Vehicle following method and device, storage medium, electronic equipment and vehicle

Abstract

The invention provides a vehicle following method and device, a storage medium, electronic equipment and a vehicle, and the method comprises the steps: obtaining multiple frames of visual target information, screening one frame of visual target information with the highest vehicle body overlapping rate from the multiple frames of visual target information, and taking the visual target information as candidate visual target information, the method comprises the steps of obtaining candidate visual target information corresponding to a front vehicle, obtaining radar target information corresponding to the front vehicle, screening out associated visual target information associated with the radar target information from multiple frames of visual target information, and if the candidate visual target information and the associated visual target information are the same visual target information, carrying out vehicle following according to the candidate visual target information. And if the candidate visual target information and the associated visual target information are different visual target information, performing vehicle following according to fusion target information obtained by fusing the radar target information and the associated visual target information. By adopting the method, the recognition accuracy of the front car following target can be improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle control technology, and in particular to a vehicle following method, device, storage medium, electronic equipment and vehicle. Background Art

[0002] Cruise following assistance is an important component of assisted driving technology. In vehicle following scenarios, accurate identification of the vehicle ahead by the ego vehicle is crucial for both safety and comfort.

[0003] To ensure accurate recognition of vehicles ahead, most related technologies use a combination of cameras and radar. This approach fuses the target acquisition results of the cameras and radar to ensure accurate recognition of vehicles ahead. However, in some scenarios where there are metal fences or metal obstacles on both sides of the road, when a vehicle is following a vehicle in a lane close to the roadside, the metal fences or obstacles can affect the radar's monitoring results, thereby affecting the recognition of the vehicle ahead. Summary of the Invention

[0004] Based on this, the present disclosure provides a vehicle following method, device, storage medium, electronic device and vehicle. By adopting this vehicle following method, the problem of inaccurate identification of the vehicle ahead caused by radar target information being affected by other environmental factors can be avoided while ensuring driving safety.

[0005] In one aspect, the present invention provides a vehicle following method, the method comprising:

[0006] Acquire multiple frames of visual target information corresponding to the vehicle ahead collected by a vehicle-mounted camera;

[0007] Screening out a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information, and using the frame of visual target information as candidate visual target information;

[0008] Acquiring radar target information corresponding to the vehicle ahead based on radar collection;

[0009] Filtering out associated visual target information associated with the radar target information from multiple frames of visual target information;

[0010] If the candidate visual target information and the associated visual target information are the same visual target information, performing vehicle following on the leading vehicle according to the candidate visual target information;

[0011] If the candidate visual target information and the associated visual target information are different visual target information, the vehicle ahead is followed according to fused target information obtained by fusing the radar target information and the associated visual target information.

[0012] Furthermore, in some embodiments, the visual target information includes first position information corresponding to the front vehicle;

[0013] The step of selecting a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information and using the frame of visual target information as candidate visual target information includes:

[0014] Calculating, based on the current position information of the self-vehicle and the first position information, a body overlap rate of the front vehicle and the self-vehicle in each visual target information;

[0015] The visual target information with the highest vehicle body overlap rate is determined as the candidate visual target information.

[0016] Furthermore, in some embodiments, each of the visual target information includes a unique first information identifier;

[0017] After selecting visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information and using the visual target information as candidate visual target information, the method further includes:

[0018] A unique second information identifier is generated for the candidate visual target information according to the first information identifier corresponding to the visual target information with the highest overlap rate.

[0019] Furthermore, in some embodiments, after screening out the associated visual target information associated with the radar target information from the multiple frames of visual target information, the method further includes:

[0020] Reversely generate the first information identifier corresponding to the visual target information with the highest overlap rate according to the second information identifier corresponding to the candidate visual target information;

[0021] If the first information identifier corresponding to the associated visual target information is the same as the first information identifier corresponding to the visual target information with the highest overlap rate, determining that the candidate visual target information and the associated visual target information are the same visual target information;

[0022] If the first information identifier corresponding to the associated visual target information is different from the first information identifier corresponding to the visual target information with the highest overlap rate, it is determined that the candidate visual target information and the associated visual target information are different visual target information.

