between the road sections, searching for the road sections connected to the target road section, and adding the target road section and the searched road sections into the array of road sections to be extracted;

[0011] Detecting whether there is an unextracted road section in the array of road sections to be extracted;

[0012] When there are unextracted road sections in the array of road sections to be extracted, a current unextracted road section is selected from the array of road sections to be extracted, and among the boundary points of the current unextracted road section, target boundary points of the current unextracted road section are screened according to the distance between each boundary point and the vehicle as the region of interest of the vehicle; and when the region of interest is screened out, the unextracted road section connected to the current unextracted road section is added to the array of road sections to be extracted, and the step of detecting whether there are unextracted road sections in the array of road sections to be extracted is returned to execute;

[0013] When there is no unextracted road section in the array of road sections to be extracted, each region of interest of the vehicle is obtained.

[0014] Optionally, the high-precision map further includes the type of each road section, and the type of each road section includes a road or an intersection;

[0015] Among the boundary points of the currently unextracted road section, according to the distance between each boundary point and the vehicle, the target boundary points of the currently unextracted road section are screened as the region of interest of the vehicle, including:

[0016] When the current unextracted road section is a road, edge lines on both sides of the current unextracted road section are obtained, and two endpoints whose distances from the vehicle are within a preset search radius are determined from points included in each edge line as regions of interest of the vehicle;

[0017] When the current unextracted road section is an intersection, all boundary points of the current unextracted road section are acquired, and the region of interest of the vehicle is determined according to the distance between each boundary point and the vehicle.

[0018] Optionally, determining two endpoints within a preset search radius from the vehicle among the points included in each edge line as the region of interest of the vehicle includes:

[0019] For any point included in an edge line, when the point meets any of the following conditions, the point is taken as the area of ​​interest of the vehicle: the current distance between the point and the vehicle is less than or equal to the preset search radius, and the point is the endpoint of the edge line; the current distance between the point and the vehicle is less than or equal to the preset search radius, and there is an adjacent point of the point whose current distance to the vehicle is greater than the preset search radius.

[0020] Optionally, when the current unextracted road section is a road, obtaining edge lines on both sides of the current unextracted road section includes:

[0021] When the currently unextracted road section is a road, determining whether the currently unextracted road section is a straight road;

[0022] When the currently unextracted road section is a straight road, edge lines on both sides of the straight road are obtained from the high-precision map as edge lines on both sides of the currently unextracted road section;

[0023] When the currently unextracted road section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and the edge lines on each side are divided into multiple straight line segments, and the obtained multiple straight line segments are used as the edge lines of that side of the currently unextracted road.

[0024] Optionally, determining the region of interest of the vehicle according to the distance between each boundary point and the vehicle includes:

[0025] Calculating the distance between each boundary point and the vehicle;

[0026] When there is a boundary point whose distance to the vehicle is less than or equal to the preset search radius, all boundary points are regarded as the region of interest of the vehicle.

[0027] Optionally, the method further includes:

[0028] Data including information of each region of interest of the vehicle is constructed and sent, wherein the information of each region of interest at least includes: the number of regions of interest, the current position of the vehicle, the number of boundary points of each region of interest, and the position information of each boundary point.

[0029] In a second aspect, an embodiment of the present invention provides a device for determining a region of interest, the device comprising:

[0030] A current position determination module, used to determine the current position of the vehicle in the map coordinate system;

[0031] A high-precision map loading module is used to load a high-precision map, wherein the high-precision map includes at least the location information of each road section and the connection relationship between each road section;

[0032] A target road section determination module, used to determine the target road section where the vehicle is currently located according to the current position of the vehicle and the position information of each road section included in the high-precision map;

[0033] A connection section search module, used for searching for sections connected to the target section according to the connection relationship between the sections, and adding the target section and the searched sections into an array of sections to be extracted;

[0034] An unextracted road section detection module is used to detect whether there is an unextracted road section in the array of road sections to be extracted;

[0035] a road section detection module, configured to select a current unextracted road section from the array of road sections to be extracted when the unextracted road section detection module determines that there are unextracted road sections in the array of road sections to be extracted, and select target boundary points of the current unextracted road section from boundary points of the current unextracted road section according to the distance between each boundary point and the vehicle as the region of interest of the vehicle; and when the region of interest is screened out, add the unextracted road section connected to the current unextracted road section to the array of road sections to be extracted, and trigger the unextracted road section detection module;

[0036] The region of interest determination module is used to obtain the regions of interest of the vehicle when the unextracted road section detection module determines that there is no unextracted road section in the array of road sections to be extracted.

[0037] Optionally, the high-precision map further includes the type of each road section, and the type of each road section includes a road or an intersection; and the road section detection module includes:

[0038] A road detection submodule, for obtaining edge lines on both sides of the current unextracted road section when the current unextracted road section is a road, and determining two endpoints within a preset search radius from the vehicle among the points included in each edge line as the region of interest of the vehicle;

[0039] The intersection detection submodule is used to obtain all boundary points of the current unextracted road section when the current unextracted road section is an intersection, and determine the area of ​​interest of the vehicle according to the distance between each boundary point and the vehicle.

