Signal lamp prompt information generation method and device, vehicle and storage medium

Abstract

The invention provides a signal lamp prompt information generation method and device, a vehicle and a storage medium, and relates to the technical field of vehicles, and the method comprises the steps: determining a target passing intersection of the vehicle according to the current vehicle position, and obtaining the current signal lamp pre-estimation information of each driving direction corresponding to the target passing intersection; extracting current signal lamp identification information from the intersection direction image information by using a preset identification model; when a preset deviation correction condition is satisfied, correcting the current signal lamp pre-estimation information based on the current signal lamp identification information; and predicting a target driving direction of the vehicle at the target traffic intersection based on the current navigation system state and a preset strategy, and generating signal lamp prompt information according to the current signal lamp pre-estimation information corresponding to the target driving direction. The comprehensiveness and accuracy of the signal lamp prompt information can be ensured, the demand direction of the user for the signal lamp information is accurately identified, and the data accuracy of the signal lamp prompt information and the demand direction identification accuracy are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicles, and in particular, to a method and device for generating signal light prompt information, a vehicle, and a storage medium. Background Art

[0002] In the traffic system, as an important traffic management facility, signal lights not only play a role in maintaining road order, but also are an important mechanism for ensuring driving safety and reducing the incidence of traffic accidents. With the continuous development of vehicle intelligence, media such as vehicle center control screens provide a new platform for the transmission of signal light information, which has an important value in ensuring traffic safety and improving the driving experience.

[0003] However, in related technologies, the signal light information displayed by vehicles for drivers often has inaccurate problems. Summary of the Invention

[0004] The problem solved by the present invention is how to improve the accuracy of signal light prompt information.

[0005] To solve the above problems, the present invention provides a method for generating signal light prompt information, including:

[0006] Determine the target passing intersection of the vehicle according to the acquired current vehicle position, and acquire the current signal light prediction information of each drivable direction corresponding to the target passing intersection;

[0007] Collect intersection direction image information, and extract the current signal light recognition information of each drivable direction from the intersection direction image information by using a preset recognition model;

[0008] When the current signal light prediction information and the current signal light recognition information meet a preset deviation correction condition, correct the current signal light prediction information based on the current signal light recognition information;

[0009] Predict the target driving direction of the vehicle at the target passing intersection based on the acquired current navigation system state and a preset strategy, and generate signal light prompt information according to the current signal light prediction information corresponding to the target driving direction.

[0010] Optionally, the preset deviation correction condition includes:

[0011] It is detected that there is a signal light color deviation between the current signal light recognition information and the current signal light prediction information corresponding to the drivable direction, and / or it is detected that there is a signal light countdown deviation between the current signal light recognition information and the current signal light prediction information corresponding to the drivable direction, and the signal light countdown deviation is greater than a preset deviation threshold.

[0012] Optionally, after correcting the current traffic signal prediction information based on the current traffic signal recognition information, the following steps are further included:

[0013] When the number of times that the traffic signal countdown deviation corresponding to the drivable direction is greater than the preset deviation threshold exceeds the first preset number of times, the target passing intersection corresponding to the drivable direction is set as the set intersection, and each traffic signal countdown deviation corresponding to the drivable direction is used as the historical deviation;

[0014] When it is determined according to the current vehicle position that the target passing intersection is switched to the set intersection and each historical deviation meets the preset consistency condition, the currently obtained traffic signal prediction information is corrected based on the historical deviation.

[0015] Optionally, the preset consistency condition includes that each historical deviation is positive or negative; the currently obtained traffic signal prediction information corrected based on the historical deviation includes:

[0016] Adding the traffic signal countdown corresponding to the currently obtained traffic signal prediction information to the historical deviation closest to the current time to obtain the corrected currently obtained traffic signal prediction information.

[0017] Optionally, after correcting the current traffic signal prediction information based on the current traffic signal recognition information, the following steps are further included:

[0018] When the number of times that the current traffic signal prediction information does not meet the preset deviation correction condition exceeds the second preset number of times, the collection of the intersection direction image information is stopped until it is detected that the traffic signal prompt information meets the preset untrusted condition; wherein, the preset untrusted condition includes that when the traffic signal color corresponding to the traffic signal prompt information is the preset prohibited passing color, it is detected that the target passing intersection has switched.

[0019] Optionally, the current navigation system state includes a navigation state or a cruise state; predicting the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and preset strategy includes:

[0020] When the current navigation system state is the navigation state, predicting the target driving direction according to the obtained current navigation information;

[0021] When the current navigation system state is the cruise state, determining the current distance between the vehicle and the target passing intersection according to the current vehicle position, and predicting the target driving direction based on the current distance.

[0022] Optionally, predicting the target driving direction based on the current distance includes:

[0023] When the current distance is greater than a preset distance threshold, predicting a target driving lane corresponding to the vehicle based on the acquired current vehicle operation information, and predicting the target driving direction according to the target driving lane;

[0024] When the current distance is less than or equal to the preset distance threshold, identifying a current driving lane corresponding to the vehicle based on the intersection direction image information and / or the current position information, and predicting the target driving direction according to the current driving lane.