[0023] Furthermore, in some embodiments, the visual target information includes first position information corresponding to the preceding vehicle, and the radar target information includes second position information corresponding to the preceding vehicle;

[0024] The step of screening out associated visual target information associated with the radar target information from multiple frames of visual target information includes:

[0025] Based on the first position information corresponding to each piece of visual target information, the visual target information closest to the second position information is screened out from multiple frames of visual target information, and the visual target information is used as the associated visual target information.

[0026] Furthermore, in some embodiments, obtaining radar target information corresponding to the vehicle ahead collected by radar includes:

[0027] Acquire multi-frame radar target information of the vehicle ahead based on radar collection;

[0028] The step of screening out associated visual target information associated with the radar target information from multiple frames of visual target information includes:

[0029] The multiple frames of radar target information and the multiple frames of visual target information are matched in pairs to filter out the most similar associated radar target information and associated visual target information.

[0030] Furthermore, in some embodiments, following the vehicle ahead according to the fused target information obtained by fusing the radar target information and the associated visual target information includes:

[0031] fusing the associated radar target information and the associated visual target information to obtain fused target information;

[0032] The vehicle ahead is followed according to the fused target information.

[0033] In another aspect, the present invention provides a vehicle following device, comprising:

[0034] A visual information acquisition module is used to acquire multiple frames of visual target information corresponding to the vehicle in front collected by the vehicle-mounted camera;

[0035] a candidate information screening module, configured to screen out a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information, and use the frame of visual target information as candidate visual target information;

[0036] A radar information acquisition module, configured to acquire radar target information corresponding to the vehicle ahead based on radar acquisition;

[0037] an associated information screening module, configured to screen out associated visual target information associated with the radar target information from multiple frames of visual target information;

[0038] a first vehicle following module, configured to follow the preceding vehicle according to the candidate visual target information if the candidate visual target information and the associated visual target information are the same visual target information;

[0039] The second vehicle following module is configured to follow the vehicle ahead according to fused target information obtained by fusing the radar target information and the associated visual target information if the candidate visual target information and the associated visual target information are different visual target information.

[0040] In another aspect, the present invention provides a storage medium storing a computer program, wherein the computer program is suitable for being loaded by a processor and executing the steps of the above method.

[0041] On the other hand, the present invention further provides an electronic device, comprising: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the steps of the above method.

[0042] On the other hand, the present invention further provides a vehicle comprising the above-mentioned vehicle following device or electronic device.

[0043] The vehicle following method provided by the present invention collects visual target information and radar target information of a leading vehicle respectively through a camera and a radar during the vehicle following process, and first determines a candidate visual target information from the collected multiple frames of visual target information, and then determines associated visual target information that is associated and fused with the radar target information in the multiple frames of visual target information based on the collected radar target information. When the candidate visual target information and the associated visual target information are the same visual target information, the leading vehicle is followed only based on the candidate visual target information, which can avoid the problem of inaccurate recognition of the leading vehicle caused by the radar being affected by other environmental factors. When the candidate visual target information and the associated visual target information are different visual target information, the leading vehicle is followed based on the fused target information obtained by fusing the radar target information and the associated visual target information, thereby ensuring driving safety.

[0044] It should be understood that the contents described in the Summary of the Invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of a vehicle following scenario provided by an embodiment of the present invention;

[0046] Figure 2 A schematic flow chart of a vehicle following method provided by an embodiment of the present invention;

[0047] Figure 3 A schematic structural diagram of a vehicle following device provided by an embodiment of the present invention;

[0048] Figure 4 A schematic structural diagram of a vehicle following device provided by an embodiment of the present invention;

[0049] Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] In the description of one or more embodiments of the present invention, the term "including" and similar terms should be understood as open inclusion, i.e., "including but not limited to." The term "based on" should be understood as "based at least in part on." The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment." The terms "first," "second," etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0052] In related technologies, in order to ensure the recognition accuracy of the vehicle in front, most of them adopt a front vehicle recognition solution that combines cameras and radars. That is, the target information of the vehicle in front is collected by cameras and radars at the same time, and the collected visual target information and radar target information are fused. The vehicle is followed according to the fused target recognition results to ensure the recognition accuracy of the vehicle in front.

