Vehicle lane change detection method, device, vehicle and storage medium

Abstract

Embodiments of the present invention provide a vehicle lane change detection method, device, vehicle and storage medium. The vehicle lane change detection method includes: acquiring a two-dimensional structure of a target vehicle in an image, and segmenting the two-dimensional structure to determine the tire landing point of the target vehicle; acquiring the distance between the tire landing point and the lane line based on the identified lane line and tire landing point; calculating the lane change weight of the target vehicle; and acquiring the lane change probability of the target vehicle based on the distance between the tire landing point and the lane line and the lane change weight. The embodiment of the present invention combines the lane change weight and the distance between the tire landing point and the lane line to obtain the lane change intention of the target vehicle. The lane change intention of the target vehicle obtained by using this solution has higher accuracy.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a vehicle lane change detection method, device, vehicle and storage medium. Background Art

[0002] Autonomous driving (also known as automated driving or driving automation) means that the vehicle continuously performs some or all dynamic driving tasks in an automatic manner. The "Automobile Automated Driving Classification" (GB / T 40429-2021) divides the levels of autonomous driving into 0 to 5, with Level 0 being emergency assisted driving and Level 5 being fully autonomous driving.

[0003] During the driving process of an autonomous vehicle, predicting the behavior of surrounding vehicles is an important manifestation of perception ability, and the lane change behavior of surrounding vehicles is an important factor affecting driving. Currently, the vehicle position is mainly obtained through lidar, combined with the lane line area of ​​the semantic map to predict whether the vehicle intends to change lanes. However, in the absence of a semantic map and when the vehicle turns, it is difficult for the current method to obtain accurate prediction results.

[0004] In the prior art, there are also solutions based on image analysis technology to identify whether a vehicle has the intention to change lanes. For example, in one prior method, the image captured by the vehicle-mounted camera is analyzed, and the lane line is identified from the captured image by using the characteristic that the lane line is white, and the wheel in the image is identified based on the image characteristics of the wheel, so as to determine the distance between the wheel and the lane line. When the change trend of this distance is shortening and the change length is greater than a preset length, it can be determined that the target vehicle has a tendency to change lanes.

[0005] However, in the above-mentioned method, the tire landing point is detected through the wheel frame, and due to the viewing angle problem, the landing point may have deviations. Therefore, the accuracy of the existing method needs to be improved. Summary of the invention

[0006] The main purpose of the present invention is to provide a vehicle lane change detection method, device and storage medium, which can improve the recognition accuracy of the vehicle lane change intention.

[0007] In order to achieve the above-mentioned object of the invention, the present invention provides a vehicle lane change detection method, the method comprising the following steps:

[0008] Acquire a two-dimensional structure of a target vehicle in an image, and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle;

[0009] According to the lane line and the tire landing point, obtaining the distance between the tire landing point and the lane line;

[0010] Calculating a lane change weight of the target vehicle;

[0011] The lane changing probability of the target vehicle is determined according to the distance between the tire landing point and the lane line and the lane changing weight.

[0012] Optionally, the calculating the lane change weight of the target vehicle comprises the following steps:

[0013] identifying a lane line type of the lane line, and acquiring a first lane change weight according to the lane line type; and / or,

[0014] Obtaining a road type where the target vehicle is located, and obtaining a second lane change weight according to the road type; and / or,

[0015] A third lane change weight is obtained according to whether the target vehicle is traveling in a tunnel; and / or,

[0016] obtaining a fourth lane change weight according to whether there is a moving vehicle within the specified range of the target vehicle;

[0017] The first lane change weight and / or the second lane change weight and / or the third lane change weight and / or the fourth lane change weight are accumulated to obtain the lane change weight of the target vehicle.

[0018] Optionally, extracting the two-dimensional structure of the target vehicle from the acquired image and segmenting the two-dimensional structure to determine the landing point of the tire of the target vehicle includes:

[0019] According to the three-dimensional structure of the target vehicle detected by the radar, obtaining the two-dimensional structure of the target vehicle from the image;

[0020] Inputting the two-dimensional structure into a segmentation model to obtain the coordinates of the contact points between the tires of the target vehicle and the ground; and

[0021] The coordinates of the contact point between the tire of the target vehicle and the ground are converted into the coordinates of the tire of the target vehicle in the image, and the coordinates are the landing point of the tire of the target vehicle.

