CN111324120A - A method for automatic driving front car cut-in and cut-out scene extraction - Google Patents

Abstract

Translated fromChinese

本发明提供了一种自动驾驶前车切入切出场景提取方法，包括：S1、在本车上安装数据采集设备，用于采集本车和目标车辆的以及车道线的数据信息；S2、结合步骤S1采集的数据信息，判断目标车辆的切入、切出，实现对该处视频或者图片的提取。本发明所述的自动驾驶前车切入切出场景提取方法可直接从场景数据库中获取满足条件的场景。无需人工比照视频来提取换道场景数据，节省人人力和时间成本。

The present invention provides a method for extracting a cut-in and cut-out scene of an automatic driving front vehicle, comprising: S1, installing a data collection device on the vehicle for collecting data information of the vehicle, the target vehicle and the lane line; S2, combining the steps The data information collected by S1 determines the cut-in and cut-out of the target vehicle, and realizes the extraction of the video or picture there. The method for extracting the cut-in and cut-out scene of the automatic driving front vehicle according to the present invention can directly obtain the scene that meets the conditions from the scene database. There is no need to manually compare videos to extract lane change scene data, saving manpower and time costs.

Description

Translated fromChinese

一种自动驾驶前车切入切出场景提取方法A method for automatic driving front car cut-in and cut-out scene extraction

技术领域technical field

本发明属于无人驾驶技术领域，尤其是涉及一种自动驾驶前车切入切出场景提取方法。The invention belongs to the technical field of unmanned driving, and in particular relates to a method for extracting cut-in and cut-out scenes of an automatic driving front vehicle.

背景技术Background technique

目前自动驾驶技术已经成为汽车最前沿的技术之一，美国SAE(汽车工程学会)将自动驾驶技术分为6级，其中0级为报警，如FCW,LDW等，只是在一定情况下给驾驶员声、光提醒，不对车辆进行控制。1级为单向辅助驾驶，如ACC(自适应续航功能)功能，可以辅助车辆进行纵向控制，LKA(车道保持系统)功能，可以辅助车辆进行横向控制。2级为双向辅助驾驶，控制系统可以同时对车辆进行横向和纵向控制。3级为部分自动驾驶功能。4级为特定场景下自动驾驶。5级为完全自动驾驶。At present, autonomous driving technology has become one of the most cutting-edge technologies in automobiles. The American SAE (Society of Automotive Engineering) divides autonomous driving technology into 6 levels, of whichlevel 0 is an alarm, such as FCW, LDW, etc., which are only given to the driver under certain circumstances. Sound and light reminder, do not control the vehicle.Level 1 is one-way assisted driving, such as the ACC (Adaptive Range Function) function, which can assist the vehicle in longitudinal control, and the LKA (Lane Keeping System) function, which can assist the vehicle in lateral control.Level 2 is two-way assisted driving, and the control system can control the vehicle laterally and vertically at the same time. Level 3 is partially autonomous driving. Level 4 is autonomous driving in specific scenarios. Level 5 is fully autonomous driving.

随着自动驾驶等级的逐渐提高，自动驾驶控制系统逐渐代替驾驶员来控制车辆完成驾驶行为。为了保证自动驾驶系统的高安全性和可靠性，越来越多的开发者采用道路测试与仿真测试相结合的方法来验证系统的性能，场景作为道路测试和仿真测试的基础原来越收到重视。目前主要的场景获取方法有自然驾驶行为下场景数据采集、基于驾驶模拟器的极限场景采集、交通事故提出的场景、交通法规规定的测试场景。但获取的数据都是大规模连续的数据，自动驾驶开发者在测试系统时，需要针对特定场景如前车切入，本车变道等进行测试，本发明提出了一种可以快速从大量数据中提出测试特定场景的方法，为自动驾驶系统提供研发、测试场景，加快自动驾驶系统的开发。With the gradual improvement of the level of automatic driving, the automatic driving control system gradually replaces the driver to control the vehicle to complete the driving behavior. In order to ensure the high safety and reliability of the autonomous driving system, more and more developers use a combination of road testing and simulation testing to verify the performance of the system. As the basis for road testing and simulation testing, scenarios have received more and more attention. . At present, the main scene acquisition methods include scene data collection under natural driving behavior, extreme scene collection based on driving simulators, scenarios proposed by traffic accidents, and test scenarios stipulated by traffic regulations. However, the acquired data are large-scale and continuous data. When testing the system, autonomous driving developers need to test for specific scenarios such as the preceding vehicle cutting in, the own vehicle changing lanes, etc. Propose methods for testing specific scenarios, provide R&D and testing scenarios for autonomous driving systems, and speed up the development of autonomous driving systems.