[0040] Optionally, the road detection submodule is specifically used to: for a point included in any edge line, when the point meets any of the following conditions, use the point as the area of ​​interest of the vehicle: the current distance between the point and the vehicle is less than or equal to a preset search radius, and the point is the endpoint of the edge line; the current distance between the point and the vehicle is less than or equal to a preset search radius, and there is an adjacent point of the point whose current distance to the vehicle is greater than the preset search radius.

[0041] Optionally, the road detection submodule is specifically used for:

[0042] When the currently unextracted road section is a road, determining whether the currently unextracted road section is a straight road;

[0043] When the currently unextracted road section is a straight road, edge lines on both sides of the straight road are obtained from the high-precision map as edge lines on both sides of the currently unextracted road section;

[0044] When the currently unextracted road section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and the edge lines on each side are divided into multiple straight line segments, and the obtained multiple straight line segments are used as the edge lines of that side of the currently unextracted road.

[0045] Optionally, the intersection detection submodule is specifically used for:

[0046] Calculating the distance between each boundary point and the vehicle;

[0047] When there is a boundary point whose distance to the vehicle is less than or equal to the preset search radius, all boundary points are regarded as the region of interest of the vehicle.

[0048] Optionally, the device further comprises:

[0049] The information building module is used to build and send data including information of each area of ​​interest of the vehicle, wherein the information of each area of ​​interest includes at least: the number of areas of interest, the current position of the vehicle, the number of boundary points of each area of ​​interest, and the position information of each boundary point.

[0050] From the above content, it can be seen that the method and device for determining the area of ​​interest provided by the embodiments of the present invention can, after locating the target section where the vehicle is currently located, filter out the vehicle's area of ​​interest based on the connection relationship between the sections and the distance between the boundary points of the sections and the vehicle. In other words, some sections can be filtered out as areas of interest from all sections within the perception range, so that when performing environmental perception, only the perception information of the area of ​​interest can be calculated. Compared with calculating all the perception information, the complexity of the calculation can be reduced, that is, the difficulty of environmental perception can be reduced, and the accuracy of the environmental perception results can be improved.

[0051] The innovative features of the embodiments of the present invention include:

[0052] 1. After locating the target road section where the vehicle is currently located, the area of ​​interest of the vehicle is screened out according to the connection relationship between the sections and the distance between the boundary points of each section and the vehicle. In other words, some sections can be screened out as the area of ​​interest from all the sections within the perception range, so that when performing environmental perception, only the perception information of the area of ​​interest can be calculated. Compared with calculating all the perception information, the complexity of the calculation can be reduced, that is, the difficulty of environmental perception can be reduced, and the accuracy of the environmental perception results can be improved.

[0053] 2. For road-type sections, the traffic conditions are usually simple. Therefore, for road-type sections, the area of ​​interest can be determined based on the edge lines on both sides to save computing resources. For intersection-type sections, the traffic conditions are usually complex. Therefore, for intersection-type sections, the area of ​​interest can be determined based on all their boundary points to obtain more sufficient road information and improve the accuracy of environmental perception results.

[0054] 3. For a curved road, after it is divided into multiple straight sections, the region of interest determination method for straight sections can be used to determine the region of interest of the curved road, thereby reducing the difficulty of determining the region of interest of the curved road and improving the accuracy of determining the region of interest.

[0055] 4. For intersection-type sections, when there are boundary points within the vehicle search radius, all boundary points are regarded as areas of interest, that is, the information of the entire intersection can be processed as useful perception information, which can improve the accuracy of environmental perception and thus improve the safety of vehicle driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work.

[0057] Figure 1 A schematic diagram of a flow chart of a method for determining a region of interest provided by an embodiment of the present invention;

[0058] Figure 2 A schematic diagram of information stored in a high-precision map in an embodiment of the present invention;

[0059] Figure 3 Another schematic diagram of a flow chart of a method for determining a region of interest provided by an embodiment of the present invention;

[0060] Figure 4 A schematic diagram of the boundary point determination results for the road type section;

[0061] Figure 5 A schematic diagram of the boundary point determination result for the intersection type road section;

[0062] Figure 6 This is a schematic diagram of the extraction result of a region of interest;

[0063] Figure 7 This is another schematic diagram of the extraction result of the region of interest;

[0064] Figure 8 A schematic diagram of information constructed in an embodiment of the present invention;

[0065] Fig. 9 A schematic structural diagram of a device for determining a region of interest provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0068] The embodiment of the present invention discloses a method and device for determining a region of interest, which can reduce the difficulty of environmental perception and improve the accuracy of environmental perception results. The embodiment of the present invention is described in detail below.

[0069] Figure 1 A flowchart of a method for determining an area of ​​interest provided in an embodiment of the present invention. The method can be applied to electronic devices, and specifically can be used in a processor with a data processing function in an autonomous driving vehicle, or any processor connected to an autonomous driving vehicle, which is not limited in the embodiment of the present invention. The method specifically includes the following steps.

[0070] S110: Determine the current position of the vehicle in the map coordinate system.

