KR20150124889A - Lane assignment method and system using multi-lane detection and lane change detection - Google Patents

Abstract

A lane recognition method using multi-lane detection and lane change detection comprises the following steps of: acquiring a front image from a vision sensor pre-installed in a vehicle which is being driven; applying a multi-lane detection method so as to estimate a lane of the vehicle which is being driven on the basis of the acquired front image; updating a score histogram on the basis of the estimated lane as a result of the application; and determining the lane of the vehicle which is being driven on the basis of the updated score histogram.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a multi-lane detection and a lane change detection based lane recognition method and system,

More particularly, the present invention relates to a method and system for recognizing a lane of a moving vehicle, and more particularly, to a method and system for recognizing a lane of a moving vehicle using a multi-lane detection technique and a lane change detection technique. The present invention relates to a technique for recognizing a user.

The existing technology for recognizing the lane of the moving vehicle uses the GPS system provided in the vehicle.

Therefore, there is a disadvantage that the technology of recognizing the lane of the existing traveling car is greatly influenced by the connection state between the GPS system and the satellite.

Therefore, the present specification proposes a technique of recognizing a lane of a running vehicle by using a vision sensor installed in advance in a running vehicle instead of using a GPS system.

A method and system for recognizing a lane according to an embodiment provides a technique of recognizing a lane of a running vehicle by using a vision sensor installed in advance in a running vehicle.

In particular, a method and system for recognizing a lane according to an embodiment provides a technique of determining a lane of a driving vehicle by applying a multi-lane detection technique based on a front image obtained from a vision sensor installed in advance in a running vehicle do.

In addition, a method and system for recognizing a lane according to an embodiment determines a lane of a driving vehicle as a wrong lane by using a score histogram updated based on the estimated lane in the process of applying the multi lane detection technique Provides a technique to lower probability.

In addition, a method and system for recognizing a lane according to an exemplary embodiment provides a technique for determining a lane of a running vehicle in consideration of changing a lane in a running vehicle by using a lane change detection technique.

According to an embodiment of the present invention, there is provided a multi-lane detection and lane change detection based lane recognition method, comprising: obtaining a front image from a vision sensor installed in a running vehicle; Applying a multi-lane detection technique to estimate a lane of the driving vehicle based on the obtained front image; Updating the score histogram based on the estimated lane as a result of the application; And determining a lane of the running vehicle based on the updated score histogram.

As a result of the application, updating the score histogram based on the estimated lane may include a lane change detection technique to determine whether the lane of the driving vehicle is changed based on the obtained front image Selectively applying; And initializing the score histogram when the lane of the driving vehicle is changed as a result of the determination.

The step of selectively applying the lane change detection technique may include calculating a departure value calculated using a vanishing point for the obtained front image, a center point of a vision sensor installed in advance in the running vehicle, and a threshold value associated with the estimated lane, value is changed in units of frames, it may be determined whether or not the lane of the moving vehicle is changed.

Wherein the applying the multi-lane detection technique comprises: transforming the obtained front image into a top-view image; Performing edge detection from the top view image; Performing a hough transform to detect a set of at least one candidate lane based on the image on which the edge detection is performed; And estimating a lane of the running vehicle based on the set of the at least one candidate lane.

The step of performing the Hough transform may further include calculating a position of the driving vehicle in the image subjected to the Hough transform, a width of the lane in the image in which the Hough transform is performed, and a number of lanes in the image in which the Hough transform is performed And detecting the set of at least one candidate lane from the image on which the Huff transform is performed.

Wherein estimating the lane of the moving vehicle based on the set of at least one candidate lane comprises: forming a matrix comprising one row for each set of the at least one candidate lane; Calculating and summing a difference value between a gray level histogram for each of at least one candidate lane area defined by the set of at least one candidate lane for each row of the matrix; Determining at least one candidate lane having the smallest value as a final lane for estimating a lane of the running vehicle; And estimating the vehicle as a lane of the running vehicle based on the determined final lane.

As a result of the application, updating the score histogram based on the estimated lane may include the appearance of a section corresponding to the estimated lane on the score histogram having each of the final lanes determined in the course of estimating the lane of the running vehicle Wherein the step of determining the lane of the running vehicle includes a step of calculating a final lane corresponding to a section having the highest frequency of occurrence of the updated score histogram as a lane of the running vehicle And a step of determining the number

Wherein the step of accumulating the number of occurrences of the section corresponding to the estimated lane on the score histogram includes a step of determining whether the number of occurrences is greater than a preset threshold value, ), And then initializing the multiplication result.

The multi-lane detection and lane change detection based lane recognition system according to an embodiment includes an acquiring unit for acquiring a forward image from a vision sensor previously installed in a running vehicle; A multi-lane detection technique applying a multi-lane detection technique to estimate a lane of the running vehicle based on the obtained front image; An update unit updating the score histogram based on the estimated lane as a result of the application; And a determination unit determining a lane of the running vehicle based on the updated score histogram.