[0022] Optionally, obtaining the distance between the tire landing point and the lane line according to the identified lane line and the tire landing point includes:

[0023] Mapping the lane line and the tire landing point to a world coordinate system;

[0024] The distance between the tire landing point and the lane line is obtained from the world coordinate system.

[0025] Optionally, the tire landing point includes a front wheel landing point and a rear wheel landing point, and the distance between the tire landing point and the lane line includes: the distance between the front wheel of the target vehicle and the lane line, and the distance between the rear wheel of the target vehicle and the lane line.

[0026] Optionally, determining the lane changing probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane changing weight comprises the following steps:

[0027] Inputting the distance between the tire landing point and the lane line into the network model to obtain the first lane change probability of the target vehicle;

[0028] Acquire a second lane-changing probability of the target vehicle according to the lane-changing weight;

[0029] The first lane change probability and the second lane change probability are multiplied to obtain the lane change probability of the target vehicle.

[0030] In addition, to achieve the above-mentioned purpose, the present invention also proposes a vehicle lane change detection device, comprising:

[0031] A landing point recognition module is used to obtain a two-dimensional structure of a target vehicle in an image and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle;

[0032] A distance calculation module, used for obtaining the distance between the tire landing point and the lane line according to the lane line and the tire landing point;

[0033] A weight calculation module, used to calculate the lane change weight of the target vehicle;

[0034] The lane change probability calculation module is used to determine the lane change probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane change weight.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also proposes a vehicle, which includes: a memory, a processor, and a vehicle lane change detection program stored in the memory and executable on the processor, wherein the vehicle lane change detection program is configured to implement the steps of any one of the vehicle lane change detection methods described above.

[0036] In addition, to achieve the above-mentioned purpose, the present invention also proposes a computer-readable storage medium storing a computer program, wherein the computer program can be executed by a processor to implement the steps of any vehicle lane change detection method described above.

[0037] Beneficial effects of the embodiments of the present invention:

[0038] In the embodiment of the present invention, the tire landing point of the target vehicle is determined based on segmentation of the two-dimensional structure of the target vehicle extracted from the image, so that the position information of the landing point can be directly obtained without processing the wheel frame; at the same time, the lane change weight is obtained according to the environment where the target vehicle is located, and the lane change intention of the target vehicle is obtained according to the lane change weight and the distance between the tire landing point and the lane line. The lane change intention of the target vehicle obtained by adopting this solution has higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic flow chart of a vehicle lane change detection method provided by the present invention.

[0040] Figure 2 A schematic diagram of a flow chart for obtaining the lane change weight of a target vehicle provided by the present invention.

[0041] Figure 3 A schematic diagram of a flow chart for determining a tire landing point of a target vehicle provided by the present invention.

[0042] Figure 4 A schematic diagram of a flow chart for obtaining the lane change probability of a target vehicle provided by the present invention.

[0043] Figure 5 It is a schematic structural diagram of an embodiment of a vehicle lane change detection device of the present invention.

[0044] Figure 6 It is a schematic diagram of the structure of a transportation tool of the hardware operating environment involved in the embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer and more understandable, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0046] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0047] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0048] The present invention is described below by way of embodiments with reference to the accompanying drawings.

[0049] like Figure 1FIG. 1 is a flow chart of an embodiment of a vehicle lane change detection method of the present invention. Figure 1 The scheme is used to identify whether the target vehicle has the intention to change lanes, such as whether the vehicle in the adjacent lane ahead has the intention to change lanes to the lane ahead, such as whether the vehicle on the right / left ahead has the intention to change lanes to the lane ahead. Specifically, Figure 1 The solution may include the following steps:

[0050] Step S1, obtaining a two-dimensional structure of a target vehicle in an image, and segmenting the two-dimensional structure to determine a landing point of a tire of the target vehicle.

[0051] The vehicle refers to the vehicle controlled by the autonomous driving system, which is equipped with software and hardware systems that support autonomous driving functions, such as autonomous driving operating systems, laser radars, cameras, etc. Autonomous driving here includes full autonomous driving and assisted driving, such as L0 to L5 autonomous driving.

[0052] In this step, the vehicle-mounted camera can be used to photograph the environment in which the vehicle is located, thereby obtaining an image of the environment.