[0025] In the present invention, determining the target passing intersection of the vehicle according to the current vehicle position is beneficial to grasping the position of the signal lamp device affecting the vehicle passing, so that the current signal lamp estimation information of each drivable direction corresponding to the obtained target passing intersection can provide a data basis for the generation of subsequent signal lamp prompt information. Since the obtained current signal lamp estimation information is restricted by factors such as computing power and lack of data sample size, the estimation result may be incorrect. In addition to obtaining the current signal lamp estimation information in the present invention, a preset recognition model is also used to extract the current signal lamp recognition information of each drivable direction from the intersection direction image information actually collected by the vehicle, providing an effective reference basis for the accuracy evaluation of the current signal lamp estimation information. The present invention realizes the accuracy evaluation of the current signal lamp estimation information by judging whether the signal lamp estimation information and the current signal lamp recognition information meet the preset deviation correction condition. When the preset correction condition is met, it indicates that the current signal lamp estimation information does not match the actual signal lamp situation, and it is necessary to correct the current signal lamp estimation information based on the current signal lamp recognition information to ensure the accuracy of the subsequent signal lamp prompt information. On this basis, based on the obtained current navigation system state and preset strategy, the target driving direction of the vehicle at the target passing intersection is predicted, and the signal lamp prompt information is generated according to the current signal lamp estimation information corresponding to the target driving direction. Whether the current navigation system state is in the navigation state or in the cruise state, accurate signal lamp prompt information can be provided for the user. On the one hand, the current signal lamp estimation information can make up for the limitations of the current signal lamp recognition information in terms of distance, space, etc. (such as in the case of long distance, space occlusion, bad weather, etc., it is difficult to extract the current signal lamp recognition information from the intersection direction image information), ensuring the timeliness and comprehensiveness of the signal lamp prompt information. On the other hand, the current signal lamp recognition information extracted based on the intersection direction image information can make up for the deficiency of the current signal lamp estimation information in terms of accuracy. In this way, by obtaining multi-channel signal lamp information and presetting deviation correction conditions in the present invention, the comprehensiveness and accuracy of the subsequent signal lamp prompt information can be ensured. And on this basis, predicting the target driving direction of the vehicle at the target passing intersection based on the current navigation system state and preset strategy can ensure the accurate identification of the user's demand direction for signal lamp information, and further comprehensively improve the data accuracy of the signal lamp prompt information and the matching degree of the demand direction.

[0026] The present invention also provides a signal lamp prompt information generation device, including:

[0027] An estimation module, which is used to determine the target passing intersection of the vehicle according to the obtained current vehicle position, and obtain the current signal lamp estimation information of each drivable direction corresponding to the target passing intersection;

[0028] An identification module, which is used to collect intersection direction image information and extract the current signal light identification information of each drivable direction from the intersection direction image information by using a preset identification model;

[0029] A correction module, which is used to correct the current signal light prediction information based on the current signal light identification information when the current signal light prediction information and the current signal light identification information meet a preset deviation correction condition;

[0030] A generation module, which is used to predict the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and a preset strategy, and generate a signal light prompt information according to the current signal light prediction information corresponding to the target driving direction.

[0031] The signal light prompt information generation device provided by the present invention and the signal light prompt information generation method have basically the same advantages compared with the prior art, and will not be elaborated here.

[0032] The present invention also provides a vehicle, including a memory and a processor;

[0033] The memory is used to store a computer program;

[0034] The processor is used to implement the signal light prompt information generation method as described above when executing the computer program.

[0035] The vehicle provided by the present invention and the signal light prompt information generation method have basically the same advantages compared with the prior art, and will not be elaborated here.

[0036] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the signal light prompt information generation method as described above is implemented.

[0037] The computer-readable storage medium provided by the present invention and the signal light prompt information generation method have basically the same advantages compared with the prior art, and will not be elaborated here. Description of the Drawings

[0038] Figure 1 It is a flow chart of the signal light prompt information generation method according to an embodiment of the present invention;

[0039] Figure 2 It is a schematic diagram of target driving direction prediction according to an embodiment of the present invention;

[0040] Figure 3 It is a schematic diagram of the structure of the signal light prompt information generation device according to an embodiment of the present invention;

[0041] Figure 4Schematic diagram of the structure of the vehicle according to an embodiment of the present invention. Detailed implementation manners

[0042] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0043] It should be understood that the various steps recorded in the method embodiments of the present invention can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this regard.

[0044] The term "including" and its variants used herein are open-ended, that is, "including but not limited to". The term "based on" is "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiment". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules, or units, and are not used to limit the order of functions performed by these devices, modules, or units or their interdependent relationships.

[0045] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0046] In the related art, after the user sets the destination in the navigation system, the navigation system can use methods such as displaying on the central control screen or voice prompting to provide the user with corresponding navigation information according to the planned route (such as prompting the user to turn or go straight at the intersection ahead). With the continuous development of information technology, some navigation systems can synchronously display the signal light information of the intersection ahead for the user (such as estimating the color and even the countdown of the signal light according to the collected big data or the signal light information publicly available on the relevant information platform), thereby reducing the probability of potential traffic violations and accidents.

[0047] However, on the one hand, the estimated information provided by the navigation system is often restricted by factors such as computing power and data volume, making it difficult to ensure accuracy. On the other hand, when the user does not set a destination, the navigation system cannot know the user's needs and cannot provide the signal light information corresponding to the expected driving direction of the user.

[0048] In view of the problems existing in the above related technologies, an embodiment of the present invention provides a method, device, vehicle and storage medium for generating signal light prompt information.

[0049] As Figure 1 shown, a method for generating signal light prompt information provided by an embodiment of the present invention includes the following steps:

[0050] S11: Determine the target passing intersection of the vehicle according to the obtained current vehicle position, and obtain the current signal light estimated information of each drivable direction corresponding to the target passing intersection.

[0051] Specifically, in this embodiment, the current vehicle position can be obtained based on the positioning device of the vehicle and may include the longitude and latitude coordinates of the vehicle. The target passing intersection of the vehicle referred to in this embodiment means the intersection that the vehicle is about to pass through, and each drivable direction corresponding to the target passing intersection means the directions corresponding to the roads connected by the target passing intersection (such as at least one of the straight-ahead direction, left-turn direction and right-turn direction).