[0071] In an embodiment of the present invention, the electronic device can perform real-time positioning operations on the vehicle. For example, the current position of the vehicle can be determined according to a set period, such as 1 millisecond, 3 milliseconds, 5 milliseconds, etc.

[0072] In one implementation, the vehicle can be positioned by GPS (Global Positioning System) to obtain the position of the vehicle in the GPS coordinate system, that is, the longitude, latitude, and altitude coordinates of the vehicle can be obtained. In an embodiment of the present invention, in order to determine the road section where the vehicle is located in the high-precision map, and to match the coordinates of various elements in the high-precision map, such as lane lines, lane edges, etc., to determine the area of ​​interest around the vehicle, that is, the area that plays an important role in vehicle decision-making. After obtaining the position of the vehicle in the GPS coordinate system, it can be converted to the map coordinate system. The above-mentioned area of ​​interest can also be called a geographic fence, which refers to a range around the vehicle, and the road conditions within this range play an important role in the vehicle's decision-making.

[0073] For example, the coordinates (x, y, z) of the vehicle in the map coordinate system can be converted using the following formula:

[0074]

[0075] Among them, lon is the longitude of the vehicle in the GPS coordinate system, lat is the latitude of the vehicle in the GPS coordinate system, height is the height of the vehicle in the GPS coordinate system, and a and e are known constants.

[0076] S120: Loading a high-precision map, where the high-precision map at least includes location information of each road section and connection relationships between the road sections.

[0077] In an embodiment of the present invention, a high-precision map of a closed scene can be pre-built and stored. Figure 2 As shown in the figure, the information stored in the high-precision map is as shown in the figure. Taking segment 2 (segment2) as an example, the high-precision map can store the following information: segment 2 includes road 1 and road 2, wherein road 1 is composed of a left lane line, a right lane line, and a road center line; road 2 is also composed of a left lane line, a right lane line, and a road center line; the segments connected to segment 2 are the upper segment 1 (segment1) and the lower segment 3 (segment3). Among them, the specific information of segment 1 and segment 3 is similar to that of segment 2, and the embodiment of the present invention will not be repeated.

[0078] S130: Determine the target road section where the vehicle is currently located based on the current position of the vehicle and the position information of each road section included in the high-precision map.

[0079] For example, the current position of the vehicle can be input into the HD map through the HD map interface. The current position of the vehicle is known, and the HD map stores the position information of each road section, so the road section number where the vehicle is located can be determined, which can be called the target road section.

[0080] S140: searching for each road segment connected to the target road segment according to the connection relationship between each road segment, and adding the target road segment and each searched road segment into an array of road segments to be extracted.

[0081] It can be understood that the perception information that plays a key role in vehicle decision-making is the information around the vehicle, and the road information far away from the vehicle has little effect on the vehicle's decision-making. Therefore, in the embodiment of the present invention, after determining the target road section where the vehicle is currently located, the vehicle's area of ​​interest can be determined in the road section near the target road section.

[0082] Specifically, the electronic device can search for each road section connected to the target road section according to the connection relationship between each road section, and add the target road section and each searched road section to the road section array to be extracted. The above-mentioned road section array to be extracted is a data set, which stores all road section information that needs to be extracted from the area of ​​interest, such as the road section number, each lane line information included in the road section, etc. For example, the information stored in the high-precision map of the target road section and each searched road section can be directly copied and stored in the road section array to be extracted.

[0083] For example, Figure 2 For example, when it is determined that the current target section of the vehicle is Section 2, according to the connection relationship stored in the high-precision map, the sections connected to Section 2 are Section 1 and Section 3. The relevant information of Section 2, Section 1 and Section 3 stored in the high-precision map can be added to the array of sections to be extracted.

[0084] S150: Detect whether there are unextracted road sections in the array of road sections to be extracted.

[0085] For example, each road segment information may be marked, for example, an unextracted road segment may be marked as 0, and an extracted road segment may be marked as 1. In this case, when the electronic device detects whether there is an unextracted road segment in the array of road segments to be extracted, it may detect whether there is a road segment marked as 0 in the array of road segments to be extracted, and if so, it may be determined that there is an unextracted road segment.

[0086] It is understandable that it is possible to mark each road section as extracted or not extracted in other ways, such as using different letters, numbers, or symbols for marking, as long as the extracted road section and the unextracted road section can be distinguished. The embodiment of the present invention does not specifically limit this.

[0087] S160: When there are unextracted sections in the array of sections to be extracted, a current unextracted section is selected in the array of sections to be extracted, and among the boundary points of the current unextracted section, the target boundary points of the current unextracted section are screened according to the distance between each boundary point and the vehicle as the area of ​​interest of the vehicle; and when the area of ​​interest is screened out, the unextracted section connected to the current unextracted section is added to the array of sections to be extracted, and the process returns to execute step S150.

[0088] When there are unextracted sections in the array of sections to be extracted, a current unextracted section can be selected from the array of sections to be extracted. For example, the earliest stored section can be selected as the current unextracted section according to the storage order of the sections; or, the section with the smallest number can be selected as the current unextracted section according to the numbering order of the sections; or, the current unextracted section can be selected according to other rules. All of these are possible and are not limited to this in the embodiments of the present invention.

