KR101826628B1 - Preceding vehicle recognizing method and apparatus - Google Patents

Abstract

The present invention relates to a method and an apparatus to recognize a preceding vehicle. The present invention includes: a step in which a control unit calculates position information of objects and traffic lane information by processing an image inputted through a camera unit; a step in which the control unit generates a trajectory by respectively tracking the objects; a step in which the control unit checks if an overlapping trajectory is generated in the objects; a step in which the control unit estimates a distance between objects and left/right traffic lanes based on a host vehicle when the traffic lane information is calculated, and calculates probability that the object is on the traffic lane having a host vehicle; and a step in which the control unit estimates a preceding vehicle of the host vehicle based on one or more of predetermined information which is calculated or generated, when the probability of a middle traffic lane where the host vehicle exists among position probabilities of each object exceeds a predetermined threshold value, recognizes the corresponding object as a preceding vehicle, and recognizes a near object as a preceding vehicle in a case of two or more objects. The present invention is provided to stabilize traffic lane recognition and recognize a preceding vehicle based on the position probability of a vehicle when having traffic lane information and recognize a preceding vehicle by using the trajectory of the preceding vehicle when having no vehicle information.

Description

[0001] PRECEDING VEHICLE RECOGNIZING METHOD AND APPARATUS [0002]

More particularly, the present invention relates to a method and apparatus for recognizing a preceding vehicle, and more particularly, to a method for recognizing a preceding vehicle, And more particularly, to a method and apparatus for recognizing a preceding vehicle that can recognize a preceding vehicle using a trajectory.

Recently launched vehicles are equipped with a Smart Cruise Control system. The smart cruise control system uses a radar sensor to measure distance and relative speed with respect to the preceding vehicle, thereby maintaining the inter-vehicle distance automatically and constantly.

For example, a vehicle equipped with a Smart Cruise Control system automatically operates the brakes when the preceding vehicle is slow, keeps braking without turning off if necessary, and automatically accelerates the vehicle when the preceding vehicle is accelerating, When the cornering is carried out, a stable function of recognizing the rotation angle of the steering wheel and recognizing that there is no preceding vehicle by mistake during cornering and not accelerating is included.

In the smart cruise control system as well as the autonomous travel system and the augmented reality HUD (Head Up Display) system, a host vehicle (i.e., a car) among the objects Is very important.

However, in the past, in detecting lane information to distinguish a preceding vehicle (that is, in detecting a forward vehicle traveling on the same lane), when lane information detected using a camera has a lot of noise and can not recognize a lane There is a limit in recognizing the preceding vehicle (i.e., the vehicle traveling in front of the same lane on which the vehicle is traveling).

Therefore, in order to solve such a problem, there is a demand for a method for more reliable and accurate recognition of the preceding vehicle.

BACKGROUND ART [0002] The background art of the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2009-0039501 (published on April 22, 2009, method of detecting a preceding vehicle for safe driving).

According to an aspect of the present invention, there is provided an information processing apparatus, which is created in order to solve the above-described problems, and which stabilizes lane recognition when there is lane information, recognizes a preceding vehicle based on a location probability of the vehicle, The present invention is directed to a method and apparatus for recognizing a preceding vehicle that can recognize a preceding vehicle using a trajectory of a preceding vehicle.

According to an aspect of the present invention, there is provided a preceding vehicle recognition method, comprising: a control unit processing an image input through a camera unit to calculate lane information and position information of objects; The control unit tracing the objects to generate a trajectory; Checking whether a superimposed trajectory is generated in the objects; Estimating a distance between the left / right lane and each object based on the host vehicle, and calculating a probability that the object is in a lane with the host vehicle, when the lane information is calculated; And estimating a preceding vehicle of the host vehicle based on at least one or more previously designated information calculated or generated by the control unit. When the probability of the host vehicle is greater than or equal to a predetermined threshold, Recognizing an object as a preceding vehicle, and recognizing a nearby object as a preceding vehicle when the number of objects is two or more.