[0052] In one embodiment, after obtaining the current vehicle position, the target passing intersection of the vehicle can be determined based on the current vehicle position, and the current signal light estimated information of each drivable direction corresponding to the target passing intersection can be obtained. For example, the position of the section where the vehicle is located in the electronic map can be determined according to the current vehicle position, and the current driving direction of the vehicle on this section can be determined according to the change direction of the current vehicle position. Then, based on the current driving direction, it can be determined which intersection of this section the vehicle is about to pass through to obtain the target passing intersection. On this basis, based on the position of the target passing intersection in the electronic map, the color and / or countdown of the signal light device of each drivable direction corresponding to the target passing intersection (that is, the signal light color and / or the signal light countdown) can be obtained from, for example, the navigation platform, and then the current signal light estimated information can be obtained.

[0053] S12: Collect the intersection direction image information, and use a preset recognition model to extract the current signal light recognition information of each drivable direction from the intersection direction image information.

[0054] Specifically, the intersection direction image information referred to in this embodiment may represent the forward intersection direction image information captured by an image acquisition device (such as a camera) pre-installed on the vehicle for photographing the signal lamp device. Before using the preset recognition model to extract the current signal lamp recognition information, historical image information of multiple signal lamps can be obtained in advance, and information such as the signal lamp device, each signal lamp color, signal lamp countdown, and direction identifier in each historical image information can be marked and processed to obtain each label information corresponding to each historical image. Based on the historical image information and its corresponding label information, a training database is constructed. On this basis, the preset initial model (such as a convolutional neural network model or other models) can be trained using the training database to obtain the preset recognition model.

[0055] In one embodiment, after collecting the intersection direction image information, the intersection direction image information can be input into the preset recognition model, and then the current signal lamp recognition information corresponding to each drivable direction can be extracted from the intersection direction image information. For example, assuming that the signal lamp device in the intersection direction image information includes three groups of signal lamps, and each group of signal lamps displays the direction identifier (such as a straight arrow, a left-turn arrow, and a right-turn arrow) it indicates and the corresponding countdown. Then, the direction identifier, the color of the direction identifier (i.e., the signal lamp color), and the number of the countdown (i.e., the signal lamp countdown) of each group of signal lamps in the intersection direction image information can be extracted using the preset recognition model, and the current signal lamp recognition information corresponding to each drivable direction can be matched based on the direction identifier.

[0056] Optionally, before using the preset recognition model to extract the current signal lamp recognition information corresponding to each drivable direction from the intersection direction image information, it further includes:

[0057] When the intersection direction image information does not meet the preset quality condition, a preset image processing operation is performed on the intersection direction image information. In this embodiment, the preset quality condition may include at least one of the resolution of the intersection direction image information being greater than the preset resolution and the signal-to-noise ratio of the intersection direction image information being less than the preset signal-to-noise ratio. After obtaining the intersection direction image information, it can be determined whether information such as the resolution and / or signal-to-noise ratio of the intersection direction image information meets the preset quality condition. If so, it means that the quality of the intersection direction image information is high, and the current signal lamp recognition information corresponding to each drivable direction can be directly extracted from the intersection direction image information using the preset recognition model. If not, it means that the quality of the intersection direction image information is low, and a preset image processing operation (such as at least one of grayscale processing, noise reduction processing, and contrast enhancement processing) needs to be performed on the intersection direction image information, and then the current signal lamp recognition information corresponding to each drivable direction is extracted from the processed intersection direction image information, which is beneficial to ensuring the accuracy and reliability of the recognition result.

[0058] S2: When the current signal light estimation information and the current signal light recognition information meet the preset deviation correction condition, correct the current signal light estimation information based on the current signal light recognition information.

[0059] Specifically, in this embodiment, both the current signal estimation information and the current signal light recognition information refer to the signal light information corresponding to the target passing intersection of the vehicle. The current signal estimation information can be directly obtained from, for example, a navigation platform, while the current signal light recognition information is collected by the vehicle's image acquisition device, and the data sources of the two are different. For the current signal light estimation information, the navigation platform may be restricted by factors such as insufficient computing power and insufficient sample quantity, resulting in deviations in the estimation results of the signal light device corresponding information at some intersections, such as the estimated signal light color not matching the actual signal light color, the estimated signal light countdown not matching the actual signal light countdown, etc. For the current signal light recognition information, since it is extracted based on the image information of the intersection direction actually collected by the vehicle, its credibility is higher. Then when the two meet the preset deviation condition, the current signal light estimation information can be corrected based on the current signal light recognition information.

[0060] In one embodiment, the preset deviation condition can be set in advance. For example, the status information of the vehicle's current navigation system (i.e., the current navigation system status, which may include the navigation status or the cruise status) can be obtained. When the current navigation system status is the navigation status, the target driving direction of the vehicle at the target passing intersection (assumed to be going straight) can be determined based on the navigation information of the navigation system (such as the navigation planned path). The current signal light estimation information in the straight-ahead direction (such as including the estimated signal light color and the estimated signal light countdown) and the current signal light recognition information in the straight-ahead direction (such as including the recognized signal light color and the recognized signal light countdown) can be obtained respectively. When the estimated signal light color is different from the recognized signal light color, and / or, the estimated signal light countdown is different from the recognized signal light countdown, it indicates that the two meet the preset deviation correction condition, and the current signal light estimation information can be corrected based on the current signal light recognition information. For example, when the estimated signal light color is different from the recognized signal light color, the real-time updated recognized signal light color can be used as the estimated signal light color, and / or, when the estimated signal light countdown is different from the recognized signal light countdown, the real-time updated recognized signal light countdown can be used as the estimated signal light countdown until the obtained current signal light estimation information and the recognized current signal light recognition information do not meet the preset deviation correction condition, and / or, the preset recognition model cannot extract the current signal light estimation information from the intersection direction image (such as the signal light device is not recognized in the intersection direction image information). Optionally, when it is detected that the two meet the preset deviation correction condition, it can be determined again whether the updated current signal light estimation information and the current signal light recognition information meet the preset deviation correction condition, so as to avoid inaccurate judgment results caused by abnormal information such as the accuracy error of the preset recognition model or network transmission delay.