[0089] It can be understood that the entire road section is usually a large area, in which each small area has different attributes, such as different road conditions, different distances from the vehicle, etc., and their role in vehicle decision-making is not exactly the same. Therefore, in the unextracted road section, the area that plays an important role in vehicle decision-making can be determined, which is called the region of interest.

[0090] Specifically, the target boundary points of the currently unextracted road section can be selected as the vehicle's area of ​​interest based on the distance between each boundary point and the vehicle. The boundary points of each road section can be directly obtained from the information stored in the high-precision map. Figure 2 In the road section 2 shown, its boundary points are the left lane line of road 1 and the right lane line of road 2, that is, the two edge lines of road section 2.

[0091] For example, the distance between each boundary point and the vehicle can be calculated, and the boundary points whose distance is less than or equal to the preset distance can be used as target boundary points. Multiple target boundary points can define an area as the area of ​​interest of the vehicle. The preset distance can be set according to actual conditions, such as the size and type of the closed scene, or the decision accuracy of the autonomous driving vehicle. This is possible and is not limited in the embodiments of the present invention.

[0092] After the currently unextracted road section is extracted, the mark of the currently unextracted road section can be modified to extracted according to a preset marking rule.

[0093] When the current unextracted road segment screens out the region of interest, it indicates that there is a region in the current unextracted road segment that is close to the vehicle, and there may also be a region that meets the distance requirement in the unextracted road segment connected to the current unextracted road segment. In this case, the unextracted road segment connected to the current unextracted road segment can be added to the array of road segments to be extracted, so as to extract enough regions of interest that meet the distance requirement and improve the accuracy of determining the region of interest.

[0094] The above steps S150 and S160 are executed cyclically until the regions of interest are extracted for all the road segments in the road segment array to be extracted.

[0095] S170: When there is no unextracted road section in the array of road sections to be extracted, obtaining each region of interest of the vehicle.

[0096] When there is no unextracted road section in the array of road sections to be extracted, it indicates that the road sections around the vehicle have been extracted, and the extraction of the regions of interest is completed. All the regions of interest extracted in the above step S160 can be used as the regions of interest of the vehicle.

[0097] From the above content, it can be seen that after locating the target section where the vehicle is currently located, this embodiment can filter out the vehicle's area of ​​interest based on the connection relationship between the sections and the distance between the boundary points of the sections and the vehicle. That is to say, some sections can be filtered out as areas of interest among all sections within the perception range, so that when performing environmental perception, only the perception information of the area of ​​interest can be calculated. Compared with calculating all the perception information, the complexity of the calculation can be reduced, that is, the difficulty of environmental perception can be reduced, and the accuracy of the environmental perception results can be improved.

[0098] It can be understood that in actual applications, different road section types have different effects on vehicle decision-making. For example, road sections of the road type usually have simpler road conditions, while intersection sections usually have more complex road conditions. The above road types are types where vehicles can only travel in one direction, and the above intersection types are other types other than road types, such as three-way intersections, crossroads, roundabouts, etc. Intersection types with complex road conditions have a greater impact on vehicle decision-making. Therefore, in embodiments of the present invention, different methods can be used to determine regions of interest for different types of road sections to improve the accuracy of determining regions of interest, thereby improving the accuracy of environmental perception.

[0099] As an implementation method of an embodiment of the present invention, the high-precision map also includes the type of each road section, and the type of each road section includes a road or an intersection. Figure 3 Another flowchart of a method for determining a region of interest provided by an embodiment of the present invention is provided, the method comprising the following steps:

[0100] S310: Determine the current position of the vehicle in the map coordinate system.

[0101] S320: Loading a high-precision map, where the high-precision map at least includes location information of each road section and connection relationships between the road sections.

[0102] S330: Determine the target road section where the vehicle is currently located based on the current position of the vehicle and the position information of each road section included in the high-precision map.

[0103] S340: searching for each road segment connected to the target road segment according to the connection relationship between each road segment, and adding each searched road segment to an array of road segments to be extracted.

[0104] S350: Detect whether there are unextracted road sections in the array of road sections to be extracted.

[0105] S360: When there are unextracted road sections in the array of road sections to be extracted, a currently unextracted road section is selected in the array of road sections to be extracted.

[0106] S370: When the current unextracted road section is a road, obtain the edge lines on both sides of the current unextracted road section, and determine the two endpoints within the preset search radius from the vehicle among the points included in each edge line as the area of ​​interest of the vehicle; and when the area of ​​interest is screened out, add the unextracted road section connected to the current unextracted road section to the array of sections to be extracted, and return to execute step S350.

[0107] In one implementation, the endpoint of each edge line can be retrieved using the following conditions. For any point included in an edge line, when the point meets any of the following conditions, the point is used as the area of ​​interest of the vehicle: the current distance between the point and the vehicle is less than or equal to a preset search radius, and the point is the endpoint of the edge line; the current distance between the point and the vehicle is less than or equal to the preset search radius, and there is an adjacent point of the point whose current distance to the vehicle is greater than the preset search radius.