In the present invention, the at least one previously designated information includes at least one of lane information, trajectory information of an object, and corresponding lane information with an object.

In the present invention, the lane information and the object position information are information calculated based on the host vehicle.

In the present invention, in order to generate the trajectory of the object, the control unit performs coordinate transformation to create a relative coordinate system at a start position of the host vehicle, and the control unit controls the movement amount and heading information of the host vehicle And calculates object position information and lane information in the relative coordinate system.

In the present invention, the movement amount and heading information of the host vehicle are calculated using the vehicle speed sensor information, the steering angle sensor information, and the yaw sensor information.

In the present invention, in order to distinguish at least one object from each other, the control unit can designate an ID for each object, and can designate a color for each object for display of the preceding vehicle.

According to the present invention, in order to stabilize the lane information, the controller removes noise of a lane using a predetermined filter, reduces a capacity of a storage space, and increases And sampling is performed at regular intervals.

In the present invention, the control unit tracks objects between frames of a continuous image input through the camera unit based on distance and trajectory direction, masks a continuous object as a 'tracking object', and generates a newly generated object Masking the object as a 'generated object', and masking each object as a 'disappearing object'.

In the present invention, in the step of tracing the objects to generate trajectory, the trajectory represents a movement path of the object, and the control unit reduces the capacity of the storage space required for generating the trajectory for the tracing object, The moving path of the object is sampled at predetermined predetermined intervals.

In the present invention, the step of checking whether the overlapping trajectory is generated may include the steps of finding two closest trajectory samples that are closest to the position of an object, calculating a distance between the object and the line segment connecting the two trajectory samples, It is determined whether or not it is a superimposed trajectory according to whether or not it is less than the set threshold value.

In the present invention, in the step of calculating the probability, the control unit calculates a probability with a normal distribution at a position of each object based on the distance between the estimated object and the lane.

In the present invention, if the lane information is not detected or the preceding vehicle can not be recognized based on probability, the control unit selects a preceding vehicle candidate object based on a trajectory, Checks whether the host vehicle overlaps the locus when the preceding vehicle candidate object overlaps the locus of another object and checks whether the host vehicle overlaps the locus. If the host vehicle also overlaps the locus, the host vehicle judges the candidate object as a preceding vehicle And when the host vehicle does not overlap the corresponding trajectory, the candidate object is not determined as a preceding vehicle.

In the present invention, in order to select a preceding vehicle candidate on the basis of the trajectory, the control unit calculates a distance from the trajectory of each object at the host vehicle position to find trajectories that fall below a predetermined threshold value, The trajectory of the trajectory is selected as the trajectory of the preceding vehicle and the trajectory of the trajectory of the other trajectory is selected as the preceding vehicle candidate object if the trajectory is less than or equal to the threshold value .

According to another aspect of the present invention, there is provided a preceding vehicle recognizing apparatus including: a camera section for photographing a forward image of a host vehicle; A sensor unit including a vehicle speed sensor, a steering angle sensor, and a yaw sensor for detecting movement amount and heading information of the host vehicle; Estimating objects in the lane on which the host vehicle exists among the objects, estimating objects in the lane on which the host vehicle exists, processing the forward image to calculate lane information and position information of the objects, And a controller for recognizing the object closest to the vehicle as a preceding vehicle.

In the present invention, the controller estimates the distance between the left and right lane and each object based on the host vehicle, calculates the probability that the object is in the lane with the host vehicle, The object is recognized as a preceding vehicle if the probability is greater than or equal to a predetermined threshold value, and the nearby object is recognized as a preceding vehicle when the object is two or more.

In the present invention, the object is a vehicle that runs on the front side of the host vehicle, the object that runs on the front of the host vehicle is a preceding vehicle, and all the objects other than the object corresponding to the preceding vehicle are all vehicles .

