阅读说明：本技术 一种车辆智能前照灯系统 (Intelligent headlamp system of vehicle ) 是由 张炳力 王亮 江尚 佘亚飞 秦浩然 郑杰禹 于 2021-09-18 设计创作，主要内容包括：本发明公开了一种车辆智能前照灯系统,是先生成车辆前方探测范围内的点云数据和车辆前方的图像数据；再根据输入的点云数据和图像数据,通过点云与图像数据的融合算法获知车辆前方目标的准确信息和准确位置；然后根据输入的车辆前方目标的准确位置,判断是否有对向来车,若有对向来车则生成对左大灯照明模组和/或右大灯照明模组的自适应调节指令,并控制左大灯驱动模块和/或右大灯驱动模块根据自适应调节指令驱动左大灯照明模组和/或右大灯照明模组,在对向来车驶入左大灯照明模组和/或右大灯照明模组的照明范围内前,在照明范围内形成照明暗区。本发明能避免对向来车的驾驶员产生眩目,提高夜间行车安全性。(The invention discloses an intelligent headlamp system of a vehicle, which is characterized in that point cloud data in a detection range in front of the vehicle and image data in front of the vehicle are generated firstly; then according to the input point cloud data and the image data, obtaining the accurate information and the accurate position of the target in front of the vehicle through a fusion algorithm of the point cloud and the image data; and then judging whether an opposite vehicle is coming according to the input accurate position of the target in front of the vehicle, if so, generating a self-adaptive adjustment instruction for the left headlamp lighting module and/or the right headlamp lighting module, and controlling the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp lighting module and/or the right headlamp lighting module according to the self-adaptive adjustment instruction, so that a lighting dark area is formed in a lighting range before the opposite vehicle drives into the lighting range of the left headlamp lighting module and/or the right headlamp lighting module. The invention can avoid dazzling drivers of oncoming vehicles and improve the driving safety at night.)

1. A vehicle intelligent headlamp system, comprising: the device comprises a laser radar module, a forward-looking camera module, a signal processing module, a headlamp controller, a whole vehicle CAN bus, a left headlamp driving module, a right headlamp driving module, a left headlamp lighting module and a right headlamp lighting module;

the laser radar module is arranged right above the vehicle and used for acquiring point cloud data in a laser radar detection range in front of the vehicle and sending the point cloud data to the signal processing module through the whole vehicle CAN bus;

the front-view camera module is arranged right above the inside rear-view mirror and used for acquiring image data in the field of view of the front camera of the vehicle and sending the image data to the signal processing module through the whole vehicle CAN bus;

the signal processing module receives point cloud data and image data through a whole vehicle CAN bus, performs data preprocessing and fusion algorithm processing, acquires target identification information in front of a vehicle, and classifies the target identification information in front of the vehicle, including: the front vehicle front left direction running vehicle, the front vehicle front right direction running vehicle and the front vehicle front right direction running vehicle are classified, and the classified front vehicle front target identification information is sent to the headlamp controller through the whole vehicle CAN bus;

the headlamp controller is respectively connected with the left headlamp driving module and the right headlamp driving module, the left headlamp driving module is connected with the left headlamp lighting module, and the right headlamp driving module is connected with the right headlamp lighting module;

the headlamp controller converts the target identification information in front of the vehicle into a corresponding self-adaptive adjusting instruction, and controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp lighting module and/or the right headlamp lighting module to form a lighting dark area in the lighting range of the left headlamp lighting module and/or the right headlamp lighting module according to the self-adaptive adjusting instruction before the opposite vehicle drives into the lighting range of the left headlamp lighting module and/or the right headlamp lighting module;

the headlamp controller judges whether a target vehicle in front of the vehicle drives out of the illumination dark area or not according to the identification information in front of the vehicle, and if so, the headlamp controller controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp illumination module and/or the right headlamp illumination module to recover to normal illumination; otherwise, adjusting the lighting effect of the left headlamp lighting module and/or the right headlamp lighting module in real time according to the target identification information in front of the vehicle updated in real time.