[0053] There are many ways to identify lane lines. For example, the lane lines can be identified from the image by utilizing the characteristic that the lane lines are white as described in the background technology. However, the method of obtaining the lane line area through color information is very susceptible to the influence of other factors, and the lane lines fitted by this method are likely to differ greatly from the actual ones. Therefore, this embodiment preferably uses the Lane ATT network model to identify the lane lines in the image. Among them, the Lane ATT network model is a real-time, high-performance lane line detection algorithm model, which is implemented based on anchors and applies the attention mechanism. Of course, lane line detection schemes based on other network models are also applicable to this embodiment.

[0054] Among them, the target vehicle mainly refers to the vehicle located in the adjacent lane in front of the vehicle (including: left lane and / or right lane). These are the vehicles that need to be paid attention to in autonomous driving. In autonomous driving, the vehicle needs to pay attention to whether the vehicle in the adjacent lane in front has the intention to change lanes. Among them, there may be several target vehicles, for example, including vehicle A, vehicle B and vehicle C in the front left lane, etc. In this embodiment and other embodiments, one target vehicle is used as an example for illustration, but those skilled in the art know that this embodiment and other embodiments are also suitable for scenarios with multiple target vehicles.

[0055] In addition, radar can be used to lock the target vehicle. The radar installed on the vehicle can detect the position of other vehicles in a certain range (in front, behind, to the side, beside, etc.) and the corresponding distance. Combining this information, it can be determined which vehicles need special attention, and the target vehicle can be locked.

[0056] In this step, the tire landing point of the target vehicle is directly identified based on segmentation instead of using the wheel frame, thereby improving the accuracy of identification.

[0057] Specifically, Figure 3 As shown in FIG. 1 , it is a schematic diagram of an implementation method of step S1. Figure 3 As shown, in step S301, based on the three-dimensional structure of the target vehicle detected by the radar (i.e., the 3D image), the two-dimensional structure of the target vehicle, i.e., the 2D frame structure, is obtained from the image acquired by the autonomous driving vehicle. The 2D frame structure can be represented by (x, y).

[0058] In step S302, the two-dimensional structure is input into the segmentation model to obtain the coordinates (x, y coordinates) of the contact point between the tire of the target vehicle and the ground. Finally, in step S303, the coordinates of the tire of the target vehicle in the image (u=x+x', v=y+y') are obtained according to the coordinates of the contact point between the tire of the target vehicle and the ground. This coordinate is the identified landing point of the tire of the target vehicle.

[0059] The segmentation model used above may be, for example, a U-net network model. In addition, as shown above, the coordinates of the tire of the target vehicle in the image include: an x-coordinate and a y-coordinate. In the prior art, the method of identifying the landing point of the tire by the wheel frame usually only determines the y-coordinate without paying attention to the x-coordinate, while the tire landing point determined by the method of this embodiment includes the x-coordinate and the y-coordinate, so it will have higher accuracy.

[0060] Step S2: acquiring the distance between the tire landing point and the lane line according to the lane line and the tire landing point.

[0061] In one embodiment, the distance between the tire of the target vehicle and the lane line is determined as follows:

[0062] First, the target vehicle's tire landing point is mapped to the world coordinate system, and each point on the lane line is also mapped to the world coordinate system; then the distance between the target vehicle's tire and the lane line is obtained from the world coordinate system. In addition, the distance between the tire and the lane line includes: the distance d1 between the target vehicle's front wheel and the lane line, and the distance d2 between the rear wheel and the lane line. Specifically, since the tires and lane lines are on the ground, the tire points and lane line points of each vehicle can be mapped to the world coordinate system by combining the internal and external parameters of the camera and using the projection transformation formula.

[0063] Step S3: Calculate the lane change weight of the target vehicle.

[0064] When a vehicle is driving on different roads, the probability of lane change is different. That is, under different road conditions, the distance between the vehicle tire landing point and the lane line is the same, but the intention of lane change will be different. For example, when the distance between the vehicle tire landing point and the lane line is the same on the yellow solid line and the white dotted line, the lane change intention of the target vehicle will also be different. The lane change intention when driving on the white dotted lane line will be greater than the lane change intention when driving on the yellow solid lane line. Therefore, it is necessary to obtain the lane change weight of the vehicle in real time according to the actual situation of the vehicle. To obtain the lane change weight of the vehicle, refer to Figure 2 The process described.