According to another aspect of the present invention, there is provided a preceding vehicle recognizing apparatus including: a camera section for photographing a forward image of a host vehicle; A sensor unit including a vehicle speed sensor, a steering angle sensor, and a yaw sensor for detecting movement amount and heading information of the host vehicle; The control unit may process the forward image to calculate lane information and position information of the objects, generate trajectories by tracing the objects, check whether overlapping trajectories are generated in the objects, Estimates the distance between the object and the object, calculates the probability that the object is present in the host vehicle, estimates the host vehicle's preceding vehicle based on the estimated lane information, object locus information, A controller for recognizing the object as a preceding vehicle if the probability of the host vehicle is greater than or equal to a predetermined threshold, and recognizing a nearby object as a preceding vehicle when there are two or more objects; And a control unit.

In the present invention, the controller may perform a coordinate transformation to generate a locus of the object to create a relative coordinate system at a start position of the host vehicle, compensate for the amount of movement and heading information of the host vehicle, And calculates object position information and lane information in a coordinate system.

In the present invention, in order to stabilize the lane information, the control unit removes noise of the lane information by using an exponential filter or a Kalman filter, and further reduces the capacity of the storage space , And the lanes are sampled at predetermined intervals in order to increase the efficiency of the calculation process.

In the present invention, in order to check whether the superimposed trajectory is generated, the control unit finds two trajectory samples closest to the position of an object, calculates a distance between the object and the line segment connecting the two trajectory samples And determining whether or not the superimposed trajectory is a trajectory according to whether or not the trajectory is less than or equal to a predetermined threshold value.

In the present invention, in calculating the probability that the object is in a lane with the host vehicle, the control unit calculates the probability from the position of each object to the normal distribution based on the distance between the estimated object and the lane .

In the present invention, if the lane information is not detected or the preceding vehicle can not be recognized based on the probability, the controller selects a preceding vehicle candidate based on a trajectory, and if the preceding vehicle candidate object is a trajectory of another object Checking whether the host vehicle overlaps the locus when the trajectory of another object overlaps the trajectory of the other object, and if the host vehicle also overlaps the trajectory, judges the candidate object as the preceding vehicle, And the host vehicle does not judge the candidate object as a preceding vehicle when the host vehicle does not overlap the corresponding trajectory.

According to one aspect of the present invention, the present invention stabilizes lane recognition in the presence of lane information, recognizes a preceding vehicle based on the location probability of the vehicle, and uses the trajectory of the preceding vehicles So that it is possible to recognize the preceding vehicle more stably and accurately.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is an exemplary diagram showing a schematic configuration of a preceding vehicle recognition apparatus according to an embodiment of the present invention; Fig.
2 is a flowchart illustrating a preceding vehicle recognition method according to an embodiment of the present invention.
3 is an exemplary view showing lane information and object position information calculated based on a host vehicle in FIG.
FIG. 4 is an exemplary diagram for explaining a method for checking an object in which a superimposed trajectory is generated, in FIG. 2; FIG.
FIG. 5 is an exemplary diagram showing a relationship between objects in which overlapping trajectories are generated, in the form of a table in FIG. 4; FIG.
FIG. 6 is an exemplary view for explaining a method of calculating a probability that a corresponding gem exists in each lane based on a distance between two lanes of a left / right and an object in FIG. 2; FIG.
FIG. 7 is an exemplary diagram for illustrating a method of displaying a recognized preceding vehicle on the HUD according to an embodiment of the present invention; FIG.

Hereinafter, an embodiment of a method and apparatus for recognizing a preceding vehicle according to the present invention will be described with reference to the accompanying drawings.

In this process, the thicknesses of the lines and the sizes of the components shown in the drawings may be exaggerated for clarity and convenience of explanation. In addition, the terms described below are defined in consideration of the functions of the present invention, which may vary depending on the intention or custom of the user, the operator. Therefore, definitions of these terms should be made based on the contents throughout this specification.

FIG. 1 is an exemplary diagram showing a schematic configuration of a preceding vehicle recognizing apparatus according to an embodiment of the present invention.

1, the preceding vehicle recognizing apparatus according to the present embodiment includes a camera unit 110, a sensor unit 120, a control unit 130, an information input unit 140, and an information output unit 150 .