Wherein the update unit applies a lane change detection technique selectively to a lane change detection technique to determine whether the lane of the driving vehicle is changed based on the obtained front image; And an initialization unit for initializing the score histogram when the lane of the driving vehicle is changed as a result of the determination.

A method and system for recognizing a lane according to an embodiment can provide a technique of recognizing a lane of a running vehicle by using a vision sensor installed in advance in a running vehicle.

In particular, a method and system for recognizing a lane according to an embodiment provides a technique of determining a lane of a driving vehicle by applying a multi-lane detection technique based on a front image obtained from a vision sensor installed in advance in a running vehicle can do.

In addition, the method and system for recognizing a lane according to an embodiment may determine a lane of a driving vehicle as a wrong lane by using a score histogram updated based on the estimated lane in the process of applying the multi lane detection technique It is possible to provide a technique of lowering the probability.

In addition, the method and system for recognizing a lane according to an embodiment can provide a technique of determining a lane of an on-going automobile in consideration of changing a lane on the road by using a lane change detection technique.

1 is a view for explaining a lane of a running vehicle according to an embodiment and expressing the position information of the running vehicle as a lane.
FIG. 2 is a diagram for explaining a process of applying a multi-track detection method according to an embodiment.
FIG. 3 is a diagram for explaining a process of estimating a lane of a running vehicle based on at least one set of candidate lanes described with reference to FIG. 2. FIG.
4 is a diagram for explaining a process of updating a score histogram based on an estimated lane according to an embodiment and determining a lane of a running vehicle.
FIG. 5 is a view for explaining a process of using a car change detection technique according to an embodiment.
6 is a flowchart showing a lane recognizing method according to an embodiment.
FIG. 7 is a flowchart specifically illustrating a step of applying the multi-lane detection technique shown in FIG.
8 is a block diagram illustrating a lane identification system in accordance with one embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. In addition, the same reference numerals shown in the drawings denote the same members.

Also, terminologies used herein are terms used to properly represent preferred embodiments of the present invention, which may vary depending on the user, intent of the operator, or custom in the field to which the present invention belongs. Therefore, the definitions of these terms should be based on the contents throughout this specification.

1 is a view for explaining a lane of a running vehicle according to an embodiment and expressing the position information of the running vehicle as a lane.

Referring to FIG. 1, the car recognition system according to an exemplary embodiment of the present invention is provided in advance in a car in motion, and recognizes the car of the car in motion by using a pre-installed vision sensor. have.

For example, in the case of (a), the lane recognition system recognizes that the vehicle 110 travels in the second lane by using a vision sensor installed in advance in the lane 110, The information can be expressed as a second lane.

Likewise, in the case of (b), the lane recognition system recognizes that the vehicle 120 travels in the third lane by using the vision sensor installed in advance in the running vehicle 120, It can be expressed as the third lane.

Hereinafter, the lane identification system can be utilized in advanced driver assistance systems (ADAS) or unmanned vehicle technology because it expresses the position information of the vehicle by the recognized lane by performing the lane recognition method described later.

FIG. 2 is a diagram for explaining a process of applying a multi-track detection method according to an embodiment.

Referring to FIG. 2, the lane recognition system according to one embodiment first obtains a front image 210 from a vision sensor installed in advance in a running vehicle. By applying the multi-lane detection technique based on the obtained front image 210, the lane of the vehicle under driving can be estimated. Hereinafter, a series of processes described with reference to FIGS. 2 and 3 means respective steps included in the multi-lane detection technique.

The car recognition system then converts the acquired front image 210 into a top-view image 220 and then performs edge detection from the top-view image 220. At this time, the lane recognizing system uses a Sobel operator to perform edge detection, but various filters can be used without limitation or limitation.

The lane recognition system may perform a hough transform to detect a set of at least one candidate lane based on the image 230 on which the edge detection has been performed. Here, the lane recognition system determines the position of the moving vehicle in the image 240 in which the Huff transform has been performed, the width of the lane in the image 240 in which the Hough transform has been performed, A set of at least one candidate lane may be detected from the image 240 in which the Hough transform is performed.

으로 설정한 후, (a) 경우와 같이, 좌측의 2개 차선들을

및

로 각각 설정함으로써, 제1 후보 차선의 세트로

,

및

을 검출할 수 있다. 이 때,

는 허프 변환이 수행된 이미지(240)에서의 차로의 폭을 의미한다. 또한, 차로 인식 시스템은 허프 변환이 수행된 이미지(240)에서 주행 중인 자동차의 위치를 참고하여, 자동차의 우측 선을 기준 선인

으로 설정한 후, (b) 경우와 같이, 좌우측의 1개씩 차선-총 2개의 차선들-을

및

로 각각 설정함으로써, 제2 후보 차선의 세트로,

,

및

을 검출할 수 있다.For example, the lane recognizing system may select adjacent lanes based on the location of the driving vehicle in the image 240 in which the Hough transform is performed. Hereinafter, two lanes are selected as adjacent lanes based on the right-hand line of the vehicle being driven, but the number of adjacent lanes selected and the right-hand line or the left-hand line of the vehicle under travel But are not limited to. The lane recognizing system refers to the position of the vehicle running in the image 240 in which the Hough transform is performed,

(A), the left two lanes

And

By setting each of the first candidate lane

,

And

Can be detected. At this time,

Quot; refers to the width of the lane in the image 240 in which the Hough transform is performed. In addition, the lane recognizing system refers to the position of the running vehicle in the image 240 in which the Hough transform is performed,

, And then, as in the case of (b), one lane on the left and one lane on the left and right - two lanes total

And

Respectively, to the second candidate lane set,

,

And

Can be detected.