[0053] However, in some scenarios with metal fences or metal obstacles on both sides of the road, when a vehicle is following a vehicle in a lane close to the roadside, the metal fences or obstacles on the roadside will affect the radar's detection results, causing the radar to identify the vehicle in front as closer to the ego vehicle. At the same time, when the camera is affected by environmental factors such as light, the identified target distance will also be less than the actual distance. When the radar is interfered with by the metal fence, causing the identified vehicle in front to be closer to the ego vehicle, and the distance between the ego vehicle and the vehicle in front identified by the camera is also less than the actual distance, the ego vehicle will obtain the recognition result that the ego vehicle is closer to the vehicle in front based on the fusion detection results of the camera and radar, so that the vehicle following system will trigger unnecessary braking and deceleration, affecting the user's driving experience and driving safety.

[0054] Based on this, the present invention proposes a vehicle following method. During the vehicle following process, visual target information and radar target information of a vehicle in front are collected respectively by a camera and a radar, and first a candidate visual target information is determined in the collected multiple frames of visual target information. Then, based on the collected radar target information, associated visual target information is determined in the multiple frames of visual target information and is associated and fused with the radar target information. When the candidate visual target information and the associated visual target information are the same visual target information, the vehicle in front is followed only based on the candidate visual target information. This can avoid the problem of inaccurate identification of the vehicle in front caused by the radar target information being affected by other environmental factors. When the candidate visual target information and the associated visual target information are different visual target information, the vehicle in front is followed based on the fused target information obtained by fusing the radar target information and the associated visual target information to ensure driving safety.

[0055] See Figure 1 , is a schematic diagram of a vehicle following scenario provided by an embodiment of the present invention. Figure 1 As shown, the figure includes an ego vehicle 01 and a front vehicle 02. Ego vehicle 01 follows front vehicle 02. During the following process, ego vehicle 01 uses its camera and radar to monitor the vehicle ahead in real time. This allows the vehicle following control system in the ego vehicle to perform following control based on the monitoring results. The detection results include visual target information collected by the camera and radar target information collected by the radar.

[0056] See Figure 2 , is a flow chart of a vehicle following method provided by an embodiment of the present invention. Figure 1 ,For the process shown in FIG. 2 ,a detailed explanation is given. The vehicle following method may ,concretely include the following steps:

[0057] Step S102, obtaining multi-frame visual target information corresponding to the vehicle ahead collected by the vehicle-mounted camera;

[0058] When the vehicle is in vehicle following driving mode, the ego vehicle will monitor the road information of the road ahead through the vehicle camera and radar, identify the position of the vehicle ahead in real time, and calculate the driving speed of the vehicle ahead based on the continuously identified position information, so that the ego vehicle can follow the vehicle based on this information.

[0059] In one or more embodiments of the present invention, when a vehicle is in a vehicle following driving mode, the ego vehicle periodically collects visual target information of the vehicle ahead based on an onboard camera. In one collection cycle, the onboard camera collects multiple frames of visual target information.

[0060] Among them, the on-board camera can be installed on the front body of the self-vehicle to collect visual target information in front of the self-vehicle. The visual targets include the vehicles in front of the self-vehicle in the same lane and adjacent lanes. The visual target information may include the type of the front vehicle, lateral and longitudinal positions, lateral and longitudinal speeds, lateral and longitudinal accelerations, vehicle length, width, height and vehicle heading angle.

[0061] Step S104, selecting a frame of visual target information having the highest overlap rate with the ego vehicle from the multiple frames of visual target information, and using the frame of visual target information as candidate visual target information;

[0062] In one or more embodiments of the present invention, after obtaining multiple frames of visual target information collected by the vehicle-mounted camera within one acquisition cycle, based on the information related to the vehicle in front carried in each visual target information, a frame of visual target information with the highest overlap rate with the body of the self-vehicle is screened out from the multiple frames of visual target information, and the visual target information is used as candidate visual target information.

[0063] It can be understood that vehicle following primarily involves automatically following the closest vehicle in front of the ego vehicle, with some overlap between the ego vehicle and the preceding vehicle. After obtaining multiple frames of visual target information, the visual target information with the highest overlap between the preceding vehicle and the ego vehicle is selected as the candidate visual target information based on the position information of the preceding vehicle carried in the visual target information. The candidate visual target information is visual target information determined purely through visual analysis and can be used by the ego vehicle for vehicle following.