[0108] Specifically, the coordinates of each point included in any edge line can be obtained. For the first point, its distance from the vehicle is calculated, and it is determined whether the distance is less than or equal to the search radius. If it is less than or equal to the search radius, it is directly used as the endpoint of the edge line. If it is greater, it is not used as the endpoint, and the judgment continues for the next point.

[0109] For each intermediate point, the following steps can be used to make a judgment: calculate the distance between the current intermediate point and the vehicle. If the distance between the current intermediate point and the vehicle is greater than the search radius, do not use it as an endpoint and continue to judge the next intermediate point; if the distance between the current intermediate point and the vehicle is less than or equal to the search radius, obtain the distance between the previous point and the vehicle. If the distance between the previous point and the vehicle is greater than the search radius, use the current intermediate point as the endpoint. If the distance between the previous point and the vehicle is less than or equal to the search radius, calculate the distance between the next point and the vehicle. If the distance between the next point and the vehicle is less than or equal to the search radius, do not use the current intermediate point as an endpoint. If the distance between the next point and the vehicle is greater than the search radius, use the current intermediate point as the endpoint.

[0110] For the last point, its distance to the vehicle can be calculated, and it can be determined whether the distance is less than or equal to the search radius. If it is less than or equal to, it is directly used as the endpoint of the edge line. If it is greater, it is not used as the endpoint.

[0111] like Figure 4 As shown in the figure, the circle is a circle drawn with the vehicle as the center according to the search radius. In the edge line on one side of the road where the vehicle is located, point 410 is the first point of the current edge line and is located within the search radius, so it is retained as the endpoint of the edge line; point 420 is within the search radius, and the next point of this point is not within the search radius, so this point is retained as another endpoint of the edge line. Point 430 is within the distance, but they are neither the first or last point, and the next point is not outside the search distance, so it is not an endpoint; point 440 is outside the distance and is discarded. Finally, the endpoints 410 and 420 of the edge line on this side are obtained. Using a similar method for the edge line on the other side, edge lines 450 and 460 can be obtained. In other words, for this road section, the final area of ​​interest is endpoints 410, 420, 450, and 460.

[0112] It can be understood that in practical applications, road-type sections may include straight roads and may also include curved roads. In an embodiment of the present invention, in order to simplify the algorithm complexity and improve the algorithm accuracy, when the current unextracted section is a road, when the edge lines on both sides of the current unextracted section are obtained, it can be determined whether the current unextracted section is a straight road; when the current unextracted section is a straight road, the edge lines on both sides of the straight road are obtained from the high-precision map as the edge lines on both sides of the current unextracted section; when the current unextracted section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and for each side of the edge line, it is divided into multiple straight segments, and the obtained multiple straight segments are used as the edge lines of that side of the current unextracted road.

[0113] In one implementation, the curvature of the currently unextracted road segment can be used to determine whether it is a straight road or a curved road. For example, a road with a curvature less than a certain value can be determined as a straight road, and a road with a curvature greater than a certain value can be determined as a curved road. The above-mentioned fixed value can be set according to demand, and the embodiment of the present invention does not specifically limit this. When dividing a curved road into multiple straight segments, for example, a section of road with a curvature within a certain range can be divided into a straight segment.

[0114] For a curved road, after it is divided into multiple straight sections, the region of interest of the curved road can be determined by using the region of interest determination method for straight sections, thereby reducing the difficulty of determining the region of interest of the curved road and improving the accuracy of determining the region of interest. In addition, for a curved section with a large curvature, dividing it into straight sections can simplify the road conditions and further reduce the difficulty of determining the region of interest.

[0115] S380: When the current unextracted road section is an intersection, all boundary points of the current unextracted road section are obtained, and the area of ​​interest of the vehicle is determined according to the distance between each boundary point and the vehicle; and when the area of ​​interest is screened out, the unextracted road section connected to the current unextracted road section is added to the array of sections to be extracted, and the execution of step S350 is returned.

[0116] For example, the distance between each boundary point included in the intersection and the vehicle can be calculated, and when the distance between the boundary points and the vehicle that meet a certain ratio is less than or equal to the search radius, all boundary points included in the intersection are determined as the area of ​​interest of the vehicle. The above ratio may include 40%, 50%, 60%, etc., and the embodiment of the present invention is not limited to this.

[0117] In one implementation, the electronic device may calculate the distance between each boundary point and the vehicle; when there is a boundary point whose distance to the vehicle is less than or equal to a preset search radius, all boundary points are regarded as the area of ​​interest of the vehicle.

[0118] That is to say, as long as there is a boundary point within the preset search radius, the intersection can be used as the vehicle's area of ​​interest to improve the accuracy of vehicle decision-making.

[0119] For intersection-type road sections, when there are boundary points within the vehicle search radius, all boundary points are regarded as regions of interest, that is, the information of the entire intersection can be processed as useful perception information, which can improve the accuracy of environmental perception and thus improve the safety of vehicle driving.

[0120] like Figure 5As shown, the vehicle is located at segment 2 (segment 2), segment 5 (segment 5) is a segment connected to segment 2, segment 5 is an intersection, and there are boundary points within the radius, so all boundary points of segment 5 can be used as the vehicle's area of ​​interest.