[0061] S3: Predict the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system status and the preset strategy, and generate a signal light prompt message according to the current signal light estimation information corresponding to the target driving direction.

[0062] Specifically, the current navigation system state referred to in this embodiment can represent the current working state of the vehicle's navigation system, which may include the navigation state or the cruise state, and can be determined by detecting whether the user has set a destination. For example, when a preset destination is detected, the current navigation system state is the navigation state; when no preset destination is detected, the current navigation state is the cruise state. The target driving direction referred to in this embodiment represents the drivable direction selected by the vehicle at the target passing intersection. After obtaining the current navigation system state, the target driving direction of the vehicle can be predicted based on the current navigation system state and a preset strategy. For example, when the current navigation system state is the navigation state, the planned direction of the vehicle at the target passing intersection (such as the left-turn direction) can be determined according to the navigation information of the navigation system (such as the planned route), and the drivable direction corresponding to the planned direction is used as the target driving direction. Another example is that when the current navigation system state is the cruise state, the collected intersection direction image information can be input into a preset recognition model for road sign extraction (such as lane signs on road signs and lane signs on the road surface), and then the lane where the vehicle is currently located can be determined by methods such as visually analyzing road signs and the vehicle's perspective. On this basis, the target driving direction of the vehicle can be predicted according to the lane where the vehicle is currently located. Assuming that the current lane is one of the straight lane, the left-turn and straight lane, or the right-turn and straight lane, the predicted target driving direction of the vehicle is the straight direction; assuming that the current lane is the left-turn lane, the predicted target driving direction of the vehicle is the left-turn direction; assuming that the current lane is the right-turn lane, the predicted target driving direction of the vehicle is the right-turn direction.

[0063] In one embodiment, after predicting the target driving direction, the signal light prompt information can be generated according to the current signal light estimation information corresponding to the target driving direction. For example, the current signal light estimation information corresponding to the drivable direction matching the target driving direction is screened out to obtain the signal light prompt information. And the signal light prompt information is displayed to the user through a central control screen, voice broadcast, etc. For example, assuming that the signal light prompt information includes the signal light color and the signal light countdown, a preset color identifier matching the signal light color can be displayed on the central control screen, and the signal light countdown number can be displayed near the preset color identifier.

[0064] It should be understood that through predicting the target driving direction in this embodiment, signal light prompt information matching the user's needs can be provided to the user throughout the whole section. Even when the vehicle is far from the target passing intersection (such as the distance is greater than 100m), accurate and reliable signal light information corresponding to the target passing direction can be provided to the user. Especially when the user learns in advance according to the signal light prompt information that the color of the signal light corresponding to the target passing direction is about to change (such as the remaining time of the green light countdown is 2s), the user can control the vehicle to decelerate in advance, thereby effectively reducing the probability of safety accidents.

[0065] In this embodiment, determining the target passing intersection of the vehicle according to the current vehicle position is beneficial to knowing the position of the signal lamp device affecting the vehicle passing, so that the current signal lamp estimation information of each drivable direction corresponding to the obtained target passing intersection can provide a data basis for the generation of subsequent signal lamp prompt information. Since the obtained current signal lamp estimation information is restricted by factors such as computing power and lack of data samples, the estimation result may be incorrect. In this embodiment, in addition to obtaining the current signal lamp estimation information, a preset recognition model is also used to extract the current signal lamp recognition information of each drivable direction from the intersection direction image information actually collected by the vehicle, providing an effective reference basis for the accuracy evaluation of the current signal lamp estimation information. In this embodiment, the accuracy evaluation of the current signal lamp estimation information is realized by judging whether the signal lamp estimation information and the current signal lamp recognition information meet the preset deviation correction condition. When the preset correction condition is met, it indicates that the current signal lamp estimation information does not match the actual signal lamp situation, and it is necessary to correct the current signal lamp estimation information based on the current signal lamp recognition information to ensure the accuracy of the subsequent signal lamp prompt information. On this basis, based on the obtained current navigation system state and preset strategy, the target driving direction of the vehicle at the target passing intersection is predicted, and the signal lamp prompt information is generated according to the current signal lamp estimation information corresponding to the target driving direction. Whether the current navigation system state is in the navigation state or in the cruise state, accurate signal lamp prompt information can be provided for the user. On the one hand, the current signal lamp estimation information can make up for the limitations of the current signal lamp recognition information in terms of distance, space, etc. (such as in the case of long distance, spatial occlusion, bad weather, etc., it is difficult to extract the current signal lamp recognition information from the intersection direction image information), ensuring the timeliness and comprehensiveness of the signal lamp prompt information. On the other hand, the current signal lamp recognition information extracted based on the intersection direction image information can make up for the deficiency of the current signal lamp estimation information in terms of accuracy. In this way, in this embodiment, by obtaining multi-channel signal lamp information and presetting deviation correction conditions, the comprehensiveness and accuracy of the subsequent signal lamp prompt information can be ensured. And on this basis, predicting the target driving direction of the vehicle at the target passing intersection based on the current navigation system state and preset strategy can also ensure accurate identification of the user's demand direction for signal lamp information, thereby comprehensively improving the data accuracy of the signal lamp prompt information and the matching degree of the demand direction.

[0066] Optionally, the preset deviation correction condition includes:

[0067] It is detected that there is a signal lamp color deviation between the current signal lamp recognition information and the current signal lamp estimation information corresponding to the drivable direction, and / or, it is detected that there is a signal lamp countdown deviation between the current signal lamp recognition information and the current signal lamp estimation information corresponding to the drivable direction, and the signal lamp countdown deviation is greater than the preset deviation threshold.