2. The intelligent headlamp system of claim 1, wherein the signal processing module obtains the identification information of the object in front of the vehicle according to the following process:

step a, carrying out space matching and time matching on the point cloud data and the image data so as to project the point cloud data onto the image data;

b, identifying a front target vehicle in the image data by using a convolutional neural network, and outputting detection frames of the front target vehicle and the image data confidence of each detection frame;

step c, counting the number of point clouds in each detection frame, obtaining the ratio of the number of point clouds in each detection frame to the maximum number of point clouds in each detection frame, and using the ratio as the confidence coefficient of point cloud data of each detection frame;

d, performing weighted calculation on the image data confidence coefficient and the point cloud data confidence coefficient to obtain a fusion confidence coefficient, and determining that no front target vehicle exists in the detection frame when the fusion confidence coefficient is smaller than a set threshold value; when the fusion confidence coefficient is larger than or equal to the threshold value, determining that a front target vehicle exists in the detection frame, and obtaining target identification information in front of the vehicle;

step e, classifying the target identification information in front of the vehicle, including: the vehicle front object identification information after classification is obtained from a vehicle left front co-running vehicle, a vehicle left front counter-running vehicle, a vehicle right front co-running vehicle, and a vehicle right front counter-running vehicle.

3. The intelligent headlamp system for the automobile as claimed in claim 1, wherein the adaptive adjustment command comprises: if the identification information is that the vehicle runs in the same direction in front of the left side of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp illumination module; if the identification information is that the vehicle runs in the left front direction of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp lighting module and the automobile right headlamp lighting module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile right headlamp illumination module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjusting instruction is to reduce the lighting brightness of the left headlamp lighting module and the right headlamp lighting module of the automobile.

Technical Field

The invention relates to the technical field of automotive electronics, in particular to an intelligent headlamp system of a vehicle.

Background

When the vehicle is driving at night, the headlamps need to be turned on to help the driver see the road ahead. However, when a vehicle is meeting or following at night, if the brightness of the headlamp is too high, the headlamp can dazzle a driver and cause traffic accidents because the driver cannot see the road clearly.

The current intelligent headlamp system can detect the oncoming vehicle according to the forward-looking camera, so as to adjust the brightness of the headlamp, however, the imaging quality of the camera caused by factors such as severe night illumination conditions is not high, the camera cannot be improved by optimizing an image processing algorithm, the problems of false detection, missing detection and the like are easily caused, the illumination effect of the headlamp is finally influenced, and the driving safety at night is influenced.

Disclosure of Invention

The invention aims to solve the defects in the prior art, and provides an intelligent headlamp system for a vehicle, so that the problems of false detection, missed detection and the like of a front-view camera in the existing intelligent headlamp system can be effectively solved, the illumination effect of the headlamp is improved, the occurrence rate of traffic accidents on roads at night is reduced, and the safety of driving at night is guaranteed.

In order to achieve the purpose, the invention adopts the following technical scheme:

the invention relates to an intelligent headlamp system of a vehicle, which is characterized by comprising the following components: the device comprises a laser radar module, a forward-looking camera module, a signal processing module, a headlamp controller, a whole vehicle CAN bus, a left headlamp driving module, a right headlamp driving module, a left headlamp lighting module and a right headlamp lighting module;

the laser radar module is arranged right above the vehicle and used for acquiring point cloud data in a laser radar detection range in front of the vehicle and sending the point cloud data to the signal processing module through the whole vehicle CAN bus;

the front-view camera module is arranged right above the inside rear-view mirror and used for acquiring image data in the field of view of the front camera of the vehicle and sending the image data to the signal processing module through the whole vehicle CAN bus;

the signal processing module receives point cloud data and image data through a whole vehicle CAN bus, performs data preprocessing and fusion algorithm processing, acquires target identification information in front of a vehicle, and classifies the target identification information in front of the vehicle, including: the front vehicle front left direction running vehicle, the front vehicle front right direction running vehicle and the front vehicle front right direction running vehicle are classified, and the classified front vehicle front target identification information is sent to the headlamp controller through the whole vehicle CAN bus;

the headlamp controller is respectively connected with the left headlamp driving module and the right headlamp driving module, the left headlamp driving module is connected with the left headlamp lighting module, and the right headlamp driving module is connected with the right headlamp lighting module;

the headlamp controller converts the target identification information in front of the vehicle into a corresponding self-adaptive adjusting instruction, and controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp lighting module and/or the right headlamp lighting module to form a lighting dark area in the lighting range of the left headlamp lighting module and/or the right headlamp lighting module according to the self-adaptive adjusting instruction before the opposite vehicle drives into the lighting range of the left headlamp lighting module and/or the right headlamp lighting module;

the headlamp controller judges whether a target vehicle in front of the vehicle drives out of the illumination dark area or not according to the identification information in front of the vehicle, and if so, the headlamp controller controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp illumination module and/or the right headlamp illumination module to recover to normal illumination; otherwise, adjusting the lighting effect of the left headlamp lighting module and/or the right headlamp lighting module in real time according to the target identification information in front of the vehicle updated in real time.