[0121] S390: When there is no unextracted road section in the array of road sections to be extracted, obtaining each region of interest of the vehicle.

[0122] The road conditions of road-type sections are usually simple. Therefore, for road-type sections, the area of ​​interest can be determined based on the edge lines on both sides to save computing resources. The road conditions of intersection-type sections are usually complex. Therefore, for intersection-type sections, the area of ​​interest can be determined based on all their boundary points to obtain more sufficient road information and improve the accuracy of environmental perception results.

[0123] like Figure 6 The figure shows the results of extracting the region of interest. The circle in the figure is a circle drawn with the vehicle as the center and according to the search radius. The vehicle is located in segment 2, and segments 5 and 6 are connected to segment 2.

[0124] The region of interest ROI-2 is retrieved in segment 2, which is composed of 4 asterisk boundary points. All boundary points of segment 6 are outside the search radius, so the region of interest of the vehicle is not retrieved in segment 6. Segment 5 is an intersection, and there are boundary points within the radius, so all boundary points of segment 5 can be used as the region of interest of the vehicle, that is, ROI-5, which is composed of multiple circular boundary points. In addition, after the region of interest of segment 5 is determined, segments 1, 3, and 4 (segment1, segment3, and segment4) connected to segment 5 can be added to the array of segments to be extracted, and the region of interest can be extracted for segments 1, 3, and 4 in turn.

[0125] For example, after extracting section 1, ROI-1 is obtained, which is composed of 4 asterisk boundary points. Section 7 (not shown in the figure) connected to section 1 can also be added to the array of sections to be extracted. When extracting section 7, all boundary points are outside the search radius, so no ROI is extracted from section 7. The same operation is performed on sections 3 and 4, with the same results.

[0126] The final regions of interest are ROI-2, ROI-5, ROI-1, ROI-3, and ROI-4. Among them, ROI-1, ROI-2, ROI-3, and ROI-4 are all composed of 4 asterisk boundary points, and ROI-5 is composed of multiple circular boundary points.

[0127] Figure 7 This is another schematic diagram of the extraction result of the region of interest. The circle in the figure is a circle drawn with the vehicle as the center and according to the search radius. The vehicle is located in segment 1 (segment1), and segment 5 (segment5) and segment 2 (segment2) are segments connected to segment 1.

[0128] The region of interest ROI-1 is retrieved in segment 1, which consists of 4 asterisk boundary points. All boundary points of segment 5 are outside the search radius, so the region of interest of the vehicle is not retrieved in segment 5. Segment 2 is an intersection, and there are boundary points within the radius, so all boundary points of segment 2 can be used as the region of interest of the vehicle, that is, ROI-2, which consists of multiple circular boundary points. In addition, after the region of interest of segment 2 is determined, segments 3 and 4 (segment3, segment4) connected to segment 2 can be added to the array of segments to be extracted, and the region of interest of segments 3 and 4 can be extracted in turn.

[0129] For example, after extracting the road section 3, the region of interest ROI-3 is obtained, which is composed of 4 asterisk boundary points. After extracting the road section 4, the region of interest ROI-4 is obtained, which is composed of 4 asterisk boundary points.

[0130] The regions of interest finally extracted are ROI-1, ROI-2, ROI-3, and ROI-4. Among them, ROI-1, ROI-3, and ROI-4 are all composed of 4 asterisk boundary points, and ROI-2 is composed of multiple circular boundary points.

[0131] As an implementation method of an embodiment of the present invention, after determining the area of ​​interest of the vehicle, the electronic device can also send the obtained information to the decision-making module and other modules that need to use the perception information, so that the module receiving the information can filter the information perceived by the sensor according to the area of ​​interest of the vehicle, and make decisions only based on the perception information in the area of ​​interest, thereby reducing the amount of calculation and improving the accuracy of the decision.

[0132] In one implementation, the electronic device may assemble and send data including information about each region of interest of the vehicle, wherein the information about each region of interest includes at least: the number of regions of interest, the current position of the vehicle, the number of boundary points of each region of interest, and the position information of each boundary point.

[0133] like Figure 8 As shown in the figure, after extracting each region of interest ROI-1-ROI-5 of the vehicle, the information shown in the figure can be constructed and sent to the decision module. The constructed information includes: information header and each region of interest, where the information header includes: the number of ROIs and the vehicle coordinates; the specific content in each ROI includes the number of points and the coordinates of each point. The number of points included in the region of interest is the number of boundary points.

[0134] Specifically, the number of regions of interest included in the information header is 5; the number of points included in ROI-1 is 4, the number of points included in ROI-2 is 4, the number of points included in ROI-3 is 4, the number of points included in ROI-4 is 4, and the number of points included in ROI-5 is 24.