The camera unit 110 is mounted on a front surface of the vehicle and photographs a forward image of the vehicle. Further, the camera unit 110 provides a lane shape (road shape) in front, information for determining the size and type of the front vehicle.

The control unit 130 detects a trajectory, a lane, and the like of an object (i.e., a preceding vehicle) using the image frame captured by the camera unit 110. [

The sensor unit 120 measures a distance between an object (that is, a front vehicle) running on the front side and a host vehicle using a radar sensor or an infrared sensor.

The information input unit 140 may be configured to control the attitude, speed, and heading angle of the vehicle using sensors (e.g., a vehicle speed sensor, a steering angle sensor, a gyro sensor, a yaw rate sensor, And the like.

Here, the yaw rate sensor is used to check whether the vehicle is turning or leaving the driving lane, and the gyro sensor is used to determine whether to enter the slope section or the curve section, and measures the slope.

The information input unit 140 is also connected to a navigation terminal (not shown) to receive information on the shape of the road (e.g., straight line, slope, curved line) Or map information) can be received.

The information output unit 150 displays the preceding vehicle information through a display unit of an AVN (Audio Video Navigation) unit or a Head Up Display (HUD).

The control unit 130 processes the forward image photographed through the camera unit 110 to calculate lane information and position information of the objects, estimates objects in the lane on which the host vehicle exists among the objects, Recognizes the object (i.e., the front object) closest to the host vehicle among the objects estimated to be in the lane with the vehicle as a preceding vehicle.

The controller 130 determines the lane, the object (i.e., the preceding vehicle) and the preceding vehicle that are recognized (or judged) and displays the information through the information output unit 150 (e.g., AVN or HUD) (See FIG. 7), or may output the preceding vehicle information to another system (e.g., a smart cruise control system, an autonomous driving system, and the like).

Hereinafter, a specific operation of the controller 130 will be described with reference to the flowchart of FIG.

2 is a flowchart illustrating a method of recognizing a preceding vehicle according to an embodiment of the present invention.

Referring to FIG. 2, the control unit 130 processes the image input through the camera unit 110 to calculate lane information and object (i.e., forward vehicle) position information.

Here, the lane information and object (i.e., forward vehicle) position information are information calculated based on the host vehicle (i.e., a vehicle).

The control unit 130 performs coordinate transformation to generate a trajectory of the object (i.e., the preceding vehicle).

The control unit 130 creates a relative coordinate system at the start position of the host vehicle and compensates the movement amount and heading information of the host vehicle as shown in FIG. Information is calculated (S101).

The calculated object (i.e., forward vehicle) position information and lane information may be displayed through the information output unit 150 (e.g., display means, HUD, etc.) as shown in FIG.

Here, the movement amount and heading information of the host vehicle can be calculated using the vehicle speed sensor information, the steering angle sensor information, and the yaw sensor information.

FIG. 3 is an exemplary diagram showing lane information and object information (ie, forward vehicle) position information calculated based on the host vehicle in FIG. 2; FIG.

As shown in FIG. 3, the control unit 130 may process an image input through the camera unit 110 in a predetermined manner to calculate lane information and object (i.e., forward vehicle) position information.

At this time, the controller 130 may designate an ID for distinguishing each object (that is, a forward vehicle), and may specify a color for each object (i.e., a forward vehicle) for display.

Here, an object (for example, ID 2) running ahead of the host vehicle in the same lane on which the host vehicle travels becomes the preceding vehicle among the objects (for example, ID 1 to ID 4).

Referring back to FIG. 2, the controller 130 removes noise from the lane using an exponential filter or a Kalman filter to stabilize lane information (or lane data) The lane is estimated (S102).

At this time, the controller 130 samples the lanes at predetermined predetermined intervals, as shown in FIG. 3, in order to reduce the capacity of the storage space (or memory space) and increase the efficiency of the calculation process.