[0068] Specifically, in this embodiment, both the current signal light recognition information and the current signal light estimation information may include at least one of the signal light color and the signal light countdown. The existence of a signal light color deviation referred to in this embodiment means that the signal light colors corresponding to the same drivable direction (such as the straight-ahead direction) in the current signal light recognition information and the current signal light estimation information are different. The existence of a signal light countdown deviation referred to in this embodiment means that the signal light countdowns corresponding to the same drivable direction (such as the right-turn direction) in the current signal light recognition information and the current signal light estimation information are different, and the signal light countdown deviation between the two is greater than a preset deviation threshold (such as 3s).

[0069] In this embodiment, the current signal light estimation information obtained based on the current vehicle position and the current signal light recognition information extracted based on the intersection direction image information can both be used to indicate information such as the signal light color and / or the signal light countdown. When there is a signal light color deviation and / or a signal light countdown deviation between the signal light information obtained by the two methods, it is very likely that the current signal light estimation information may be incorrect. And an error in the signal light color in the current signal light information has a greater impact on the running safety of the vehicle than an error in the signal light countdown. Therefore, in this embodiment, when it is detected that there is a signal light color deviation between the two, the preset deviation correction condition can be satisfied. And after it is detected that there is a signal light countdown deviation between the two, the preset deviation correction condition can only be satisfied when the signal light countdown deviation is greater than the preset deviation threshold. On the basis of ensuring that the signal light prompt information does not affect driving safety, it is also beneficial to improve the signal light information generation efficiency.

[0070] Optionally, after correcting the current signal light estimation information based on the current signal light recognition information, it further includes:

[0071] When the number of times that the signal light countdown deviation corresponding to the drivable direction is greater than the preset deviation threshold exceeds the first preset number of times, the target passing intersection corresponding to the drivable direction is taken as the set intersection, and each signal light countdown deviation corresponding to the drivable direction is taken as the historical deviation;

[0072] When it is determined according to the current vehicle position that the target passing intersection is switched to the set intersection and each historical deviation meets the preset consistency condition, the current signal light estimation information obtained by correcting based on the historical deviation is used.

[0073] Specifically, for each target passing intersection, the number of times that the signal light countdown deviation corresponding to each drivable direction is greater than the preset deviation threshold is respectively accumulated (for the convenience of understanding and expression, it is denoted as the first number of times). When it is detected that the first number of times corresponding to any drivable direction is greater than the first number of times threshold (such as 3 times), the target passing intersection corresponding to the drivable direction can be taken as the set intersection, and each signal light countdown deviation corresponding to the drivable direction is taken as the historical deviation.

[0074] In one embodiment, the current vehicle position can be obtained every preset period (such as 50 to 100 ms), and the target passing intersection of the vehicle can be determined again. When the target passing intersection determined last time is inconsistent with the target passing intersection determined this time, it indicates that the target passing intersection has switched, reflecting that the vehicle has left the target passing intersection determined last time. In this embodiment, when it is determined according to the current vehicle position that the target passing intersection switches to the set intersection, it means that after the set intersection is determined, the vehicle travels to the section corresponding to the set intersection again (such as in the scenario where the vehicle has a relatively fixed commuting route every day and needs to pass through the set intersection every time it commutes). At this time, it can be judged whether each historical deviation meets the preset consistency condition (such as the variance of each historical deviation is less than the preset variance threshold). If so, the currently obtained traffic signal prediction information is corrected based on the historical deviation. If not, no relevant correction operation is performed.

[0075] In this embodiment, for the scenario where the vehicle has a relatively fixed route, when the number of times that the countdown deviation of the traffic signal corresponding to the drivable direction exceeds the preset deviation threshold exceeds the first preset number, it indicates that the currently obtained traffic signal prediction information corresponding to the target passing intersection may be inaccurate due to insufficient historical sample data, resulting in inaccurate prediction results repeatedly. When the vehicle travels on the section corresponding to the target passing intersection multiple times, the countdown of the traffic signal in the currently obtained traffic signal prediction information deviates from the actual countdown of the traffic signal. In this embodiment, the target passing intersection corresponding to the drivable direction is used as the set intersection, and the countdown deviations of each traffic signal corresponding to the drivable direction are used as historical deviations, which is beneficial to identifying intersections with repeated information deviations. When it is determined according to the current vehicle position that the target passing intersection switches to the set intersection, it means that the vehicle travels to the section corresponding to the set intersection again. At this time, if it is determined that each historical deviation meets the preset consistency condition, the currently obtained traffic signal prediction information can be corrected in advance based on the historical deviation. Even when the current traffic signal recognition information is not extracted (such as when the distance between the vehicle and the set intersection is far and the traffic signal device has not been photographed), the current traffic signal prediction information can be corrected in advance based on the deviation law specific to the set intersection, providing a reliable information reference for the user, and thus improving driving safety.

[0076] Optionally, the preset consistency condition includes that each historical deviation is positive or negative; correcting the currently obtained traffic signal prediction information based on the historical deviation includes:

[0077] Adding the countdown of the traffic signal corresponding to the currently obtained traffic signal prediction information to the historical deviation closest to the current time to obtain the corrected currently obtained traffic signal prediction information.

[0078] Specifically, in this embodiment, the preset consistency condition may include that all historical deviations are positive or all are negative, which can be determined by judging whether all historical deviations are greater than 0 or less than 0. Optionally, to further improve the accuracy of the finally generated signal light prompt information, the preset consistency condition in this embodiment may further include that the time intervals corresponding to all historical deviations are less than a preset interval (such as 72h).