[0065] Step S201: Identify the lane line type of the lane line, and obtain a first lane change weight according to the lane line type.

[0066] Perform image recognition on the lane lines in the acquired image to obtain the lane line type of the lane lines. According to the lane lines acquired in the image acquired by the autonomous driving vehicle, the acquired lane lines are then identified by type. When performing type identification, a network model can be used for identification, and the specific network model used is not limited by this technical solution. Lane line types include: white dashed line, white solid line, yellow dashed line, single yellow solid line, double yellow solid line, yellow dashed and solid line, double white dashed line, double white solid line. The specific lane line type can be set according to actual needs.

[0067] You can also obtain the current location information in real time through the positioning device of the autonomous vehicle, and then combine the distance and direction of the target vehicle with the current autonomous vehicle by combining the distance and direction of the target vehicle with the current autonomous vehicle through the location information of the current autonomous vehicle. Use the location information of the target vehicle to query the high-precision map in real time to obtain the lane line type of the current target vehicle.

[0068] Each lane line type corresponds to a lane change weight. The larger the lane change weight value, the greater the probability of the vehicle changing lanes when driving on this lane line. The lane change weight corresponding to each lane line type is shown in the following table:

[0069] Serial number Lane Type First lane change weight 1 White dotted line 50 2 Solid white line 40 3 Yellow dotted line 35 4 Solid yellow line 20 5 Double solid yellow line 10 6 Yellow dotted line 28 7 Double white dashed line 30 8 Double solid white line 25

[0070] The weight corresponding to each lane line type can be set according to actual needs, and this technical solution does not impose any restrictions.

[0071] Step S202: Obtain the type of road on which the target vehicle is located, and obtain a second lane change weight according to the road type.

[0072] The image acquired by the autonomous driving vehicle is processed to determine whether the road where the target vehicle is located is a straight road or a turning road. The network model is used for image recognition to determine whether the road where the target vehicle is located is a straight road or a turning road. The specific network model used is not limited in this technical solution.

[0073] You can also obtain the current location information in real time through the positioning device of the autonomous vehicle, and then combine the distance and direction of the target vehicle with the current autonomous vehicle by combining the distance and direction of the target vehicle with the current autonomous vehicle through the location information of the current autonomous vehicle. Use the location information of the target vehicle to query the high-precision map in real time to obtain whether the road where the current target vehicle is located is a straight road or a turning road.

[0074] Each road type corresponds to a lane change weight. The larger the lane change weight value, the greater the probability of the vehicle changing lanes when driving on this road. The lane change weight corresponding to each road type is shown in the following table:

[0075] Serial number Road Type Second lane change weight 1 straight road 50 2 Turning Road 30

[0076] The weight corresponding to each road type can be set according to actual needs, and this technical solution does not impose any restrictions.

[0077] Step S203: obtaining a third lane change weight according to whether the target vehicle is traveling in a tunnel.

[0078] The probability of a vehicle changing lanes when traveling in a tunnel is relatively low, so it is necessary to determine whether the current target vehicle is traveling in a tunnel.

[0079] The image acquired by the autonomous driving vehicle is processed to determine whether the road where the target vehicle is located is a tunnel. The network model is used for image recognition to determine whether the road where the target vehicle is located is a tunnel. The specific network model used is not limited in this technical solution.

[0080] The current location information can also be obtained in real time through the positioning device of the autonomous vehicle, and then the distance and direction between the target vehicle and the current autonomous vehicle can be obtained by combining the distance and direction between the target vehicle and the current autonomous vehicle through the location information of the current autonomous vehicle. The location information of the target vehicle can be used to query the high-precision map in real time to determine whether the road where the current target vehicle is located is a tunnel.

[0081] Whether the driving road is a tunnel or not, it corresponds to different lane change weights. The larger the lane change weight value, the greater the probability of the vehicle changing lanes when driving on this road. The lane change weights corresponding to whether the driving road is a tunnel are shown in the following table:

[0082] Serial number tunnel Third lane change weight 1 no 50 2 yes 20

[0083] The lane change weight corresponding to whether the driving road is a tunnel can be set according to actual needs, and this technical solution does not limit it.

[0084] Step 204: Obtain a fourth lane change weight according to whether there is a moving vehicle within the specified range of the target vehicle.