As described above, the control unit 130 processes an image photographed through the camera unit 110 and allocates an ID to the recognized object (i.e., forward vehicles) and tracks the object (S103).

For example, the control unit 130 tracks objects between consecutive image frames on the basis of the distance and the trajectory direction, masks the continuity object as a 'tracking object', masks the newly generated object as a 'generated object' Each object is masked as a 'destroying object'.

Also, as shown in FIG. 3, the controller 130 generates a trajectory for each of the 'tracking objects' (S104).

Here, the trajectory represents the movement path of each object, and sampling is performed at predetermined predetermined intervals to reduce the capacity of the storage space (or memory space) required for generating the trajectory and increase the efficiency of the computation process.

Also, the controller 130 searches (or checks) whether the trajectories of the objects overlap each other (i.e., whether overlapping trajectories have been generated) (S105).

For example, the step of checking (or checking) whether the superimposed trajectory is generated (step S105) (i.e., determining whether the position of each object overlaps with the trajectory of another object) Finds two trajectory samples that are closest to the position of the object (i.e., the preceding vehicle) (e.g., ID 2), calculates the distance between the object and the line segment connecting the two trajectory samples and determines whether the distance is below a predetermined threshold value It can be judged whether or not it is a superimposed trajectory.

FIG. 4 is a diagram for explaining a method of checking an object in which a superimposed trajectory is generated in FIG. 2, and FIG. 5 is a diagram illustrating a relationship between objects in which superimposed trajectories are generated, Fig.

Referring to FIG. 4 and FIG. 5, it can be easily known what object (for example, ID 1 and ID 2) the generated overlapping locus is.

Further, the controller 130 estimates the distance between the left / right lane and each object (i.e., the object-lane distance) based on the host vehicle when the lane is recognized (or when the lane information is calculated) ).

For example, as shown in FIG. 6, the controller 130 finds two lane samples closest to the position of each object, calculates a distance between the line segment connecting the left and right two lane samples and the object Or estimated). Here, the calculation of the distance between the left and right lanes and the object is performed for each of the left and right lanes.

Also, the controller 130 calculates the probability that the object (i.e., the preceding vehicle) is in the lane when the lane is recognized (or the lane information is calculated) (S107).

For example, assuming that there are three lanes of a left lane, a middle lane, and a right lane on the basis of a host vehicle (that is, a child car), the control unit 130 calculates a probability that an object exists for each lane of the third lane. (Or probability distribution) at the position of each object based on the estimated object-to-lane distance as shown in FIG. 6, and calculates the probability calculated for each object at a specified time interval Accumulates the accumulated binning value by normalizing the binarized binarized binarization probability. The lane on which the object exists is estimated by virtue of the accumulated value.

FIG. 6 is an exemplary diagram for explaining a method of calculating a probability that a corresponding object exists in each lane based on a distance between two lanes of a left / right and an object, in FIG.

Also, the controller 130 estimates the preceding vehicle of the host vehicle based on the estimated information (e.g., lane, object trajectory, corresponding lane with the object, etc.) (S108).

For example, when the lane is recognized (or the lane information is detected), the controller 130 determines that the probability that the center probability is equal to or greater than a predetermined threshold value (e.g., a threshold value for the voting value) Recognizes an object as a preceding vehicle, and recognizes a nearby object as a preceding vehicle if there are two or more objects.

However, if the lane is not recognized (or the lane information is not detected) or if the preceding vehicle can not be recognized based on probability, the controller 130 selects the preceding vehicle candidate based on the trajectory.

For example, the controller 130 calculates distances to the respective trajectories at the host vehicle position and finds trajectories that fall below a preset threshold value. At this time, if there is one trajectory that falls below the threshold value, the object of the trajectory becomes the preceding vehicle candidate, and if there are two or more trajectories that fall below the threshold value (e.g., two or more vehicles in the same lane travel continuously back and forth The object that overlaps the trajectory of another object becomes the preceding vehicle candidate object. Next, the controller 130 checks whether the preceding vehicle candidate object overlaps the trajectory of another object. If the preceding vehicle candidate object of the two or more objects overlaps the trajectory of another object, the control unit 130 determines whether the host vehicle also overlaps the corresponding trajectory If the host vehicle does not overlap the locus, the host vehicle does not judge the candidate object as the preceding vehicle.