[0064] In a feasible implementation manner, the visual target information includes the first position information corresponding to the front vehicle, and a frame of visual target information with the highest body overlap rate with the self-vehicle is screened out from multiple frames of visual target information, and the visual target information is used as candidate visual target information. It can be as follows: based on the current position information and the first position information of the self-vehicle, the body overlap rate of the front vehicle and the self-vehicle in each visual target information is calculated respectively, and the visual target information with the highest body overlap rate is determined as the candidate visual target information.

[0065] The higher the body overlap rate between the front vehicle and the self vehicle in the visual target information, the more accurate the determination of the front vehicle as the following target based on the visual target information, and the higher the following safety of the vehicle.

[0066] Step S106, obtaining radar target information corresponding to the vehicle ahead based on radar collection;

[0067] In one or more embodiments of the present invention, when a vehicle is in a vehicle-following driving mode, the ego vehicle periodically collects radar target information of the leading vehicle based on an onboard radar.

[0068] Among them, vehicle-mounted radars include but are not limited to lidar, ultrasonic radar, millimeter-wave radar, etc.

[0069] Radar target information may include the position information, size information, and type information of the vehicle ahead.

[0070] Step S108, filtering out associated visual target information associated with the radar target information from the multiple frames of visual target information;

[0071] In one or more embodiments of the present invention, based on a recognition solution that fuses camera and radar detection results, after acquiring multiple frames of visual target information and radar target information, a preset association algorithm is used to determine, within the multiple frames of visual target information, associated visual target information associated with the radar target information. The associated visual target information is then fused with the radar target information to generate fused target information, enabling a vehicle following system of the ego vehicle to follow the ego vehicle according to the fused target information.

[0072] It can be understood that the camera collects multiple frames of visual target information, and selects the most relevant visual target information with the radar target information from the multiple frames for fusion. Compared with directly collecting one frame of visual target information and fusing it, it can further improve the recognition accuracy of the vehicle in front.

[0073] Step S110: If the candidate visual target information and the associated visual target information are the same visual target information, then following the vehicle ahead is performed according to the candidate visual target information;

[0074] In one or more embodiments of the present invention, after determining the candidate visual target information and the associated visual target information, it is further determined whether the candidate visual target information and the associated visual target information are the same frame visual target information. If the candidate visual target information and the associated visual target information are the same frame visual target information, it indicates that this visual target information has been selected regardless of whether it is based on a pure vision solution or a vision + radar fusion solution. At this time, vehicle following is performed only based on the candidate visual target information, which can avoid the problem of inaccurate recognition of the vehicle in front caused by the radar being affected by surrounding environmental factors, and thus avoid the problem of unnecessary braking of the vehicle caused by incorrect target recognition, which can improve the user's driving experience and driving safety.

[0075] Step S112: If the candidate visual target information and the associated visual target information are different visual target information, the vehicle ahead is followed according to the fused target information obtained by fusing the radar target information and the associated visual target information.

[0076] In one or more embodiments of the present invention, after determining the candidate visual target information and the associated visual target information, a further determination is made as to whether the candidate visual target information and the associated visual target information are from the same frame. If the candidate visual target information and the associated visual target information are from different frames, this indicates that the associated visual target information selected based on the radar target information is different from the previously determined candidate visual target information. This may be due to inaccurate visual target information collected by the camera due to environmental interference. To fully ensure driving safety, the vehicle ahead is followed using the fused target information obtained by fusing the radar target information and the associated visual target information.

[0077] Based on the vehicle following method provided by the present invention, during the vehicle following process, visual target information and radar target information of the vehicle in front are respectively collected by a camera and a radar, and first, candidate visual target information is determined in the collected multiple frames of visual target information, and then associated visual target information that is associated and fused with the radar target information is determined in the multiple frames of visual target information based on the collected radar target information. When the candidate visual target information and the associated visual target information are the same visual target information, the vehicle in front is followed only based on the candidate visual target information, which can avoid the problem of inaccurate identification of the vehicle in front caused by the radar target information being affected by other environmental factors. When the candidate visual target information and the associated visual target information are different visual target information, the vehicle in front is followed based on the fused target information obtained by fusing the radar target information and the associated visual target information, thereby ensuring driving safety.