[0079] The above scenario will be specifically described with an example:

[0080] (1) Suppose there are two drivable directions (such as the left-turn direction and the straight-ahead direction) at a certain target passing intersection. Among them, the number of times that the signal light countdown deviation corresponding to the left-turn direction continuously exceeds the preset deviation threshold is three times. Then, this target passing intersection can be used as the set intersection, and the three signal light countdown deviations corresponding to the left-turn direction (for the convenience of understanding and expression, they are respectively denoted as the first deviation, the second deviation, and the third deviation in chronological order) can be used as historical deviations;

[0081] (2) The current vehicle position is obtained every 100 ms, and the target passing intersection is re-determined. When it is determined according to the current vehicle position that the target passing intersection switches to the set intersection, it is judged whether the first deviation, the second deviation, and the third deviation are all positive or all negative.

[0082] (3) When it is detected that the first deviation, the second deviation, and the third deviation are all negative, it is indicated that all historical deviations meet the preset consistency condition.

[0083] (4) The signal light countdown corresponding to the obtained current signal light prediction information is added to the third deviation (such as -3s) closest to the current time to obtain the corrected current signal light prediction information.

[0084] In this embodiment, by judging whether all historical deviations meet the preset consistency condition, it is beneficial to evaluate whether all historical deviations have consistent regularity. If all historical deviations do not meet the preset consistency condition, it means that the inaccuracy of the current signal light prediction information corresponding to this set intersection may be caused by accidental factors. When the vehicle passes through this set intersection next time, the probability that the signal light prediction information is still inaccurate is relatively small. On the contrary, when all historical deviations meet the preset consistency condition, when the vehicle passes through this set intersection next time, the probability that the signal light prediction information is still inaccurate is relatively high. At this time, adding the signal light countdown corresponding to the obtained current signal light prediction information to the historical deviation closest to the current time is beneficial to ensuring the accuracy of the corrected current signal light prediction information.

[0085] Optionally, after correcting the current signal light prediction information based on the current signal light recognition information, it further includes:

[0086] When the number of times that the current signal light prediction information and the current signal light recognition information do not meet the preset deviation correction condition exceeds the second preset number of times, stop collecting the intersection direction image information until it is detected that the signal light prompt information meets the preset untrustworthy condition; wherein, the preset untrustworthy condition includes that when the signal light color corresponding to the signal light prompt information is the preset prohibited passing color, it is detected that the target passing intersection has switched.

[0087] Specifically, accumulate the number of times that the current signal light prediction information and the current signal light recognition information do not meet the preset deviation correction condition (for the convenience of understanding and expression, denote it as the second number of times). When it is detected that the second number of times exceeds the second preset number of times, it indicates that the credibility of the current signal light prediction information is relatively high. The collection of the intersection direction image information can be stopped, and the signal light prompt information can be generated only relying on the obtained current signal light prediction information. After stopping the collection of the intersection direction image information, it can be determined whether the signal light prompt information meets the preset untrustworthy condition. If so, return to the step of collecting the intersection direction image information. If not, continue to stop collecting the intersection direction image information.

[0088] In this embodiment, when the signal light color corresponding to the signal light prompt information is the preset prohibited passing color (such as red), if it is detected that the target passing intersection has switched (that is, the continuously determined target passing intersections are different), it indicates that the user's actual driving decision does not match the signal light indication information (such as when the signal light color corresponding to the target passing intersection is red, the vehicle passes through the target passing intersection), and the signal light prompt information is likely to have a large deviation. At this time, it is necessary to re-obtain the intersection direction image information, so as to continue to make up for the error of the prediction result based on the visual recognition result of the signal light information and ensure the accuracy of the signal light prompt information.

[0089] Optionally, the current navigation system state includes a navigation state or a cruise state; predicting the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and the preset strategy includes:

[0090] When the current navigation system state is the navigation state, predict the target driving direction according to the obtained current navigation information;

[0091] When the current navigation system state is the cruise state, determine the current distance between the vehicle and the target passing intersection according to the current vehicle position, and predict the target driving direction based on the current distance.

[0092] Specifically, the current navigation information referred to in this embodiment may represent information such as the planned route provided by the navigation system according to the destination set in advance by the user. When the current navigation system is in the navigation state, the target driving direction of the vehicle at the target passing intersection can be predicted based on information such as the planned route. When the current navigation system is in the cruise state, in this embodiment, the current distance between the vehicle and the target passing intersection needs to be determined based on the current vehicle position, and the target driving direction is predicted based on the current distance. For example, the current distance between the two can be determined according to the current vehicle position and the positioning information of the target passing intersection on the electronic map. Assume that when it is detected that the current lane where the vehicle is located is a left-turn through lane based on at least one of, for example, the current vehicle position, the intersection direction image information, or the lane positioning provided by the navigation system. When the current distance is greater than the preset distance threshold (such as 30 m), the target driving direction is the straight-ahead direction. When the current distance is less than or equal to 30 m, the target driving direction is the left-turn direction.

[0093] In this embodiment, when the current navigation system is in the navigation state, the current navigation information can provide a relatively accurate reference for the vehicle's direction selection at the target passing intersection. At this time, based on the current navigation information, it is beneficial to improve the accuracy of predicting the target driving direction. When the current navigation system is in the cruise state, it is difficult to directly grasp the user's intention of the vehicle's direction selection. Since when the vehicle is driving on a section of the road, if the user has a turning intention, it is necessary to drive into the lane corresponding to the corresponding drivable direction in advance at a certain distance from the target passing intersection. Then when the vehicle is driving on a section of the road, the closer the vehicle is to the target passing intersection, the more the drivable direction corresponding to the current lane where the vehicle is located can reflect the direction expected by the user. In this case, determining the current distance between the vehicle and the target passing intersection based on the current vehicle position and predicting the target driving direction based on the current distance is beneficial to ensuring the rationality and reliability of predicting the target driving direction, thereby comprehensively improving the matching degree between the signal light prompt information and the actual needs of the user.

[0094] Optionally, predicting the target driving direction based on the current distance includes:

[0095] When the current distance is greater than the preset distance threshold, predict the target driving lane corresponding to the vehicle based on the obtained current vehicle operation information, and predict the target driving direction according to the target driving lane;

[0096] When the current distance is less than or equal to the preset distance threshold, identify the current driving lane corresponding to the vehicle based on the intersection direction image information and / or the current position information, and predict the target driving direction according to the current driving lane.