7, when the lane, the object (i.e., the forward vehicle) and the preceding vehicle are judged (recognized) as described above, the control unit 130 controls the information output unit 150 For example, AVN, HUD, etc.) to display a preceding vehicle (e.g., a visual indication such as a border to distinguish it from a preceding vehicle), or voice to perform a guidance operation on the preceding vehicle information.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, I will understand the point. Accordingly, the technical scope of the present invention should be defined by the following claims.

110:
120:
130:
140: Information input unit
150: Information output section

Claims (22)

The control unit processes the image input through the camera unit to calculate the lane information and the position information of the objects;
The control unit tracing the objects to generate a trajectory;
Checking whether a superimposed trajectory is generated in the objects;
Estimating a distance between the left / right lane and each object based on the host vehicle, and calculating a probability that the object is in a lane with the host vehicle, when the lane information is calculated; And
The control unit estimates a preceding vehicle of the host vehicle based on at least one or more previously designated information calculated or generated, and when the host vehicle is present among the position probabilities of the respective objects, the vehicle probability is equal to or greater than a predetermined threshold, And recognizing a nearby object as a preceding vehicle when the number of objects is two or more,
If the lane information is not detected or the preceding vehicle can not be recognized based on probability,
Wherein,
A candidate for a preceding vehicle candidate is selected based on a trajectory, and a check is made as to whether the candidate candidate for the preceding vehicle overlaps the trajectory of another object. If the candidate for the preceding vehicle overlaps the trajectory of another object, and,
Wherein the host vehicle also determines the candidate object as a preceding vehicle when the corresponding trajectory is superimposed, and does not determine the candidate object as a preceding vehicle when the host vehicle does not overlap the corresponding trajectory .
The method of claim 1, wherein the at least one pre-
Lane information, trajectory information of an object, and corresponding lane information with an object.
2. The method of claim 1, wherein the lane information and the location information of the objects include:
Wherein the information is calculated based on the host vehicle.
2. The method of claim 1, further comprising:
The control unit performs coordinate transformation to create a relative coordinate system at a start position of the host vehicle,
Wherein the control unit calculates object position information and lane information in the relative coordinate system by compensating movement amount and heading information of the host vehicle.
5. The method of claim 4, wherein the amount of movement and heading information of the host vehicle
The vehicle speed sensor information, the steering angle sensor information, and the yaw sensor information.
The method of claim 1, further comprising:
Wherein,
(ID) of each object can be specified, and the color of each object can be designated for display of the preceding vehicle.
The method according to claim 1, wherein, in order to stabilize the lane information,
The control unit removes noise in the lane using a predetermined filter,
Wherein the lanes are sampled at predetermined intervals in order to reduce the capacity of the storage space and to increase the efficiency of the calculation process.
The apparatus of claim 1,
Tracks the objects between frames of a continuous image inputted through the camera unit based on the distance and the trajectory direction, masks the continuity object as a 'tracking object', masks the newly generated object as a 'generated object' Wherein each of the objects is masked as a 'destruction object'.
2. The method according to claim 1, wherein, in the step of tracing each of the objects to generate a trajectory, the trajectory represents a movement path of the object,
Wherein,
Wherein the moving path of the object is sampled at predetermined intervals in order to reduce the capacity of the storage space required for creating the trajectory of the tracked object and to increase the efficiency of the calculation process.
2. The method of claim 1, wherein checking whether the overlapping trajectory is generated comprises:
The two trajectory samples that are closest to the position of any one object are searched and the distance between the line segment and the object connecting the two trajectory samples is calculated to determine whether or not the trajectory is a superimposed trajectory depending on whether the difference is less than a predetermined threshold value Of the preceding vehicle.
2. The method of claim 1, wherein, in calculating the probability,
Wherein,