[0078] In one embodiment, the visual target information includes the first position information corresponding to the front vehicle. In step S104, a frame of visual target information with the highest body overlap rate with the self-vehicle is screened out from multiple frames of visual target information, and the visual target information is used as candidate visual target information. This can be done by: calculating the body overlap rate of the front vehicle and the self-vehicle in each visual target information based on the current position information and the first position information of the self-vehicle, and determining the visual target information with the highest body overlap rate as the candidate visual target information.

[0079] The first position information is the image of the vehicle ahead, captured by the ego vehicle's onboard camera. This information is calculated by constructing a coordinate system and then calculating the position of the vehicle ahead. The current position information is the ego vehicle's position within the same coordinate system. Based on the ego vehicle's current position information and the first position information of the vehicle ahead, it is possible to calculate whether there is overlap between the ego vehicle and the vehicle ahead, as well as the overlap ratio.

[0080] The higher the overlap between the front vehicle and the ego vehicle in the visual target information, the more accurately the front vehicle is identified as a following target based on that visual target information, and the safer the vehicle is when following the vehicle. By selecting the visual target information with the highest overlap as the candidate visual target information, and only using the candidate visual target information for vehicle following when the candidate visual target information and the associated visual target information are in the same frame, the radar can avoid inaccurate recognition of the front vehicle due to environmental factors, thereby avoiding unnecessary braking caused by incorrect target recognition, thereby improving the user's driving experience and driving safety.

[0081] In one embodiment, the visual target information further includes a unique first information identifier, and the first information identifier is used to distinguish multiple frames of visual target information.

[0082] A frame of visual target information with the highest overlap rate with the body of the self-vehicle is screened out from multiple frames of visual target information, and the visual target information is used as candidate visual target information. Then, a unique second information identifier is generated for the candidate visual target information based on the first information identifier corresponding to the visual target information with the highest overlap rate.

[0083] It should be noted that in actual application, it is necessary to distinguish multiple frames of visual target information, so each visual target information has a unique first information identifier. For example, the first information identifier can be a digital number. When there are 16 frames of visual target information, the first information identifier of each visual target information can be numbered 1 to 16. In actual application, the visual target acquisition controller controls the vehicle-mounted camera to collect visual target information. After collecting multiple frames of visual target information and determining candidate visual target information therefrom, the multiple frames of visual target information and the candidate visual target information are sent to the radar target acquisition controller via the CAN bus. The radar target acquisition controller is used to control the radar to collect radar target information. After receiving the multiple frames of visual target information and the candidate visual target information, the radar target acquisition controller screens the multiple frames of visual target information for associated visual target information based on the radar target information. To avoid screening the candidate visual target information together during the screening process, a unique second information identifier is generated for the candidate visual target information to avoid screening the candidate visual target information as associated visual target information. For example, the first information identifier of each visual target information can be numbered 1 to 16, and the second information identifier of the candidate visual target information can be 64 added to the first information identifier of the visual target information with the highest determined overlap rate, so as to distinguish the candidate visual target information from other multi-frame visual target information, thereby avoiding using the candidate visual target information for the functional execution target of the emergency braking function (the emergency braking function has high requirements on the attributes of the target, and the candidate visual target information collected only based on the camera does not meet the requirements of the emergency braking function), and when screening the associated visual target information, based on the difference between the candidate visual target information and other multi-frame visual target information, only the associated visual target information is screened in the remaining visual target information, avoiding screening the candidate visual target information as the associated visual target information (when the associated visual target information and the candidate visual target information are consistent, the vehicle follows the candidate visual target information as the target to improve the comfort of vehicle following).