[0097] Specifically, the current vehicle operation information referred to in this embodiment may include information such as the current vehicle position, the current heading angle, the current steering wheel angle, and the current vehicle speed. When the current distance is greater than the preset distance threshold, the target driving lane of the vehicle can be predicted based on the acquired current vehicle operation information. For example, the above current operation information can be acquired every preset step length (such as 2 - 5 s), and the position of the vehicle at the next step length can be predicted based on the acquired current operation information to obtain the predicted position (for example, inputting the current operation information into an existing model such as a kinematic model for position prediction, which will not be elaborated here), and then the lane corresponding to the predicted position is used as the target driving lane of the vehicle. Assuming that when the current distance is greater than 30 m, the target driving lane at the next step length is determined to be the straight - through lane according to the predicted position (such as the position longitude and latitude coordinates) at the next step length, then the target driving direction can be predicted to be the straight - through direction according to the drivable direction corresponding to the target driving lane. When the current distance is less than or equal to the preset distance threshold, the current driving lane of the vehicle can be identified based on the intersection direction image information and / or the current position information, and then the target driving direction can be predicted. For example, assuming that when the current distance is less than or equal to 30 m, the vehicle can be lane - level positioned based on the current position information and the high - definition map, and then the current driving lane of the vehicle (such as the right - turn straight - through lane) can be identified, and then the target driving direction can be predicted to be the right - turn direction according to the current driving lane.

[0098] Now, a specific embodiment is used to further introduce the method for predicting the target driving direction based on the current distance:

[0099] In this embodiment, the schematic diagram for predicting the target driving direction is as Figure 2 shown, Figure 2 In the figure, P1 represents the scenario where the current distance is greater than the preset distance threshold (for ease of understanding and description, it is denoted as the first scenario), the lane corresponding to the current vehicle position. P2 represents the target driving lane of the vehicle at a future moment predicted based on the acquired current vehicle operation information in the first scenario. P3 represents the lane corresponding to the target driving direction predicted based on the target driving lane in the first scenario. P4 represents the scenario where the current distance is less than or equal to the preset distance threshold (for ease of understanding and description, it is denoted as the second scenario), the current driving lane where the vehicle is located. P5 represents the lane corresponding to the target driving direction predicted based on the current driving lane in the second scenario. Among them, in the first scenario, when the lane corresponding to the current vehicle position is the left - turn straight - through lane and the predicted target driving lane is the right - turn straight - through lane, the predicted target passing direction is the right - turn direction. In the second scenario, when the current driving lane is the left - turn straight - through lane, the predicted target passing direction is the left - turn direction.

[0100] In this embodiment, when the current distance is greater than the preset distance threshold, the correlation between the lane where the vehicle is currently located and the steering intention of the user driving the vehicle is relatively low, while factors such as the change of the vehicle's lane path are more relevant to its steering intention. In this embodiment, when the current distance is greater than the preset distance threshold, predicting the target driving lane corresponding to the vehicle based on the obtained current vehicle operation information is beneficial to predicting in advance whether the vehicle has the intention to change lanes, and then predicting the target driving direction according to the target driving lane. When the current distance is less than or equal to the preset distance threshold, the current driving lane where the vehicle is located has a relatively high correlation with its steering intention. At this time, identifying the current driving lane corresponding to the vehicle based on the intersection direction image information and / or the current position information is beneficial to identifying the user's direction selection intention. In this way, even when the current navigation system is in the cruise mode, this embodiment can ensure the rationality, reliability, and accuracy of the prediction of the target driving direction based on the current distance between the vehicle and the target passing intersection and the effective reference information, thereby improving the matching degree between the signal light prompt information and the user's needs.

[0101] As Figure 3 shown, a signal light prompt information generation device 300 provided by an embodiment of the present invention includes:

[0102] An estimation module 310, which is used to determine the target passing intersection of the vehicle according to the obtained current vehicle position, and obtain the current signal light estimation information of each drivable direction corresponding to the target passing intersection;

[0103] An identification module 320, which is used to collect intersection direction image information, and extract the current signal light identification information of each drivable direction from the intersection direction image information by using a preset identification model;

[0104] A correction module 330, which is used to correct the current signal light estimation information based on the current signal light identification information when the current signal light estimation information and the current signal light identification information meet a preset deviation correction condition;

[0105] A generation module 340, which is used to predict the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and a preset strategy, and generate signal light prompt information according to the current signal light estimation information corresponding to the target driving direction.

[0106] The technical effects that can be produced by the signal light prompt information generation device and the signal light prompt information generation method provided in this embodiment are basically the same, and will not be elaborated here.

[0107] As Figure 4As shown in the figure, a vehicle 400 provided by an embodiment of the present invention includes a memory 410 and a processor 420; the memory 410 is used to store a computer program; the processor 420 is used to implement the signal light prompt information generation method as described above when executing the computer program.

[0108] Or, a vehicle 400 includes a memory 410 and a processor 420 coupled to the memory 410; the memory 410 is configured to store a computer program; the processor 420 is configured to perform the following operations when executing the computer program:

[0109] Determine the target passing intersection of the vehicle according to the obtained current vehicle position, and obtain the current signal light prediction information of each drivable direction corresponding to the target passing intersection;

[0110] Collect intersection direction image information, and extract the current signal light recognition information of each drivable direction from the intersection direction image information by using a preset recognition model;

[0111] When the current signal light prediction information and the current signal light recognition information meet the preset deviation correction condition, correct the current signal light prediction information based on the current signal light recognition information;

[0112] Predict the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and preset strategy, and generate signal light prompt information according to the current signal light prediction information corresponding to the target driving direction.