发明内容SUMMARY OF THE INVENTION

有鉴于此，本发明旨在提出一种自动驾驶前车切入切出场景提取方法，以解决现有的方法采用人工提取，通过手动对照视频或图片截取出特定要求的场景，造成提取效率低的问题。In view of this, the present invention aims to propose a method for extracting the cut-in and cut-out scene of an automatic driving front vehicle, so as to solve the problem that the existing method adopts manual extraction, and manually compares videos or pictures to cut out scenes with specific requirements, resulting in low extraction efficiency. question.

为达到上述目的，本发明的技术方案是这样实现的：In order to achieve the above object, the technical scheme of the present invention is achieved in this way:

一种自动驾驶前车切入切出场景提取方法，包括：A method for extracting a cut-in and cut-out scene of an automatic driving front vehicle, comprising:

S1、在本车上安装数据采集设备，用于采集本车和目标车辆的以及车道线的数据信息；S1. Install data collection equipment on the vehicle to collect data information of the vehicle, the target vehicle and the lane lines;

S2、结合步骤S1采集的数据信息，判断目标车辆的切入、切出，实现对该处视频或者图片的提取。S2, in combination with the data information collected in step S1, determine the cut-in and cut-out of the target vehicle, and extract the video or picture there.

进一步的，所述步骤S1中，所述采集设备包括：毫米波雷达、摄像头、陀螺仪传感器、方向盘转角传感器；Further, in the step S1, the collection device includes: a millimeter wave radar, a camera, a gyroscope sensor, and a steering wheel angle sensor;

所述毫米波雷达用于采集本车与目标车之间的横向距离、纵向距离、横向相对速度、纵向相对速度、相对加速度；The millimeter-wave radar is used to collect the lateral distance, longitudinal distance, lateral relative velocity, longitudinal relative velocity, and relative acceleration between the vehicle and the target vehicle;

所述摄像头用于车道线的种类、颜色、本车与车道线的间距；The camera is used for the type, color of the lane line, and the distance between the vehicle and the lane line;

所述陀螺仪传感器用于采集本车车速、本车横摆角速度；The gyroscope sensor is used to collect the vehicle speed and the yaw rate of the vehicle;

所述方向盘转角传感器用于采集本车方向盘的转角。The steering wheel angle sensor is used to collect the rotation angle of the steering wheel of the vehicle.

进一步的，所述摄像头安装在车辆前挡风玻璃上；Further, the camera is installed on the front windshield of the vehicle;

所述毫米波雷达安装在本车前保险杠上。The millimeter-wave radar is installed on the front bumper of the vehicle.

进一步的，所述步骤S2中，识别目标车辆切入切出的方法如下：Further, in the step S2, the method for identifying the cut-in and cut-out of the target vehicle is as follows:

S201、判断本车目前是否为行驶状态，如果则进入步骤S202,如果否继续执行步骤S201；S201, determine whether the vehicle is currently in a driving state, if so, enter step S202, if not, continue to execute step S201;

S202、判断本车的行驶状态，是处于变道状态还是车道内行驶状态；如果是变道状态则返回步骤S201；如果是车道内形式状态，则进入步骤S203；S202. Determine whether the driving state of the vehicle is in a lane-changing state or in a lane-changing state; if it is a lane-changing state, return to step S201; if it is an in-lane state, then enter step S203;

S203、判断是否有目标车出现在本车车道，包括以下两种情况，一种是本车在没有目标物的情况下，相邻车道车辆变道到本车道，另一种是本车有目标车辆的情况下，相邻车道切入一辆相较于目标车辆距离更近的车辆，如果判断出有车辆变更到本车道中，则进行场景提取。S203. Determine whether there is a target vehicle in the lane of the vehicle, including the following two situations, one is that the vehicle in the adjacent lane changes lanes to the current lane when there is no target object, and the other is that the vehicle has a target In the case of a vehicle, the adjacent lane cuts into a vehicle that is closer than the target vehicle. If it is determined that a vehicle has changed to the current lane, scene extraction is performed.

进一步的，所述步骤S201中，判断车辆是否为行驶状态的方法如下：Further, in the step S201, the method for judging whether the vehicle is in a driving state is as follows:

通过判断本车速度进行判断；当自车速度为零时，为停车状态；继续执行步骤S201；The judgment is made by judging the speed of the own vehicle; when the speed of the own vehicle is zero, it is in a parking state; continue to execute step S201;

当本车速度大于零时，为行驶状态，进入步骤S202。When the speed of the own vehicle is greater than zero, it is in the running state, and the process proceeds to step S202.