And calculating a probability from a position of each object to a normal distribution based on the distance between the estimated object and the lane.
delete The method according to claim 1,
In order to select a preceding vehicle candidate based on the trajectory,
Wherein,
Calculating a distance between the trajectory of each object and a trajectory of each object at a host vehicle position to find trajectories that fall below a predetermined threshold value and selecting an object of the trajectory as a preceding vehicle candidate if the trajectory is less than the threshold value, Is selected as a preceding vehicle candidate object when the number of trajectories to be traced is two or more.
A camera unit for photographing a forward image of the host vehicle;
A sensor unit including a vehicle speed sensor, a steering angle sensor, and a yaw sensor for detecting movement amount and heading information of the host vehicle;
Estimating objects in the lane on which the host vehicle exists among the objects, estimating objects in the lane on which the host vehicle exists, processing the forward image to calculate lane information and position information of the objects, And a controller for recognizing the object closest to the vehicle as a preceding vehicle,
Wherein,
If the lane information is not detected or the preceding vehicle can not be recognized based on the probability, the leading vehicle candidate is selected based on the trajectory,
Checking whether the preceding vehicle candidate object overlaps the trajectory of another object, checking whether the host vehicle also overlaps the trajectory when the preceding vehicle candidate object overlaps the trajectory of another object,
Wherein the host vehicle also determines the candidate object as a preceding vehicle when the corresponding trajectory is overlapped and does not determine the candidate object as a preceding vehicle when the host vehicle does not overlap the corresponding trajectory. .
15. The apparatus of claim 14,
Estimating a distance between the left / right lane and each object based on the host vehicle, calculating a probability that the object is present in the host vehicle,
Recognizing the object as a preceding vehicle when the probability of the host vehicle is greater than a predetermined threshold, and recognizing a nearby object as a preceding vehicle when the number of objects is two or more. A preceding vehicle recognition device.
15. The method of claim 14,
As a vehicle running in front of the host vehicle,
Wherein the object to be driven immediately before the host vehicle is a preceding vehicle, and all objects other than the object corresponding to the preceding vehicle are forward vehicles.
A camera unit for photographing a forward image of the host vehicle;
A sensor unit including a vehicle speed sensor, a steering angle sensor, and a yaw sensor for detecting movement amount and heading information of the host vehicle;
The control unit may process the forward image to calculate lane information and position information of the objects, generate trajectories by tracing the objects, check whether overlapping trajectories are generated in the objects, Estimates the distance between the object and the object, calculates the probability that the object is present in the host vehicle, estimates the host vehicle's preceding vehicle based on the estimated lane information, object locus information, A controller for recognizing the object as a preceding vehicle if the probability of the host vehicle is greater than or equal to a predetermined threshold, and recognizing a nearby object as a preceding vehicle when there are two or more objects; Wherein the preceding vehicle recognizing device comprises:
18. The apparatus of claim 17,
In order to generate the trajectory of the object, a coordinate transformation is performed to create a relative coordinate system at the start position of the host vehicle,
And calculates the object position information and the lane information in the relative coordinate system by compensating the movement amount and the heading information of the host vehicle.
18. The apparatus of claim 17,
In order to stabilize the lane information, noise of the lane information is removed by using an exponential filter or a Kalman filter,
Wherein the lanes are sampled at predetermined intervals in order to reduce the capacity of the storage space and to increase the efficiency of the calculation process.
18. The apparatus of claim 17,
In order to check whether the superimposed trajectory is generated, two trajectory samples that are closest to the position of an object are searched, and the distance between the object and the line segment connecting the two trajectory samples is calculated to determine whether it is below a predetermined threshold And determining whether the vehicle is a superimposed trajectory according to the detected position.
18. The apparatus of claim 17,
In computing the probability that the object is in a lane with a host vehicle,
And calculates a probability from a position of each object to a normal distribution based on the distance between the estimated object and the lane.
delete

Images