[0084] Further explanation: When the ego vehicle is in follow driving mode, it monitors the road ahead using its camera and radar, identifies the position of the vehicle ahead in real time, and calculates the vehicle's speed based on this continuously identified position information, allowing the ego vehicle to follow the vehicle based on this information. In addition to the follow function, follow driving mode also includes an emergency braking function. This function is used to perform emergency braking based on forward target information collected by the vehicle's camera and radar to ensure vehicle driving safety. In follow driving mode, if the radar environment factors lead to inaccurate identification of the vehicle ahead, the vehicle will frequently accelerate and brake, affecting user comfort. One or more embodiments of the present invention address the aforementioned issues by selecting candidate visual target information from multiple frames of collected visual target information. When the candidate visual target information and the associated visual target information are identical, the vehicle following the preceding vehicle is performed solely based on the candidate visual target information. This avoids the aforementioned issues and improves driving comfort in the following driving mode. The emergency braking function, however, has a higher safety level than the following function. The emergency braking function is the baseline for driving safety, and relying solely on candidate visual target information cannot fully guarantee driving safety. Therefore, the emergency braking function still determines whether to perform emergency braking based on fused target information generated by fusing radar target information and associated visual target information. It is understood that visual target information is forward target information collected by the on-board camera, while radar target information is forward target information collected by the radar. The fused target information generated by fusing radar target information and associated visual target information is equivalent to the fusion of information collected by the radar and the camera, and has higher accuracy and reliability. The emergency braking function collects fused target information as the basis for determining whether to perform emergency braking, which can guarantee driving safety to a greater extent. By generating a unique second information identifier for the candidate visual target information, it is possible to prevent the emergency braking function from using the candidate visual target information as a target to perform emergency braking.

[0085] In one embodiment, after selecting associated visual target information associated with radar target information from multiple frames of visual target information, the following steps are further included:

[0086] First, based on the second information identifier corresponding to the candidate visual target information, the first information identifier corresponding to the visual target information with the highest overlap rate is reversely generated, and then the reversely generated first information identifier is compared with the first information identifier corresponding to the associated visual target information; if the first information identifier corresponding to the associated visual target information and the first information identifier corresponding to the visual target information with the highest overlap rate are the same, it is determined that the candidate visual target information and the associated visual target information are the same visual target information; if the first information identifier corresponding to the associated visual target information and the first information identifier corresponding to the visual target information with the highest overlap rate are different, it is determined that the candidate visual target information and the associated visual target information are different visual target information.

[0087] In one embodiment, the visual target information includes first position information corresponding to the vehicle in front, and the radar target information includes second position information corresponding to the vehicle in front. In step S108, the associated visual target information associated with the radar target information is filtered out from the multiple frames of visual target information. This can be done by filtering out the visual target information closest to the second position information from the multiple frames of visual target information based on the first position information corresponding to each visual target information, and using the visual target information as the associated visual target information.

[0088] In one embodiment, within one acquisition cycle, multiple frames of visual target information are collected based on the vehicle-mounted camera, and multiple frames of radar target information of the vehicle in front collected based on the radar are simultaneously obtained. The multiple frames of radar target information and the multiple frames of visual target information are matched pairwise to screen out the most similar associated radar target information and associated visual target information.

[0089] In a feasible implementation, the radar target information and the visual target information may be filtered based on the position information of the vehicle ahead, to obtain the associated radar target information and the associated visual target information that are closest to the position of the vehicle ahead.

[0090] Then, the associated radar target information and the associated visual target information are fused to obtain fused target information, so as to follow the vehicle in front according to the fused target information.

[0091] In this embodiment, the radar collects multiple frames of radar target information and selects one frame of related radar target information from the multiple frames of radar information. Compared with collecting only one frame of radar target information, the radar detection accuracy can be improved.

[0092] See Figure 3 , is a schematic diagram of the structure of a vehicle following device provided by an embodiment of the present invention. Figure 3As shown, the vehicle following device 1 can be implemented as all or part of an electronic device through software, hardware, or a combination of both. According to some embodiments, the vehicle following device 1 includes a visual information acquisition module 11, a candidate information screening module 12, a radar information acquisition module 13, an associated information screening module 14, a first vehicle following module 15, and a second vehicle following module 16, specifically including:

[0093] A visual information acquisition module 11 is used to acquire multiple frames of visual target information corresponding to the vehicle in front collected by a vehicle-mounted camera;

[0094] a candidate information screening module 12 for screening out a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information, and using the frame of visual target information as candidate visual target information;

[0095] A radar information acquisition module 13 is used to acquire radar target information corresponding to the vehicle ahead based on radar acquisition;

[0096] The associated information screening module 14 is configured to screen out associated visual target information associated with the radar target information from multiple frames of visual target information;

[0097] a first vehicle following module 15, configured to follow the vehicle ahead according to the candidate visual target information if the candidate visual target information and the associated visual target information are the same visual target information;

[0098] The second vehicle following module 16 is configured to follow the vehicle ahead according to fused target information obtained by fusing the radar target information and the associated visual target information if the candidate visual target information and the associated visual target information are different visual target information.