进一步的，所述步骤S202中利用到车道线的距离变化趋势来判断本车是否靠近左侧或者右侧车道线，进而判断是变道还是车道内行驶，具体方法如下：Further, in the step S202, the change trend of the distance to the lane line is used to determine whether the vehicle is close to the left or right lane line, and then to determine whether to change lanes or drive in the lane. The specific method is as follows:

以车辆车头中间位置为坐标系的原点，副驾驶方向为x轴正向，车辆行驶方向为Y轴正向，建立坐标系，用a1表示距离本车道内左侧车道线的距离，a2表示距离左侧相邻车道车道线的距离，用b1表示距离本车道内右侧车道线的距离，b2表示距离右侧相邻车道车道线的距离；Take the middle position of the front of the vehicle as the origin of the coordinate system, the co-pilot direction is the positive x-axis, and the driving direction of the vehicle is the positive Y-axis, establish a coordinate system, use a1 to represent the distance from the left lane line in the lane, and a2 to represent the distance. The distance from the lane line of the adjacent lane on the left side, with b1 representing the distance from the lane line on the right side of the current lane, and b2 representing the distance from the lane line on the adjacent lane on the right side;

将本车车道线的系数存储为历史车道线矩阵的一行，当历史车道线达到n行，其中n行对应n帧图片，就删除最早的一帧，因此保证历史车道线矩阵中存储了最近的n-1帧车道线数据；Store the coefficient of the lane line of the vehicle as a row of the historical lane line matrix. When the historical lane line reaches n rows, where n rows correspond to n frames of pictures, the earliest frame is deleted, so it is ensured that the latest lane line matrix is stored. n-1 frames of lane line data;

利用历史车道线矩阵，判断当前是车道内行驶还是变道：Use the historical lane line matrix to determine whether it is currently driving in the lane or changing lanes:

首先从当前所有的车道线中找到本车车道线，对比a1,a2大小，最大的为本车道左车道线，对比b1,b2,最小的为本车道右车道线；First, find the own lane line from all the current lane lines, compare the size of a1, a2, the largest is the left lane line of the current lane, compare b1, b2, the smallest is the right lane line of the current lane;

利用距离车道线的距离变化趋势判断车辆是否有变道、车道偏离的趋势；Use the change trend of the distance from the lane line to determine whether the vehicle has a tendency to change lanes or deviate from the lane;

其中，变道状态的判断方法如下：Among them, the method of judging the lane change status is as follows:

左车道线中，距离左车道线的距离|a1|变小，且a1从负值变成正值，此车道线从左车道线变为右车道线，则判断为有向左进行了变道；In the left lane line, the distance |a1| from the left lane line becomes smaller, and a1 changes from a negative value to a positive value, and the lane line changes from the left lane line to the right lane line, then it is judged that there is a lane change to the left ;

右车道线中，距离左车道线的距离|b1|变小，且b1从正值变成负值，此车道线从右车道线变为左车道线，则判断为向右进行了变道；In the right lane line, the distance |b1| from the left lane line becomes smaller, and b1 changes from a positive value to a negative value, and the lane line changes from the right lane line to the left lane line, then it is judged that the lane has changed to the right;

其中，车道内行驶状态的判断方法如下：Among them, the method of judging the driving state in the lane is as follows:

左车道线中，距离左车道线的距离|a1|变小，且a1没有从负值变成正值；右车道线中，距离右车道线的距离|b1|变大，则判断为有向左进行车道偏离的趋势，但仍未发生变道的动作，判定为车道内行驶；In the left lane line, the distance |a1| from the left lane line becomes smaller, and a1 does not change from a negative value to a positive value; in the right lane line, the distance from the right lane line |b1| If there is a tendency to deviate from the lane on the left, but there is no lane change, it is determined to be driving in the lane;

右车道线中，距离左车道线的距离|b1|变小，且b1没有从正值变成负值；左车道线中，距离右车道线的距离|a1|变大，则判断为有向右进行车道偏离的趋势，但仍未发生变道的动作，判定为车道内行驶。In the right lane line, the distance |b1| from the left lane line becomes smaller, and b1 does not change from a positive value to a negative value; in the left lane line, the distance from the right lane line |a1| The vehicle has a tendency to deviate from the lane on the right, but no lane change has occurred, and it is determined to be driving in the lane.

进一步的，所述步骤S203中，判断目标车切入切出的方法如下：Further, in the step S203, the method for judging that the target vehicle cuts in and out is as follows:

在车道保持状态下，进一步判断前车的状态，划分为自由行驶、跟车、前车切入、前车切出，具体的判断方法如下：In the lane keeping state, the state of the preceding vehicle is further judged, which is divided into free driving, following, cutting in, and cutting out of the preceding vehicle. The specific judgment method is as follows:

提取本车道内的最近目标Extract the nearest object in this lane

在本车为车道线内行驶状态下，利用毫米波雷达探测到的本车与目标物信息的横向距离Range_X，纵向距离Range_Y来判断；首先判断Range_X，筛选出Range_X在[-2m,2m]范围之内的目标物，之后选取Range_Y最小的目标物，存入历时目标物信息矩阵中的一行，4个变量分别为objectID、Range_X,、Range_Y,、距离目Relative Speed_X；设定Range_Y最大不超过60m；When the vehicle is driving within the lane line, the horizontal distance Range_X and the vertical distance Range_Y between the vehicle and the target detected by the millimeter-wave radar are used to judge; first judge the Range_X, and filter out the Range_X in the range of [-2m, 2m] Then select the target with the smallest Range_Y, and store it in a row in the lasting target information matrix. The four variables are objectID, Range_X, Range_Y, and Relative Speed_X of the distance target; the maximum Range_Y is set to be no more than 60m ;