[0085] When there are vehicles driving within a certain range around the vehicle, the vehicle's lane change weight will also be reduced. By performing image processing on the images obtained by the autonomous driving vehicle, it is determined whether there are vehicles driving within a certain range around the target vehicle, such as within 30 meters on the left or right side of the vehicle. To determine whether there are vehicles driving within a certain range of the target vehicle, a network model, such as a machine learning network model, can be used for judgment. The specific network model used for judgment is not limited by this technical solution.

[0086] Whether there are vehicles traveling within a certain range of the target vehicle corresponds to different lane change weights. The larger the lane change weight value, the greater the probability of the vehicle changing lanes when driving in this situation. The lane change weights corresponding to whether there are vehicles traveling within a certain range of the target vehicle are shown in the following table:

[0087] Serial number Whether there are other vehicles running within a certain range of the target vehicle Fourth lane change weight 1 no 50 2 yes 30

[0088] The lane change weight corresponding to whether there are vehicles traveling within a certain range of the target vehicle can be set according to actual needs, and this technical solution does not limit it.

[0089] Step S205: Accumulate the first lane change weight and / or the second lane change weight and / or the third lane change weight and / or the fourth lane change weight to obtain the lane change weight of the target vehicle.

[0090] The various lane change weights of the target vehicle's environment are accumulated, as shown in the following table:

[0091]

[0092]

[0093] The lane change weight of the current vehicle is 140.

[0094] Step S4: determining the lane changing probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane changing weight.

[0095] The lane change probability of the target vehicle is obtained based on the lane change weight and the distance between the tire landing point and the lane line. Figure 4 The process shown.

[0096] Step S401: Input the distance between the tire landing point of the target vehicle and the lane line into the network model to obtain the first lane-changing probability of the target vehicle.

[0097] After obtaining the distance between the tire landing point of the target vehicle and the lane line, the lane-changing probability of the target vehicle can be determined based on the determined distance. For example, according to the distance d1 from the front wheel on the same side of the target vehicle to the lane line and the distance d2 from the rear wheel to the lane line, a time series model can be used to predict whether the target vehicle will change lanes. For instance, when the values of d1 and d2 are relatively small, the target vehicle is close to the lane line; if d1 is close to d2, it means the target vehicle is parallel to the lane line and the possibility of changing lanes is relatively low; if d1 << d2, at this time the vehicle body is tilted and there is a greater possibility of changing lanes; in terms of time series, the process of d1 and d2 continuously decreasing indicates that the vehicle is approaching the lane line and has a tendency to change lanes. Therefore, the lane-changing probability of the vehicle can be obtained based on the situation where d1 << d2 and d1 and d2 are continuously decreasing in time series. The above judgment process can be implemented with the help of a Long Short-Term Memory (LSTM) model. For example, input the distance between the tire of the target vehicle and the lane line into the LSTM model, and then obtain the judgment result of the lane-changing probability of the target vehicle from the output of the LSTM model. For example, the obtained second lane-changing probability is 80%.

[0098] Step S402: Obtain the second lane-changing probability of the target vehicle according to the lane-changing weight.

[0099] Based on the driving state of the target vehicle, the lane-changing weight of the target vehicle is obtained. For example, the lane-changing weight is 140. Then compare this lane-changing weight with the sum of the maximum values of the first lane-changing weight, the second lane-changing weight, the third lane-changing weight, and the fourth lane-changing weight to obtain the third lane-changing probability. For example, 140 / (50 + 50 + 50 + 50) = 70%.

[0100] Step S403: Multiply the first lane-changing probability and the second lane-changing probability to obtain the lane-changing probability of the target vehicle.

[0101] Multiply the second lane-changing probability obtained based on the distance between the tire of the target vehicle and the lane line and the third lane-changing probability obtained based on the lane-changing weight to obtain the first lane-changing probability of the target vehicle. For example, if the second lane-changing probability is 80% and the third lane-changing probability is 70%, then the first lane-changing probability = 80% * 70% = 56%.

[0102] After obtaining the above lane-changing probability, the above lane-changing probability can be used when performing autonomous driving operations. For example, if it is recognized that the target vehicle intends to change into this lane, then control the vehicle to decelerate to maintain a reasonable distance from the target vehicle.