[0099] Optionally, the visual target information includes first position information corresponding to the front vehicle; the candidate information screening module 12 is specifically configured to:

[0100] Calculating, based on the current position information of the self-vehicle and the first position information, a body overlap rate of the front vehicle and the self-vehicle in each visual target information;

[0101] The visual target information with the highest vehicle body overlap rate is determined as the candidate visual target information.

[0102] Optionally, each visual target information includes a unique first information identifier; and the candidate information screening module 12 is further configured to:

[0103] A unique second information identifier is generated for the candidate visual target information according to the first information identifier corresponding to the visual target information with the highest overlap rate.

[0104] Optional, see Figure 4 The device further includes a visual information judgment module 17, which is used to:

[0105] Reversely generate the first information identifier corresponding to the visual target information with the highest overlap rate according to the second information identifier corresponding to the candidate visual target information;

[0106] If the first information identifier corresponding to the associated visual target information is the same as the first information identifier corresponding to the visual target information with the highest overlap rate, determining that the candidate visual target information and the associated visual target information are the same visual target information;

[0107] If the first information identifier corresponding to the associated visual target information is different from the first information identifier corresponding to the visual target information with the highest overlap rate, it is determined that the candidate visual target information and the associated visual target information are different visual target information.

[0108] Optionally, the visual target information includes first position information corresponding to the vehicle ahead, and the radar target information includes second position information corresponding to the vehicle ahead; the associated information screening module 14 is specifically configured to:

[0109] Based on the first position information corresponding to each piece of visual target information, the visual target information closest to the second position information is screened out from multiple frames of visual target information, and the visual target information is used as the associated visual target information.

[0110] Optionally, the radar information acquisition module 13 is specifically configured to:

[0111] Acquire multi-frame radar target information of the vehicle ahead based on radar collection;

[0112] The associated information screening module 14 is further configured to:

[0113] The multiple frames of radar target information and the multiple frames of visual target information are matched in pairs to filter out the most similar associated radar target information and associated visual target information.

[0114] The second vehicle following module 16 is specifically configured to:

[0115] fusing the associated radar target information and the associated visual target information to obtain fused target information;

[0116] The vehicle ahead is followed according to the fused target information.

[0117] The above-mentioned device embodiments correspond to the method embodiments. For detailed descriptions, please refer to the description of the method embodiments, which will not be repeated here. The device embodiments are obtained based on the corresponding method embodiments and have the same technical effects as the corresponding method embodiments. For detailed descriptions, please refer to the corresponding method embodiments.

[0118] The present invention also provides a storage medium that can store multiple instructions. The instructions are suitable for being loaded by a processor and executed by the vehicle following method of the above embodiments. The specific execution process can be found in the specific description of the above embodiments and will not be repeated here.

[0119] The present invention also provides a computer program product, which stores at least one instruction, and the at least one instruction is loaded by the processor and executes the vehicle following method as described in the above embodiments. The specific execution process can be found in the specific description of the above embodiments and will not be repeated here.

[0120] In one embodiment, the present invention further provides Figure 5 The structural diagram of the electronic device shown in FIG. Figure 5 At the hardware level, the electronic device includes a processor 31, an internal bus 32, a network interface 33, memory 34, and non-volatile storage 35. It may also include other hardware required for its operation. The electronic device can be installed in a vehicle, where the processor 31 reads the corresponding computer program from the non-volatile storage 35, stores it in the memory, and then executes it to implement the vehicle following method described above.

[0121] In one embodiment, the present invention further provides a vehicle, which may include the vehicle following device or electronic device as described above, so as to implement a vehicle following function by executing a vehicle following method through the vehicle following device or electronic device.

[0122] Finally, the various embodiments of the present invention are described in a progressive manner. Similar parts between the various embodiments can be referred to in detail. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiment is generally similar to the method embodiment, so its description is relatively simple. For relevant parts, refer to the description of the method embodiment.