当历史目标物信息矩阵数量达到n行，就删除最早的一帧，因此保证历史目标物信息矩阵中存储了最近的n-1帧最新目标物的信息；When the number of historical target information matrix reaches n rows, the earliest frame is deleted, so it is ensured that the latest n-1 frames of the latest target information are stored in the historical target information matrix;

利用历史目标信息，通过判断Range_Y的变化量，考虑当前和上一帧的最近目标是否为同一目标；若Range_Y的变化量没有发生突变，绝对值小于设定阈值，则为同一目标，则当前为跟车状态；若Range_Y的变化量绝对值大于设定的阈值，则判定为不是同一目标，进一步判断，若Range_Y的变化量大于设定阈值，则初步判断为前车切出；若Range_Y的变化量小于设定阈值，则初步判断为前车切入，若当前没有本车道目标，而上一帧存在历史目标，则判断为前车切出或驶离，本车自由行驶。Using the historical target information, by judging the amount of change in Range_Y, consider whether the recent target of the current and the previous frame is the same target; if the amount of change in Range_Y does not change abruptly, and the absolute value is less than the set threshold, it is the same target, and the current is Car-following state; if the absolute value of the change of Range_Y is greater than the set threshold, it is judged that it is not the same target, and further judgement, if the change of Range_Y is greater than the set threshold, it is preliminarily judged that the preceding vehicle cuts out; if the change of Range_Y If the amount is less than the set threshold, it is preliminarily judged that the preceding vehicle cuts in. If there is no current lane target, but there is a historical target in the previous frame, it is judged that the preceding vehicle cuts out or leaves, and the vehicle runs freely.

相对于现有技术，本发明所述的自动驾驶前车切入切出场景提取方法具有以下优势：Compared with the prior art, the method for extracting the cut-in cut-out scene of the automatic driving front vehicle according to the present invention has the following advantages:

本发明所述的自动驾驶前车切入切出场景提取方法可直接从场景数据库中获取满足条件的场景。无需人工比照视频来提取换道场景数据，节省人人力和时间成本。The method for extracting the cut-in and cut-out scene of the automatic driving front vehicle according to the present invention can directly obtain the scene that meets the conditions from the scene database. There is no need to manually compare videos to extract lane-changing scene data, saving manpower and time costs.

附图说明Description of drawings

构成本发明的一部分的附图用来提供对本发明的进一步理解，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：The accompanying drawings constituting a part of the present invention are used to provide further understanding of the present invention, and the exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the attached image:

图1为本发明实施例所述的自动驾驶前车切入切出场景提取方法流程图；FIG. 1 is a flowchart of a method for extracting a cut-in cut-out scene of an automatic driving front vehicle according to an embodiment of the present invention;

图2为本发明实施例所述的车道线距离信息示意图；2 is a schematic diagram of lane line distance information according to an embodiment of the present invention;

图3为本发明实施例所述的判断本车是变道还是车道内行驶的过程流程图；3 is a flowchart of a process for judging whether the vehicle is changing lanes or driving in a lane according to an embodiment of the present invention;

图4为本发明实施例所述的本车及目标车对应位置示意图；4 is a schematic diagram of the corresponding positions of the own vehicle and the target vehicle according to an embodiment of the present invention;

图5为本发明实施例所述的选取本车道最近目标物的过程流程图；FIG. 5 is a flowchart of a process for selecting the nearest target in the lane according to an embodiment of the present invention;

图6为本发明实施例所述的判断目标物是稳定跟随、切入、切出状态的流程图。FIG. 6 is a flowchart of determining whether the target object is in a stable following, cut-in, or cut-out state according to an embodiment of the present invention.

具体实施方式Detailed ways

需要说明的是，在不冲突的情况下，本发明中的实施例及实施例中的特征可以相互组合。It should be noted that the embodiments of the present invention and the features of the embodiments may be combined with each other under the condition of no conflict.