[0103] Through the embodiment of the present invention, the tire landing point of the target vehicle is determined based on segmentation of the two-dimensional structure of the target vehicle extracted from the image, so that the position information of the landing point can be directly obtained without processing the wheel frame; at the same time, the lane change weight is obtained according to the environment where the target vehicle is located, and the lane change intention of the target vehicle is obtained according to the lane change weight and the distance between the tire landing point and the lane line. The lane change intention of the target vehicle obtained by using this solution has higher accuracy.

[0104] In addition, an embodiment of the present invention further provides a vehicle lane change detection device, which corresponds to Figure 1 The embodiment shown. Figure 5 As shown, the vehicle lane change detection device includes:

[0105] A landing point recognition module 10 is used to obtain a two-dimensional structure of a target vehicle in an image and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle;

[0106] A distance calculation module 20, configured to obtain the distance between the tire landing point and the lane line according to the lane line and the tire landing point;

[0107] A weight calculation module 30, used to calculate the lane change weight of the target vehicle;

[0108] The lane change probability calculation module 40 is used to determine the lane change probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane change weight.

[0109] Through the embodiment of the present invention, the tire landing point of the target vehicle is determined based on segmentation of the two-dimensional structure of the target vehicle extracted from the image, so that the position information of the landing point can be directly obtained without processing the wheel frame; at the same time, the lane change weight is obtained according to the environment where the target vehicle is located, and the lane change intention of the target vehicle is obtained according to the lane change weight and the distance between the tire landing point and the lane line. The lane change intention of the target vehicle obtained by using this solution has higher accuracy.

[0110] Reference Figure 6 , Figure 6 A schematic diagram of the structure of a vehicle in the hardware operating environment involved in an embodiment of the present invention.

[0111] like Figure 6As shown, the vehicle may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as WI-FI, 4G, 5G interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0112] Those skilled in the art will understand that Figure 6 The structure shown in the figure does not constitute a limitation on the vehicle, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0113] like Figure 6 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a vehicle lane change detection program.

[0114] exist Figure 6 In the vehicle shown, the network interface 1004 is mainly used for data communication with an external network; the user interface 1003 is mainly used for receiving user input instructions; the vehicle calls the vehicle lane change detection program stored in the memory 1005 through the processor 1001, and performs the following operations:

[0115] Acquire a two-dimensional structure of a target vehicle in an image, and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle;

[0116] According to the lane line and the tire landing point, obtaining the distance between the tire landing point and the lane line;

[0117] Calculating a lane change weight of the target vehicle;

[0118] The lane changing probability of the target vehicle is determined according to the distance between the tire landing point and the lane line and the lane changing weight.

[0119] Optionally, the calculating the lane change weight of the target vehicle comprises the following steps:

[0120] identifying a lane line type of the lane line, and acquiring a first lane change weight according to the lane line type; and / or,

[0121] Obtaining a road type where the target vehicle is located, and obtaining a second lane change weight according to the road type; and / or,

[0122] A third lane change weight is obtained according to whether the target vehicle is traveling in a tunnel; and / or,

[0123] obtaining a fourth lane change weight according to whether there is a moving vehicle within the specified range of the target vehicle;

[0124] The first lane change weight and / or the second lane change weight and / or the third lane change weight and / or the fourth lane change weight are accumulated to obtain the lane change weight of the target vehicle.

[0125] Optionally, extracting the two-dimensional structure of the target vehicle from the acquired image and segmenting the two-dimensional structure to determine the landing point of the tire of the target vehicle includes:

[0126] According to the three-dimensional structure of the target vehicle detected by the radar, obtaining the two-dimensional structure of the target vehicle from the image;

[0127] Inputting the two-dimensional structure into a segmentation model to obtain the coordinates of the contact points between the tires of the target vehicle and the ground; and

[0128] The coordinates of the contact point between the tire of the target vehicle and the ground are converted into the coordinates of the tire of the target vehicle in the image, and the coordinates are the landing point of the tire of the target vehicle.

[0129] Optionally, obtaining the distance between the tire landing point and the lane line according to the identified lane line and the tire landing point includes:

[0130] Mapping the lane line and the tire landing point to a world coordinate system;

[0131] The distance between the tire landing point and the lane line is obtained from the world coordinate system.

[0132] Optionally, the tire landing point includes a front wheel landing point and a rear wheel landing point, and the distance between the tire landing point and the lane line includes: the distance between the front wheel of the target vehicle and the lane line, and the distance between the rear wheel of the target vehicle and the lane line.