[0135] The embodiment of the present invention provides a device for determining a region of interest, such as Fig. 9 As shown, the device comprises:

[0136] A current position determination module 910 is used to determine the current position of the vehicle in the map coordinate system;

[0137] A high-precision map loading module 920 is used to load a high-precision map, wherein the high-precision map includes at least the location information of each road section and the connection relationship between each road section;

[0138] The target road section determination module 930 is used to determine the target road section where the vehicle is currently located according to the current position of the vehicle and the position information of each road section included in the high-precision map;

[0139] A connection section search module 940 is used to search for sections connected to the target section according to the connection relationship between the sections, and add the target section and the searched sections to an array of sections to be extracted;

[0140] An unextracted road section detection module 950 is used to detect whether there is an unextracted road section in the array of road sections to be extracted;

[0141] A road segment detection module 960 is used for selecting a current unextracted road segment in the array of road segments to be extracted when the unextracted road segment detection module 850 determines that there are unextracted road segments in the array of road segments to be extracted, and screening target boundary points of the current unextracted road segment as the region of interest of the vehicle according to the distance between each boundary point and the vehicle among the boundary points of the current unextracted road segment; and when the region of interest is screened out, adding the unextracted road segment connected to the current unextracted road segment to the array of road segments to be extracted, and triggering the unextracted road segment detection module;

[0142] The region of interest determination module 970 is used to obtain the regions of interest of the vehicle when the unextracted road section detection module 850 determines that there is no unextracted road section in the array of road sections to be extracted.

[0143] Optionally, the high-precision map further includes the type of each road section, and the type of each road section includes a road or an intersection; the road section detection module 960 includes:

[0144] A road detection submodule, for obtaining edge lines on both sides of the current unextracted road section when the current unextracted road section is a road, and determining two endpoints within a preset search radius from the vehicle among the points included in each edge line as the region of interest of the vehicle;

[0145] The intersection detection submodule is used to obtain all boundary points of the current unextracted road section when the current unextracted road section is an intersection, and determine the area of ​​interest of the vehicle according to the distance between each boundary point and the vehicle.

[0146] Optionally, the road detection submodule is specifically used to: for a point included in any edge line, when the point meets any of the following conditions, use the point as the area of ​​interest of the vehicle: the current distance between the point and the vehicle is less than or equal to a preset search radius, and the point is the endpoint of the edge line; the current distance between the point and the vehicle is less than or equal to a preset search radius, and there is an adjacent point of the point whose current distance to the vehicle is greater than the preset search radius.

[0147] Optionally, the road detection submodule is specifically used for:

[0148] When the currently unextracted road section is a road, determining whether the currently unextracted road section is a straight road;

[0149] When the currently unextracted road section is a straight road, edge lines on both sides of the straight road are obtained from the high-precision map as edge lines on both sides of the currently unextracted road section;

[0150] When the currently unextracted road section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and the edge lines on each side are divided into multiple straight line segments, and the obtained multiple straight line segments are used as the edge lines of that side of the currently unextracted road.

[0151] Optionally, the intersection detection submodule is specifically used for:

[0152] Calculating the distance between each boundary point and the vehicle;

[0153] When there is a boundary point whose distance to the vehicle is less than or equal to the preset search radius, all boundary points are regarded as the region of interest of the vehicle.

[0154] Optionally, the device further comprises:

[0155] The information building module is used to build and send data including information of each area of ​​interest of the vehicle, wherein the information of each area of ​​interest includes at least: the number of areas of interest, the current position of the vehicle, the number of boundary points of each area of ​​interest, and the position information of each boundary point.

[0156] The device for determining the region of interest provided by the embodiment of the present invention can, after locating the target section where the vehicle is currently located, filter out the region of interest of the vehicle based on the connection relationship between the sections and the distance between the boundary points of the sections and the vehicle. That is to say, some sections can be filtered out as regions of interest from among all sections within the perception range, so that when performing environmental perception, only the perception information of the region of interest can be calculated. Compared with calculating all the perception information, the complexity of the calculation can be reduced, that is, the difficulty of environmental perception can be reduced, and the accuracy of the environmental perception results can be improved.

[0157] The above device embodiment corresponds to the method embodiment and has the same technical effect as the method embodiment. For specific description, please refer to the method embodiment. The device embodiment is obtained based on the method embodiment. For specific description, please refer to the method embodiment part, which will not be repeated here.

[0158] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0159] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining a region of interest, characterized in that: The method comprises: Determine the current position of the vehicle in the map coordinate system; Loading a high-precision map, wherein the high-precision map includes at least the location information of each road section and the connection relationship between the road sections; Determine the target road section where the vehicle is currently located according to the current position of the vehicle and the position information of each road section included in the high-precision map; According to the connection relationship between the road sections, searching for the road sections connected to the target road section, and adding the target road section and the searched road sections into the array of road sections to be extracted; Detecting whether there is an unextracted road section in the array of road sections to be extracted; When there are unextracted road sections in the array of road sections to be extracted, a current unextracted road section is selected from the array of road sections to be extracted, and among the boundary points of the current unextracted road section, target boundary points of the current unextracted road section are screened according to the distance between each boundary point and the vehicle as the region of interest of the vehicle; and when the region of interest is screened out, the unextracted road section connected to the current unextracted road section is added to the array of road sections to be extracted, and the step of detecting whether there are unextracted road sections in the array of road sections to be extracted is returned to execute; When there is no unextracted road section in the array of road sections to be extracted, obtaining each region of interest of the vehicle; The high-precision map further includes the type of each road section, and the type of each road section includes a road or an intersection; Among the boundary points of the currently unextracted road section, according to the distance between each boundary point and the vehicle, the target boundary points of the currently unextracted road section are screened as the region of interest of the vehicle, including: When the current unextracted road section is a road, edge lines on both sides of the current unextracted road section are obtained, and two endpoints whose distances from the vehicle are within a preset search radius are determined from points included in each edge line as regions of interest of the vehicle; When the current unextracted road section is an intersection, all boundary points of the current unextracted road section are acquired, and the region of interest of the vehicle is determined according to the distance between each boundary point and the vehicle.