在本发明的描述中，需要理解的是，术语“中心”、“纵向”、“横向”、“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”、“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。此外，术语“第一”、“第二”等仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”等的特征可以明示或者隐含地包括一个或者更多个该特征。在本发明的描述中，除非另有说明，“多个”的含义是两个或两个以上。In the description of the present invention, it should be understood that the terms "center", "portrait", "horizontal", "top", "bottom", "front", "rear", "left", "right", " The orientation or positional relationship indicated by vertical, horizontal, top, bottom, inner, outer, etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and The description is simplified rather than indicating or implying that the device or element referred to must have a particular orientation, be constructed and operate in a particular orientation, and therefore should not be construed as limiting the invention. In addition, the terms "first", "second", etc. are used for descriptive purposes only, and should not be construed as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature defined as "first", "second", etc., may expressly or implicitly include one or more of that feature. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

在本发明的描述中，需要说明的是，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域的普通技术人员而言，可以通过具体情况理解上述术语在本发明中的具体含义。In the description of the present invention, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense, unless otherwise expressly specified and limited, for example, it may be a fixed connection or a detachable connection Connection, or integral connection; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be internal communication between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood through specific situations.

下面将参考附图并结合实施例来详细说明本发明。The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

通过在行驶车辆上搭载前视摄像头、毫米波雷达、陀螺仪、方向盘转角传感器获取前面目标物信息，车道线信息，本车运动相关的参数。具体输出信息包括如下表：By installing a forward-looking camera, millimeter-wave radar, gyroscope, and steering wheel angle sensor on the driving vehicle to obtain the information of the front target, lane line information, and parameters related to the movement of the vehicle. The specific output information includes the following table:

数据场景库构建，通过在行驶车辆上搭载前视摄像头、毫米波雷达，陀螺仪、方向盘转角等传感器。其中前视摄像头安装在车辆前挡风玻璃上，可以识别实际路况下的车道线并发出本车所在车道线的距离，相邻车道线的距离，前期经过标定，确保输出信号的准确性。毫米波雷达安装在前保险杠的位置。陀螺仪对安装位置没有要求，可以放在车辆内部。方向盘转角传感器可以输出自车方向盘转角信息。所有传感器输出的信号保持同步。The data scene library is constructed by installing sensors such as front-view camera, millimeter-wave radar, gyroscope, and steering wheel angle on the driving vehicle. The front-view camera is installed on the front windshield of the vehicle, which can identify the lane line under the actual road conditions and send out the distance of the lane line where the vehicle is located, and the distance between the adjacent lane lines. The millimeter-wave radar is installed in the position of the front bumper. The gyroscope has no requirements on the installation position and can be placed inside the vehicle. The steering wheel angle sensor can output the steering wheel angle information of the vehicle. The signals output by all sensors remain synchronized.

如图1所示，具有了上述所列信号之后，分三步实现前车切入场景的提取。第一步是判断本车目前是前向行驶状态；第二步是在确认本车是在前向行驶的情况下，判断本车的行驶状态：变道、车道内行驶；第三步是确定是否有更近的目标物出现在本车车道，包括：本车在没有目标物的情况下，相邻车道车辆变道到本车道，本车有目标物的情况下，相邻车道切入一辆距离更近的车辆。As shown in Figure 1, after having the signals listed above, the extraction of the preceding vehicle cut-in scene is realized in three steps. The first step is to determine whether the vehicle is currently driving in a forward direction; the second step is to determine the driving status of the vehicle when it is confirmed that the vehicle is driving in a forward direction: changing lanes, driving in lanes; the third step is to determine Whether there is a closer target in the lane of the vehicle, including: when the vehicle has no target, the vehicle in the adjacent lane changes to the lane, and when the vehicle has a target, the adjacent lane cuts into a vehicle closer vehicles.

以下具体对每个过程详细描述，Each process is described in detail below,

(1)第一步：判断本车是否是行驶状态(1) The first step: determine whether the vehicle is in a driving state

通过判断本车速度；By judging the speed of the vehicle;

当自车速度为零时，为停车状态。When the self-vehicle speed is zero, it is in the stop state.

当自车速度大于零时，为前向行驶状态，进入第二步。When the self-vehicle speed is greater than zero, it is in the forward driving state and enters the second step.

(2)第二步：本车行驶趋势(2) The second step: the driving trend of the vehicle

本部分利用到车道线的距离变化趋势来判断本车是否靠近左侧或者右侧车道线。将本车行驶趋势分成变道、车道内行驶两个状态。This part uses the change trend of the distance to the lane line to judge whether the vehicle is close to the left or right lane line. The driving trend of the vehicle is divided into two states: changing lanes and driving in lanes.

规定本车坐标系符合右手定则标准，规定车辆车头中间位置为坐标系的原点，副驾驶方向为x正向，车辆行驶方向为Y正向。用a1表示距离本车道内左侧车道线的距离，a2表示距离左侧相邻车道车道线的距离，用b1表示距离本车道内右侧车道线的距离，b2表示距离右侧相邻车道车道线的距离；如图2所示。It is stipulated that the coordinate system of the vehicle conforms to the standard of the right-hand rule, and the center position of the front of the vehicle is defined as the origin of the coordinate system, the direction of the co-pilot is the positive x direction, and the direction of the vehicle is the positive direction of the Y. Use a1 to represent the distance from the left lane line in the current lane, a2 to represent the distance to the lane line of the adjacent lane on the left, use b1 to represent the distance from the right lane line in the current lane, and b2 to represent the distance from the adjacent lane on the right distance of the line; as shown in Figure 2.