[0133] Optionally, determining the lane changing probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane changing weight comprises the following steps:

[0134] Inputting the distance between the tire landing point and the lane line into the network model to obtain the first lane change probability of the target vehicle;

[0135] Acquire a second lane-changing probability of the target vehicle according to the lane-changing weight;

[0136] The first lane change probability and the second lane change probability are multiplied to obtain the lane change probability of the target vehicle.

[0137] Through the embodiment of the present invention, the tire landing point of the target vehicle is determined based on segmentation of the two-dimensional structure of the target vehicle extracted from the image, so that the position information of the landing point can be directly obtained without processing the wheel frame; at the same time, the lane change weight is obtained according to the environment where the target vehicle is located, and the lane change intention of the target vehicle is obtained according to the lane change weight and the distance between the tire landing point and the lane line. The lane change intention of the target vehicle obtained by using this solution has higher accuracy.

[0138] In addition, an embodiment of the present invention further provides a computer-readable storage medium, on which a vehicle lane change detection program is stored. When the vehicle lane change detection program is executed by a processor, the following operations are implemented:

[0139] Acquire a two-dimensional structure of a target vehicle in an image, and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle;

[0140] According to the lane line and the tire landing point, obtaining the distance between the tire landing point and the lane line;

[0141] Calculating a lane change weight of the target vehicle;

[0142] The lane changing probability of the target vehicle is determined according to the distance between the tire landing point and the lane line and the lane changing weight.

[0143] Optionally, the calculating the lane change weight of the target vehicle comprises the following steps:

[0144] identifying a lane line type of the lane line, and acquiring a first lane change weight according to the lane line type; and / or,

[0145] Obtaining a road type where the target vehicle is located, and obtaining a second lane change weight according to the road type; and / or,

[0146] A third lane change weight is obtained according to whether the target vehicle is traveling in a tunnel; and / or,

[0147] obtaining a fourth lane change weight according to whether there is a moving vehicle within the specified range of the target vehicle;

[0148] The first lane change weight and / or the second lane change weight and / or the third lane change weight and / or the fourth lane change weight are accumulated to obtain the lane change weight of the target vehicle.

[0149] Optionally, extracting the two-dimensional structure of the target vehicle from the acquired image and segmenting the two-dimensional structure to determine the landing point of the tire of the target vehicle includes:

[0150] According to the three-dimensional structure of the target vehicle detected by the radar, obtaining the two-dimensional structure of the target vehicle from the image;

[0151] Inputting the two-dimensional structure into a segmentation model to obtain the coordinates of the contact points between the tires of the target vehicle and the ground; and

[0152] The coordinates of the contact point between the tire of the target vehicle and the ground are converted into the coordinates of the tire of the target vehicle in the image, and the coordinates are the landing point of the tire of the target vehicle.

[0153] Optionally, obtaining the distance between the tire landing point and the lane line according to the identified lane line and the tire landing point includes:

[0154] Mapping the lane line and the tire landing point to a world coordinate system;

[0155] The distance between the tire landing point and the lane line is obtained from the world coordinate system.

[0156] Optionally, the tire landing point includes a front wheel landing point and a rear wheel landing point, and the distance between the tire landing point and the lane line includes: the distance between the front wheel of the target vehicle and the lane line, and the distance between the rear wheel of the target vehicle and the lane line.

[0157] Optionally, determining the lane changing probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane changing weight comprises the following steps:

[0158] Inputting the distance between the tire landing point and the lane line into the network model to obtain the first lane change probability of the target vehicle;

[0159] Acquire a second lane-changing probability of the target vehicle according to the lane-changing weight;

[0160] The first lane change probability and the second lane change probability are multiplied to obtain the lane change probability of the target vehicle.

[0161] Through the embodiment of the present invention, the tire landing point of the target vehicle is determined based on segmentation of the two-dimensional structure of the target vehicle extracted from the image, so that the position information of the landing point can be directly obtained without processing the wheel frame; at the same time, the lane change weight is obtained according to the environment where the target vehicle is located, and the lane change intention of the target vehicle is obtained according to the lane change weight and the distance between the tire landing point and the lane line. The lane change intention of the target vehicle obtained by using this solution has higher accuracy.