The intelligent headlamp system of the automobile is also characterized in that the signal processing module obtains the identification information of the front target of the automobile according to the following processes:

step a, carrying out space matching and time matching on the point cloud data and the image data so as to project the point cloud data onto the image data;

b, identifying a front target vehicle in the image data by using a convolutional neural network, and outputting detection frames of the front target vehicle and the image data confidence of each detection frame;

step c, counting the number of point clouds in each detection frame, obtaining the ratio of the number of point clouds in each detection frame to the maximum number of point clouds in each detection frame, and using the ratio as the confidence coefficient of point cloud data of each detection frame;

d, performing weighted calculation on the image data confidence coefficient and the point cloud data confidence coefficient to obtain a fusion confidence coefficient, and determining that no front target vehicle exists in the detection frame when the fusion confidence coefficient is smaller than a set threshold value; when the fusion confidence coefficient is larger than or equal to the threshold value, determining that a front target vehicle exists in the detection frame, and obtaining target identification information in front of the vehicle;

step e, classifying the target identification information in front of the vehicle, including: the vehicle front object identification information after classification is obtained from a vehicle left front co-running vehicle, a vehicle left front counter-running vehicle, a vehicle right front co-running vehicle, and a vehicle right front counter-running vehicle.

The adaptive adjustment instruction comprises the following steps: if the identification information is that the vehicle runs in the same direction in front of the left side of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp illumination module; if the identification information is that the vehicle runs in the left front direction of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp lighting module and the automobile right headlamp lighting module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile right headlamp illumination module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjusting instruction is to reduce the lighting brightness of the left headlamp lighting module and the right headlamp lighting module of the automobile.

Compared with the prior art, the invention has the beneficial effects that:

the intelligent vehicle headlamp system disclosed by the invention utilizes the advantages that the laser point cloud is not limited by illumination conditions, the anti-interference capability is strong, and the detection distance is longer, adopts a data fusion idea that images identify a front target and the point cloud assists in enhancing the detection precision, overcomes the defect of low night imaging quality of a forward-looking camera, effectively solves the problems of false detection, missed detection and the like of the forward-looking camera in the existing intelligent headlamp system, and fully improves the safety of road illumination and the driving comfort of a driver.

Drawings

FIG. 1 is a block diagram of an intelligent headlamp system of a vehicle according to an embodiment of the present invention;

fig. 2 is a flow chart of a fusion algorithm of point cloud data and image data.

Detailed Description

In this embodiment, as shown in fig. 1, an intelligent headlamp system for a vehicle includes: the device comprises a laser radar module, a forward-looking camera module, a signal processing module, a headlamp controller, a whole vehicle CAN bus, a left headlamp driving module, a right headlamp driving module, a left headlamp lighting module and a right headlamp lighting module;

the laser radar module is arranged right above the vehicle and used for acquiring point cloud data in a laser radar detection range in front of the vehicle and sending the point cloud data to the signal processing module through the whole vehicle CAN bus;

the front-view camera module is arranged right above the inside rearview mirror and used for acquiring image data in the visual field range of the front camera of the vehicle and sending the image data to the signal processing module through the whole vehicle CAN bus;

the input of the signal processing module is point cloud data and image data which are sent through a finished automobile CAN bus, data preprocessing is carried out firstly, and as the laser radar and the forward-looking camera are respectively installed at different positions of an automobile body, each sensor defines a coordinate system of the sensor, a world coordinate system, a laser radar coordinate system and a forward-looking camera coordinate system need to be unified. The conversion matrix among the coordinate systems can be obtained by utilizing the Matlab calibration tool box, so that different coordinate systems are converted to a uniform coordinate system, and the spatial matching of point cloud data and image data is completed; then searching image data of the nearest neighbor moment of each frame of point cloud data, projecting the point cloud data onto the image data, and completing time matching of the two data;