将自车道线的系数存储为历史车道线矩阵的一行，4个数字(a1，a2，b1，b2)，当历史车道线达到10行(对应10帧图片)(n，可根据情况选择)，就删除最早的一帧，因此保证历史车道线矩阵中存储了最近的9(n-1)帧车道线数据。Store the coefficient of the own lane line as one row of the historical lane line matrix, 4 numbers (a1, a2, b1, b2), when the historical lane line reaches 10 lines (corresponding to 10 frames of pictures) (n, can be selected according to the situation), The earliest frame is deleted, so it is guaranteed that the latest 9(n-1) frames of lane line data are stored in the historical lane line matrix.

利用历史车道线矩阵，就可以判断当前是车道内行驶还是变道。利用历史车道线的系数变化即可实现这一功能：Using the historical lane line matrix, it is possible to determine whether it is currently driving in the lane or changing lanes. This function can be achieved by using the coefficient change of the historical lane lines:

如图3所示，首先从当前所有的车道线中找到自车道线。对比a1,a2大小，最大的为本车道左车道线。对比b1,b2,最小的为本车道右车道线。As shown in Figure 3, firstly find the own lane line from all the current lane lines. Comparing the sizes of a1 and a2, the largest is the left lane line of this lane. Comparing b1 and b2, the smallest is the right lane line of this lane.

利用距离车道线的距离变化趋势判断车辆是否有变道、车道偏离的趋势。Use the change trend of the distance from the lane line to determine whether the vehicle has a tendency to change lanes or deviate from the lane.

变道：Lane change:

1)左车道线中，距离左车道线的距离|a1|变小，且a1从负值变成正值，此车道线从左车道线变为右车道线，则判断为有向左进行了变道。1) In the left lane line, the distance |a1| from the left lane line becomes smaller, and a1 changes from a negative value to a positive value, and the lane line changes from the left lane line to the right lane line, then it is judged that there is a leftward movement. Lane change.

2)右车道线中，距离左车道线的距离|b1|变小，且b1从正值变成负值，此车道线从右车道线变为左车道线，则判断为向右进行了变道。2) In the right lane line, the distance |b1| from the left lane line becomes smaller, and b1 changes from a positive value to a negative value, and the lane line changes from the right lane line to the left lane line, then it is judged that the change to the right has been made. road.

车道内行驶：Driving in the lane:

1)左车道线中，距离左车道线的距离|a1|变小，且a1没有从负值变成正值；右车道线中，距离右车道线的距离|b1|变大，则判断为有向左进行车道偏离的趋势，但仍未发生变道的动作，判定为车道内行驶。1) In the left lane line, the distance |a1| from the left lane line becomes smaller, and a1 does not change from a negative value to a positive value; in the right lane line, the distance |b1| from the right lane line becomes larger, it is judged as There is a tendency to deviate from the lane to the left, but no lane change has occurred, and it is determined to be driving in the lane.

2)右车道线中，距离左车道线的距离|b1|变小，且b1没有从正值变成负值；左车道线中，距离右车道线的距离|a1|变大，则判断为有向右进行车道偏离的趋势，但仍未发生变道的动作，判定为车道内行驶。2) In the right lane line, the distance |b1| from the left lane line becomes smaller, and b1 does not change from a positive value to a negative value; in the left lane line, the distance |a1| from the right lane line becomes larger, it is judged as There is a tendency to deviate from the lane to the right, but there is no lane change, and it is determined that the vehicle is driving in the lane.

3)具体左右车道线的距离都没有发生明显偏差，判断为车道内行驶。3) If there is no obvious deviation in the distance between the left and right lane lines, it is judged as driving in the lane.

(3)第三步：考虑是否有更近的目标物出现(3) Step 3: Consider whether there is a closer target

在车道保持状态下，进一步考虑前车的状态，划分为自由行驶、跟车、前车切入、前车切出几种。In the lane keeping state, the state of the preceding vehicle is further considered, and it is divided into free driving, following vehicle, preceding vehicle cutting in, and preceding vehicle cutting out.

1)提取本车道内的最近目标1) Extract the nearest target in this lane

在本车为车道线内行驶状态下，利用毫米波雷达探测到的本车与目标物信息的横向距离Range_X，纵向距离Range_Y来判断。首先判断Range_X，筛选出Range_X在[-2m,2m]范围之内的目标物，之后选取Range_Y最小的目标物，存入历时目标物信息矩阵中的一行，4个变量(objectID,Range_X,Range_Y,距离目Relative Speed_X)。设定Range_Y最大不超过60m(根据摄像头感知距离的精度设定，可修改，超多60m则认定本车为自由行驶状态。针对不同传感器，可设置不同的距离限制)，示意图如图4所示。When the vehicle is driving within the lane line, the horizontal distance Range_X and the vertical distance Range_Y of the vehicle and the target information detected by the millimeter-wave radar are used to judge. First judge Range_X, filter out the targets whose Range_X is within the range of [-2m, 2m], then select the target with the smallest Range_Y, and store it in a row in the temporal target information matrix, 4 variables (objectID, Range_X, Range_Y, Distance target Relative Speed_X). Set Range_Y to a maximum of no more than 60m (it can be modified according to the accuracy of the perceived distance of the camera. If it exceeds 60m, the vehicle is considered to be in a free driving state. Different distance limits can be set for different sensors), as shown in Figure 4. .