[0162] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0163] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0164] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, controller, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0165] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A vehicle lane change detection method, characterized in that: include: Acquire a two-dimensional structure of a target vehicle in an image, and segment the two-dimensional structure to determine a landing point of a tire of the target vehicle; According to the lane line and the tire landing point, obtaining the distance between the tire landing point and the lane line; Calculating a lane change weight of the target vehicle; Determining the lane change probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane change weight; Wherein, determining the lane changing probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane changing weight comprises the following steps: Inputting the distance between the tire landing point and the lane line into the network model to obtain the first lane change probability of the target vehicle; Acquire a second lane-changing probability of the target vehicle according to the lane-changing weight; Multiplying the first lane change probability and the second lane change probability to obtain a lane change probability of the target vehicle; The step of extracting a two-dimensional structure of a target vehicle from an acquired image and segmenting the two-dimensional structure to determine a landing point of a tire of the target vehicle includes: According to the three-dimensional structure of the target vehicle detected by the radar, obtaining the two-dimensional structure of the target vehicle from the image; Inputting the two-dimensional structure into a segmentation model to obtain the coordinates of the contact point between the tire of the target vehicle and the ground, including an x-coordinate and a y-coordinate; and The coordinates of the contact point between the tire of the target vehicle and the ground are converted into the coordinates of the tire of the target vehicle in the image, and the coordinates are the landing point of the tire of the target vehicle.

2. The vehicle lane change detection method according to claim 1, characterized in that: The calculating of the lane change weight of the target vehicle comprises the following steps: identifying a lane line type of the lane line, and acquiring a first lane change weight according to the lane line type; and / or, Obtaining a road type where the target vehicle is located, and obtaining a second lane change weight according to the road type; and / or, A third lane change weight is obtained according to whether the target vehicle is traveling in a tunnel; and / or, obtaining a fourth lane change weight according to whether there is a moving vehicle within the specified range of the target vehicle; The first lane change weight and / or the second lane change weight and / or the third lane change weight and / or the fourth lane change weight are accumulated to obtain the lane change weight of the target vehicle.

3. The vehicle lane change detection method according to claim 1, characterized in that: According to the identified lane line and the tire landing point, obtaining the distance between the tire landing point and the lane line includes: Mapping the lane line and the tire landing point to a world coordinate system; The distance between the tire landing point and the lane line is obtained from the world coordinate system.

4. The vehicle lane change detection method according to claim 3, characterized in that: The tire landing point includes a front wheel landing point and a rear wheel landing point, and the distance between the tire landing point and the lane line includes: the distance between the front wheel of the target vehicle and the lane line, and the distance between the rear wheel of the target vehicle and the lane line.

5. A vehicle lane change detection device, characterized in that: include: The landing point recognition module is used to obtain the two-dimensional structure of the target vehicle in the image and segment the two-dimensional structure to determine the landing point of the tire of the target vehicle, including: According to the three-dimensional structure of the target vehicle detected by the radar, obtaining the two-dimensional structure of the target vehicle from the image; Inputting the two-dimensional structure into a segmentation model to obtain the coordinates of the contact point between the tire of the target vehicle and the ground, including an x-coordinate and a y-coordinate; and Convert the coordinates of the contact point between the tire of the target vehicle and the ground into the coordinates of the tire of the target vehicle in the image, which is the landing point of the tire of the target vehicle; A distance calculation module, used for obtaining the distance between the tire landing point and the lane line according to the lane line and the tire landing point; A weight calculation module, used to calculate the lane change weight of the target vehicle; A lane change probability calculation module, used to determine the lane change probability of the target vehicle according to the distance between the tire landing point and the lane line and the lane change weight, including: Inputting the distance between the tire landing point and the lane line into the network model to obtain the first lane change probability of the target vehicle; Acquire a second lane-changing probability of the target vehicle according to the lane-changing weight; The first lane change probability and the second lane change probability are multiplied to obtain the lane change probability of the target vehicle.

6. A means of transport, characterized in that: The vehicle comprises: a memory, a processor, and a vehicle lane change detection program stored in the memory and executable on the processor, wherein the vehicle lane change detection program is configured to implement the steps of the vehicle lane change detection method according to any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that: The computer program can be executed by a processor to implement the steps of the vehicle lane change detection method according to any one of claims 1 to 4.

Images