2. The method according to claim 1, characterized in that The step of determining two endpoints whose distances from the vehicle are within a preset search radius from the points included in each edge line as the region of interest of the vehicle includes: For any point included in an edge line, when the point meets any of the following conditions, the point is taken as the area of ​​interest of the vehicle: the current distance between the point and the vehicle is less than or equal to the preset search radius, and the point is the endpoint of the edge line; the current distance between the point and the vehicle is less than or equal to the preset search radius, and there is an adjacent point of the point whose current distance to the vehicle is greater than the preset search radius.

3. The method according to claim 1, characterized in that When the current unextracted road section is a road, obtaining edge lines on both sides of the current unextracted road section includes: When the currently unextracted road section is a road, determining whether the currently unextracted road section is a straight road; When the currently unextracted road section is a straight road, edge lines on both sides of the straight road are obtained from the high-precision map as edge lines on both sides of the currently unextracted road section; When the currently unextracted road section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and the edge lines on each side are divided into multiple straight line segments, and the obtained multiple straight line segments are used as the edge lines of that side of the currently unextracted road.

4. The method according to claim 1, characterized in that Determining the region of interest of the vehicle according to the distance between each boundary point and the vehicle comprises: Calculating the distance between each boundary point and the vehicle; When there is a boundary point whose distance to the vehicle is less than or equal to the preset search radius, all boundary points are regarded as the region of interest of the vehicle.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Data including information of each region of interest of the vehicle is constructed and sent, wherein the information of each region of interest at least includes: the number of regions of interest, the current position of the vehicle, the number of boundary points of each region of interest, and the position information of each boundary point.

6. A device for determining a region of interest, characterized in that: The device comprises: A current position determination module, used to determine the current position of the vehicle in the map coordinate system; A high-precision map loading module is used to load a high-precision map, wherein the high-precision map includes at least the location information of each road section and the connection relationship between each road section; A target road section determination module, used to determine the target road section where the vehicle is currently located according to the current position of the vehicle and the position information of each road section included in the high-precision map; A connection section search module, used for searching for sections connected to the target section according to the connection relationship between the sections, and adding the target section and the searched sections into an array of sections to be extracted; An unextracted road section detection module is used to detect whether there is an unextracted road section in the array of road sections to be extracted; a road section detection module, configured to select a current unextracted road section from the array of road sections to be extracted when the unextracted road section detection module determines that there are unextracted road sections in the array of road sections to be extracted, and select target boundary points of the current unextracted road section from boundary points of the current unextracted road section according to the distance between each boundary point and the vehicle as the region of interest of the vehicle; and when the region of interest is screened out, add the unextracted road section connected to the current unextracted road section to the array of road sections to be extracted, and trigger the unextracted road section detection module; An area of ​​interest determination module, configured to obtain each area of ​​interest of the vehicle when the unextracted road section detection module determines that there is no unextracted road section in the array of road sections to be extracted; The high-precision map further includes the type of each road section, and the type of each road section includes a road or an intersection; the road section detection module includes: A road detection submodule, for obtaining edge lines on both sides of the current unextracted road section when the current unextracted road section is a road, and determining two endpoints within a preset search radius from the vehicle among the points included in each edge line as the region of interest of the vehicle; The intersection detection submodule is used to obtain all boundary points of the current unextracted road section when the current unextracted road section is an intersection, and determine the area of ​​interest of the vehicle according to the distance between each boundary point and the vehicle.

7. The device according to claim 6, characterized in that The road detection submodule is specifically configured to: for a point included in any edge line, when the point satisfies any of the following conditions, use the point as the region of interest of the vehicle: the current distance between the point and the vehicle is less than or equal to a preset search radius, and the point is an endpoint of the edge line; The current distance between the point and the vehicle is less than or equal to a preset search radius, and the current distance between an adjacent point of the point and the vehicle is greater than the preset search radius.

8. The device according to claim 6, characterized in that The road detection submodule is specifically used for: When the currently unextracted road section is a road, determining whether the currently unextracted road section is a straight road; When the currently unextracted road section is a straight road, edge lines on both sides of the straight road are obtained from the high-precision map as edge lines on both sides of the currently unextracted road section; When the currently unextracted road section is a curved road, the edge lines on both sides of the curved road are obtained from the high-precision map, and the edge lines on each side are divided into multiple straight line segments, and the obtained multiple straight line segments are used as the edge lines of that side of the currently unextracted road.

Images