当历史目标物信息矩阵数量达到10行(对应10帧图片)(n，可根据情况选择)，就删除最早的一帧，因此保证历史目标物信息矩阵中存储了最近的9(n-1)帧最新目标物的信息。流程图如图5。When the number of historical target information matrix reaches 10 lines (corresponding to 10 frames of pictures) (n, which can be selected according to the situation), the earliest frame is deleted, so it is ensured that the most recent 9 (n-1) are stored in the historical target information matrix. Frame the latest target information. The flowchart is shown in Figure 5.

2)自由行驶、跟车、切入、切出情形的划分(如图6所示)2) The division of free driving, following, cutting in, and cutting out (as shown in Figure 6)

利用历史目标信息，通过判断Range_Y的变化量，考虑当前和上一帧的最近目标是否为同一目标。若Range_Y的变化量没有发生突变，如Range_Y的变化量绝对值小于2m(可根据情况设定门槛值)则为同一目标，则当前为跟车状态。若Range_Y的变化量绝对值大于设定的门槛值，则判定为不是同一目标，进一步判断，若Range_Y的变化量大于2m(门槛值)，则初步判断为前车切出；若Range_Y的变化量小于-2m(门槛值)，则初步判断为前车切入，若当前没有本车道目标，而上一帧存在历史目标，则判断为前车切出或驶离，本车自由行驶。Using historical target information, by judging the amount of change in Range_Y, consider whether the nearest target of the current and previous frame is the same target. If there is no sudden change in the change of Range_Y, if the absolute value of the change of Range_Y is less than 2m (the threshold value can be set according to the situation), it is the same target, and the current state is following the vehicle. If the absolute value of the change of Range_Y is greater than the set threshold value, it is judged that it is not the same target, and further judgment is made. If the change of Range_Y is greater than 2m (threshold value), it is preliminarily judged that the preceding vehicle is cut out; if the change of Range_Y If it is less than -2m (threshold value), it is preliminarily judged that the preceding vehicle cuts in. If there is no current lane target, but there is a historical target in the previous frame, it is judged that the preceding vehicle cuts out or leaves, and the vehicle runs freely.

以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the present invention. within the scope of protection.

Claims (7)

1. A cut-in and cut-out scene extraction method for an automatic driving front vehicle is characterized by comprising the following steps:

s1, installing data acquisition equipment on the vehicle, and acquiring data information of the vehicle, the target vehicle and the lane line;

and S2, judging the cut-in and cut-out of the target vehicle by combining the data information collected in the step S1, and extracting the video or the picture at the position.

2. The automatic driving front vehicle cut-in and cut-out scene extraction method according to claim 1, wherein in the step S1, the acquisition device includes: the device comprises a millimeter wave radar, a camera, a gyroscope sensor and a steering wheel corner sensor;

the millimeter wave radar is used for acquiring the transverse distance, the longitudinal distance, the transverse relative speed, the longitudinal relative speed and the relative acceleration between the vehicle and the target vehicle;

the camera is used for the type and color of the lane line and the distance between the vehicle and the lane line;

the gyroscope sensor is used for acquiring the speed and the yaw rate of the vehicle;

the steering wheel corner sensor is used for collecting the corner of the steering wheel of the vehicle.

3. The automatic driving front vehicle cut-in and cut-out scene extraction method according to claim 2, characterized in that: the camera is arranged on the front windshield of the vehicle;

the millimeter wave radar is installed on the front bumper of the vehicle.

4. The method for extracting cut-in and cut-out scene of the autonomous driving vehicle according to claim 1, wherein in step S2, the method for identifying the cut-in and cut-out of the target vehicle is as follows:

s201, judging whether the vehicle is in a running state at present, if so, entering a step S202, and if not, continuing to execute the step S201;

s202, judging whether the running state of the vehicle is in a lane changing state or a driving state in a lane; if the lane change state is detected, returning to the step S201; if the lane form state is the in-lane form state, the step S203 is entered;

and S203, judging whether a target vehicle appears in the lane of the vehicle, wherein the adjacent lane vehicle changes the lane to the lane when the vehicle does not have the target object, the adjacent lane vehicle cuts in a vehicle which is closer to the target vehicle when the vehicle has the target vehicle, and the scene extraction is carried out if the adjacent lane vehicle cuts in a vehicle which is closer to the target vehicle.