then, as shown in fig. 2, the two kinds of data after successful matching are processed by a fusion algorithm, a convolutional neural network is used for detecting image data and identifying a front target vehicle therein, and the confidence degrees alpha of the detection frames of the front target vehicle and the image data of each detection frame are output, wherein alpha is more than or equal to 0 and less than or equal to 1; then counting the number p of the point clouds in each detection frame, and obtaining the number p of the point clouds in each detection frame and the maximum number p of the point clouds in each detection frame₀The ratio beta of beta is more than or equal to 0 and less than or equal to 1, (the maximum point cloud number of each detection frame is a constant and is determined by the size of the detection frame), and the value is the point cloud data confidence beta of each detection frame. Finally, determining the weight omega according to the relative precision of the laser radar and the foresight camera, wherein if the precision of the foresight camera is higher than that of the laser radar, the value of omega is more than 0.5; otherwise, the content is less than or equal to 0.5; then, the weights omega and 1-omega are respectively added into the image data confidence coefficient alpha and the point cloud data confidence coefficient beta, so that the image data confidence coefficient and the point cloud data confidence coefficient of each detection frame are weighted and averaged by using the formula (1), and the fusion confidence coefficient of each detection frame is calculatedWhen fusing confidenceWhen the confidence coefficient is fused, the front target vehicle does not exist in the detection frameAnd if so, determining that the front target vehicle exists in the detection frame.

Finally, classifying the identification information in front of the vehicle, and sending the classified identification information of the target in front of the vehicle to a headlamp controller through a whole vehicle CAN bus;

the whole vehicle CAN bus is respectively connected with the laser radar module, the front-view camera module, the signal processing module and the headlamp controller and is used for acquiring identification information in front of the vehicle;

the headlamp controller is respectively connected with the left headlamp driving module and the right headlamp driving module, the left headlamp driving module is connected with the left headlamp lighting module, and the right headlamp driving module is connected with the right headlamp lighting module;

the headlamp controller converts the target identification information in front of the vehicle into a corresponding self-adaptive adjusting instruction, and controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp lighting module and/or the right headlamp lighting module to form a lighting dark area in the lighting range according to the self-adaptive adjusting instruction before an oncoming vehicle drives into the lighting range of the left headlamp lighting module and/or the right headlamp lighting module;

the headlamp controller judges whether a target vehicle in front of the vehicle drives out of the illumination dark area or not according to the identification information in front of the vehicle, if so, the headlamp controller controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp illumination module and/or the right headlamp illumination module to recover to normal illumination; otherwise, adjusting the lighting effect of the left headlamp lighting module and/or the right headlamp lighting module in real time according to the target identification information in front of the vehicle updated in real time.

In this embodiment, an automobile intelligent headlamp system control system's working process is as follows:

step 1: in the vehicle advancing process, the laser radar module and the front-view camera module work and are respectively used for acquiring point cloud data in a detection range of the laser radar in front of the vehicle and image data in a visual field range of the camera in front of the vehicle;

step 2: the signal processing module sequentially performs space matching and time matching, fusion algorithm processing, classification and output of vehicle front identification information on point cloud data and image data sent by a finished vehicle CAN bus;

and step 3: the signal processing module sends the classified vehicle front identification information to the headlamp controller through a finished vehicle CAN bus, the headlamp controller converts the vehicle front identification information into a corresponding self-adaptive adjusting instruction of the automobile headlamp, and the self-adaptive adjusting instruction comprises the following steps: if the identification information is that the vehicle runs in the same direction in front of the left side of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp illumination module; if the identification information is that the vehicle runs in the left front direction of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile left headlamp lighting module and the automobile right headlamp lighting module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjustment instruction is to reduce the illumination brightness of the automobile right headlamp illumination module; if the identification information is that the vehicle runs in the same direction at the front right of the vehicle, the self-adaptive adjusting instruction is to reduce the lighting brightness of the left headlamp lighting module and the right headlamp lighting module of the automobile.

And 5: the headlamp control controller sends the self-adaptive adjusting instruction to the left headlamp driving module and/or the right headlamp driving module through the whole vehicle CAN bus, and controls the left headlamp driving module and/or the right headlamp driving module to drive the left headlamp lighting module and/or the right headlamp lighting module according to the self-adaptive adjusting instruction, and a lighting dark area is formed before a target vehicle drives into the lighting range of the left headlamp lighting module and/or the right headlamp lighting module.

Step 6: when the target vehicle is driven out of the bright and dark area, the left headlamp driving module is controlled to drive the left headlamp lighting module and/or the right headlamp driving module drives the right headlamp lighting module to recover to normal lighting.