[0123] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A vehicle following method, comprising: Obtain multi-frame visual target information corresponding to the vehicle in front captured by the on-board camera; Screening out a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information, and using the frame of visual target information as candidate visual target information; Acquiring radar target information corresponding to the vehicle ahead based on radar collection; Filtering out associated visual target information associated with the radar target information from multiple frames of visual target information; If the candidate visual target information and the associated visual target information are the same visual target information, performing vehicle following on the leading vehicle according to the candidate visual target information; If the candidate visual target information and the associated visual target information are different visual target information, the vehicle ahead is followed according to fused target information obtained by fusing the radar target information and the associated visual target information.

2. The method according to claim 1, wherein the visual target information includes first position information corresponding to the preceding vehicle; The step of selecting a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information and using the frame of visual target information as candidate visual target information includes: Calculating, based on the current position information of the self-vehicle and the first position information, a body overlap rate of the front vehicle and the self-vehicle in each visual target information; The visual target information with the highest vehicle body overlap rate is determined as the candidate visual target information.

3. The method according to claim 1, wherein each of the visual target information includes a unique first information identifier; After selecting visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information and using the visual target information as candidate visual target information, the method further includes: A unique second information identifier is generated for the candidate visual target information according to the first information identifier corresponding to the visual target information with the highest overlap rate.

4. The method according to claim 3, after screening out the associated visual target information associated with the radar target information from the multiple frames of visual target information, further comprising: Reversely generate the first information identifier corresponding to the visual target information with the highest overlap rate according to the second information identifier corresponding to the candidate visual target information; If the first information identifier corresponding to the associated visual target information is the same as the first information identifier corresponding to the visual target information with the highest overlap rate, determining that the candidate visual target information and the associated visual target information are the same visual target information; If the first information identifier corresponding to the associated visual target information is different from the first information identifier corresponding to the visual target information with the highest overlap rate, it is determined that the candidate visual target information and the associated visual target information are different visual target information.

5. The method according to claim 1, wherein the visual target information includes first position information corresponding to the preceding vehicle, and the radar target information includes second position information corresponding to the preceding vehicle; The step of screening out associated visual target information associated with the radar target information from multiple frames of visual target information includes: Based on the first position information corresponding to each piece of visual target information, the visual target information closest to the second position information is screened out from multiple frames of visual target information, and the visual target information is used as the associated visual target information.

6. The method according to claim 1, wherein obtaining radar target information corresponding to the vehicle ahead based on radar acquisition comprises: Acquire multi-frame radar target information of the vehicle ahead based on radar collection; The step of screening out associated visual target information associated with the radar target information from multiple frames of visual target information includes: The multiple frames of radar target information and the multiple frames of visual target information are matched in pairs to filter out the most similar associated radar target information and associated visual target information.

7. The method according to claim 6, wherein the following the vehicle ahead according to the fused target information obtained by fusing the radar target information and the associated visual target information comprises: fusing the associated radar target information and the associated visual target information to obtain fused target information; The vehicle ahead is followed according to the fused target information.

8. A vehicle following device, comprising: A visual information acquisition module is used to obtain multi-frame visual target information corresponding to the vehicle in front based on the information collected by the vehicle-mounted camera; a candidate information screening module, configured to screen out a frame of visual target information having the highest overlap rate with the body of the ego vehicle from the multiple frames of visual target information, and use the frame of visual target information as candidate visual target information; A radar information acquisition module, configured to acquire radar target information corresponding to the vehicle ahead based on radar acquisition; an associated information screening module, configured to screen out associated visual target information associated with the radar target information from multiple frames of visual target information; a first vehicle following module, configured to follow the preceding vehicle according to the candidate visual target information if the candidate visual target information and the associated visual target information are the same visual target information; The second vehicle following module is configured to follow the vehicle ahead according to fused target information obtained by fusing the radar target information and the associated visual target information if the candidate visual target information and the associated visual target information are different visual target information.

9. A storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. An electronic device comprising: A processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the steps of the method according to any one of claims 1 to 7.

11. A vehicle comprising the vehicle following device according to claim 8 or the electronic device according to claim 10.

Images