5. The automatic driving front vehicle cut-in and cut-out scene extraction method according to claim 4, characterized in that: in step S201, the method for determining whether the vehicle is in a running state is as follows:

judging by judging the speed of the vehicle; when the speed of the bicycle is zero, the bicycle is in a parking state; continuing to execute step S201;

when the vehicle speed is greater than zero, the vehicle is in a traveling state, and the process proceeds to step S202.

6. The automatic driving front vehicle cut-in and cut-out scene extraction method according to claim 4, characterized in that: in step S202, whether the vehicle approaches the left or right lane line is determined by using the trend of the change in the distance to the lane line, and then whether the vehicle is changing lanes or driving in a lane is determined, which includes the following specific steps:

establishing a coordinate system by taking the middle position of the head of the vehicle as the origin of the coordinate system, the copilot direction as the positive x-axis direction and the vehicle running direction as the positive Y-axis direction, wherein a1 represents the distance from the lane line on the left side in the lane, a2 represents the distance from the lane line on the adjacent left side, b1 represents the distance from the lane line on the right side in the lane, and b2 represents the distance from the lane line on the adjacent right side;

storing the coefficient of the lane line of the vehicle as one row of a historical lane line matrix, and deleting the earliest frame when the historical lane line reaches n rows, wherein the n rows correspond to n frames of pictures, so that the latest n-1 frames of lane line data are stored in the historical lane line matrix;

and judging whether the current driving in the lane or lane changing is performed by utilizing the historical lane line matrix:

firstly, finding out a lane line of a vehicle from all current lane lines, comparing the sizes of a1 and a2, wherein the largest line is the left lane line of the vehicle, and comparing b1 and b2, the smallest line is the right lane line of the vehicle;

judging whether the vehicle has the tendency of lane changing and lane departure by using the distance change tendency of the distance lane line;

the method for judging the lane change state comprises the following steps:

if the distance | a1| from the left lane line is smaller, a1 changes from a negative value to a positive value, and the lane line changes from the left lane line to the right lane line, it is determined that the lane change is performed to the left;

in the right lane line, the distance | b1| from the left lane line becomes smaller, b1 changes from a positive value to a negative value, and the lane line changes from the right lane line to the left lane line, so that the lane change is determined to be performed rightward;

the method for judging the driving state in the lane comprises the following steps:

in the left lane line, the distance | a1| from the left lane line becomes small, and a1 does not change from a negative value to a positive value; if the distance | b1| from the right lane line is larger in the right lane line, it is determined that the lane departure is proceeding to the left, but the lane change operation is not occurring, and it is determined that the vehicle is traveling in the lane;

in the right lane line, the distance | b1| from the left lane line becomes small, and b1 does not change from a positive value to a negative value; if the distance | a1| from the right lane line is large in the left lane line, it is determined that the lane change operation has not occurred although the lane change tends to be performed rightward, and it is determined that the vehicle is traveling in the lane.

7. The method for extracting cut-in and cut-out scene of the pre-autonomous vehicle according to claim 4, wherein in step S203, the method for determining the cut-in and cut-out of the target vehicle is as follows:

in the lane keeping state, the state of the preceding vehicle is further judged and divided into free running, following, preceding vehicle cut-in and preceding vehicle cut-out, and the specific judgment method is as follows:

extracting nearest target in own lane

Judging the transverse distance Range _ X and the longitudinal distance Range _ Y of the information between the vehicle and the target object, which are detected by a millimeter wave radar, when the vehicle is in the driving state in the lane line; firstly, judging Range _ X, screening out the target objects with Range _ X in the Range of [ -2m,2m ], then selecting the target object with the smallest Range _ Y, and storing the target object into a row in a duration target object information matrix, wherein 4 variables are object ID, Range _ X, Range _ Y and distance target related Speed _ X respectively; setting Range _ Y not to exceed 60m at maximum;

when the number of the historical target object information matrixes reaches n rows, deleting the earliest frame, thereby ensuring that the latest n-1 frames of latest target object information are stored in the historical target object information matrixes;

by utilizing historical target information, whether the current and the latest target of the previous frame are the same target or not is considered by judging the variation of Range _ Y; if the variation of Range _ Y is not mutated and the absolute value is smaller than the set threshold, the same target is set, and the vehicle following state is set currently; if the absolute value of the variation of the Range _ Y is larger than a set threshold, judging that the target is not the same target, and further judging, if the absolute value of the variation of the Range _ Y is larger than the set threshold, primarily judging that the target is cut out of the front vehicle; if the variation of Range _ Y is smaller than the set threshold, it is preliminarily determined that the vehicle is cut in, and if there is no current lane target and there is a history target in the previous frame, it is determined that the vehicle is cut out or driven away, and the vehicle is free to run.

Images