[0113] The vehicle provided in this embodiment and the signal light prompt information generation method can produce basically the same technical effects, which will not be elaborated here.

[0114] A computer-readable storage medium provided by an embodiment of the present invention has a computer program stored thereon. When the computer program is executed by a processor, the signal light prompt information generation method as described above is implemented.

[0115] Or, a non-volatile computer-readable storage medium has a computer program stored thereon. When the computer program is executed by a processor, the processor is made to perform the following operations:

[0116] Determine the target passing intersection of the vehicle according to the obtained current vehicle position, and obtain the current signal light prediction information of each drivable direction corresponding to the target passing intersection;

[0117] Collect intersection direction image information, and extract the current signal light recognition information of each drivable direction from the intersection direction image information by using a preset recognition model;

[0118] When the current traffic signal prediction information and the current traffic signal recognition information meet the preset deviation correction condition, correct the current traffic signal prediction information based on the current traffic signal recognition information;

[0119] Predict the target driving direction of the vehicle at the target passing intersection based on the obtained current navigation system state and preset strategy, and generate a traffic signal prompt information according to the current traffic signal prediction information corresponding to the target driving direction.

[0120] The computer-readable storage medium provided in this embodiment has substantially the same technical effects as the traffic signal prompt information generation method, and will not be elaborated here.

[0121] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present invention. In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0122] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will all fall within the protection scope of the present invention.

Claims

1. A method for generating signal light prompt information, characterized in that: include: Determine the target intersection for the vehicle according to the acquired current vehicle position, and obtain the estimated information of the current traffic lights for each drivable direction corresponding to the target intersection; Collecting intersection direction image information, and extracting current traffic light identification information of each drivable direction from the intersection direction image information using a preset recognition model; When the current signal light prediction information and the current signal light identification information meet a preset deviation correction condition, correcting the current signal light prediction information based on the current signal light identification information; The target driving direction of the vehicle at the target intersection is predicted based on the acquired current navigation system state and a preset strategy, and the signal light prompt information is generated according to the current signal light estimation information corresponding to the target driving direction.

2. The method for generating signal light prompt information according to claim 1, characterized in that: The preset deviation correction conditions include: It is detected that there is a signal light color deviation between the current signal light identification information corresponding to the drivable direction and the current signal light prediction information, and / or it is detected that there is a signal light countdown deviation between the current signal light identification information corresponding to the drivable direction and the current signal light prediction information, and the signal light countdown deviation is greater than a preset deviation threshold.

3. The method for generating signal light prompt information according to claim 2, characterized in that: After the current signal light prediction information is corrected based on the current signal light identification information, the method further includes: When the number of times that the countdown deviation of the traffic light corresponding to the drivable direction is greater than the preset deviation threshold exceeds a first preset number of times, the target traffic intersection corresponding to the drivable direction is used as a set intersection, and the countdown deviations of the traffic lights corresponding to the drivable direction are used as historical deviations; When it is determined according to the current vehicle position that the target traffic intersection is switched to the set intersection, and each of the historical deviations satisfies a preset consistency condition, the current traffic light estimation information obtained is corrected based on the historical deviations.

4. The method for generating signal light prompt information according to claim 3, characterized in that: The preset consistency condition includes that all the historical deviations are positive or negative; the current traffic light estimation information obtained based on the historical deviation correction includes: The traffic light countdown corresponding to the acquired current traffic light prediction information is added to the historical deviation closest to the current time to obtain the corrected current traffic light prediction information.

5. The method for generating signal light prompt information according to claim 1, characterized in that: After the current signal light prediction information is corrected based on the current signal light identification information, the method further includes: When the number of times that the current traffic light prediction information and the current traffic light identification information do not satisfy the preset deviation correction condition exceeds a second preset number of times, stop collecting the intersection direction image information until it is detected that the traffic light prompt information satisfies a preset unreliable condition; wherein the preset unreliable condition includes when the traffic light color corresponding to the traffic light prompt information is a preset prohibited passage color, detecting that the target passage intersection has switched.

6. The method for generating signal light prompt information according to claim 1, characterized in that: The current navigation system state includes a navigation state or a cruise state; and the predicting the target driving direction of the vehicle at the target intersection based on the acquired current navigation system state and a preset strategy includes: When the current navigation system state is the navigation state, predicting the target driving direction according to the acquired current navigation information; When the current navigation system state is the cruising state, the current distance between the vehicle and the target intersection is determined according to the current vehicle position, and the target driving direction is predicted based on the current distance.

7. The method for generating signal light prompt information according to claim 6, characterized in that: The predicting the target driving direction based on the current distance includes: When the current distance is greater than a preset distance threshold, predicting a target driving lane corresponding to the vehicle based on the acquired current vehicle operation information, and predicting the target driving direction according to the target driving lane; When the current distance is less than or equal to the preset distance threshold, the current driving lane corresponding to the vehicle is identified based on the intersection direction image information and / or the current position information, and the target driving direction is predicted according to the current driving lane.

8. A signal light prompt information generating device, characterized in that: include: An estimation module, which is used to determine the target intersection of the vehicle according to the acquired current vehicle position, and obtain the estimated information of the current traffic lights of each drivable direction corresponding to the target intersection; A recognition module, which is used to collect intersection direction image information and extract current traffic light recognition information of each drivable direction from the intersection direction image information using a preset recognition model; a correction module, configured to correct the current signal light prediction information based on the current signal light identification information when the current signal light prediction information and the current signal light identification information meet a preset deviation correction condition; A generation module is used to predict the target driving direction of the vehicle at the target intersection based on the acquired current navigation system state and a preset strategy, and generate signal light prompt information according to the current signal light estimation information corresponding to the target driving direction.

9. A vehicle, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is configured to implement the signal light prompt information generating method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method for generating signal light prompt information according to any one of claims 1 to 7 is implemented.

Images