The lane of the vehicle under running can be estimated based on the set of at least one candidate lane thus detected (the set of the first candidate lane and the set of the second candidate lane). A detailed description thereof will be described with reference to Fig.

FIG. 3 is a diagram for explaining a process of estimating a lane of a running vehicle based on at least one set of candidate lanes described with reference to FIG. 2. FIG.

3, the lane recognition system according to one embodiment is based on a set of at least one candidate lane (a set of first candidate lanes and a set of second candidate lanes) detected as described with reference to Fig. 2 , It is possible to estimate the lane of the moving vehicle.

Specifically, the lane-recognition system comprises a matrix constituting two rows with a set of first candidate lanes in the case (b) and a second set of candidates in the case (b), and a matrix of at least one candidate lane And one row per set.

,

및

로 구성하고, 2행을 제2 후보 차선의 세트인

,

및

로 구성할 수 있다.
For example, the lane recognizing system may be configured to calculate one row of the matrix P as a first set of candidate lanes

,

And

And the second row is constituted by a set of the second candidate lanes

,

And

.

&Quot; (1) "

Here, since lanes between lanes have the same shape and color, respectively, a gray level histogram for each lane can have similar properties. Thus, the lane recognition system may compare the gray-level histograms for each of the regions of at least one candidate lane defined by the set of at least one candidate lane of cases (a) and (b) The final lane can be determined. For example, the lane recognition system may include a first region 310, a second region 320, and a second region 320 that are regions of at least one candidate lane defined by a set of at least one candidate lane in case (a) and case (b) By comparing the gray level histograms 311, 321, and 331 with respect to each of the third areas 330, the second area 320 and the third area 330 similar to each other can be determined as the final lane.

The process of calculating such a final lane deciding process may include a step of calculating a difference value between gray level histograms for each of at least one candidate lane area in which the recognition system is defined as a set of at least one candidate lane for each row of the matrix And then summing the results of the summation and determining at least one candidate lane having the smallest value as the final lane.

및 제1 후보 차선의 세트의 두 번째 차선인 두 번째 행렬 요소와 제1 후보 차선의 세트의 세 번째 차선인 세 번째 행렬 요소로 정의되는 두 번째 차로에 대한 그레이 레벨 히스토그램

사이의 차이값

을 수학식 2와 같이 계산할 수 있다.
Specifically, the lane recognition system may calculate the difference value between the gray-level histograms for each of the regions of at least one candidate lane defined by the set of at least one candidate lane, for each row of the matrix. For example, the lane recognition system may determine, for one row of the matrix, the first matrix element that is the first lane of the first set of candidate lanes and the first matrix element that is the second lane of the second lane of the set of first candidate lanes Gray level histogram for the lth lane

And a gray level histogram for a second lane defined by a second matrix element of a second set of first candidate lanes and a third matrix element of a third set of first candidate lanes

Difference value between

Can be calculated as shown in Equation (2).

&Quot; (2) "

이 획득될 수 있다.
When the difference value between gray level histograms for each of at least one candidate lane defined by a set of at least one candidate lane for each row of the matrix is calculated, it is summed as shown in Equation (3)

Can be obtained.

&Quot; (3) "

는 매트릭스의 i행을 구성하는 적어도 하나의 후보 차선의 세트인 행렬 요소들로 정의되는 i행의 후보 차로의 영역 각각의 그레이 레벨 히스토그램의 세트를 의미하고, N은 후보 차로의 영역의 개수를 의미하며,

는 i행의 j번째 후보 차로의 영역에 대한 그레이 레벨 히스토그램을 의미한다.here,

Means a set of gray-level histograms of each of the candidate lane areas of i rows defined by matrix elements that are a set of at least one candidate lane constituting i rows of the matrix, and N denotes the number of regions in the candidate lanes In addition,

Denotes a gray-level histogram for the area of the jth candidate lane of the i-th row.

가 가장 작은 행을 구성하는 적어도 하나의 차선 후보 세트로 정의되는 적어도 하나의 후보 차로를 수학식 4와 같이 최종 차로

로 결정할 수 있다.
Thus, the lane identification system determines at least one candidate lane having the smallest value as a result of summing out at least one candidate lane defined by at least one candidate lane set for each row of the matrix, The lane can be estimated. For example, the lane recognition system may calculate a difference value between gray level histograms for each of at least one candidate lane defined as a set of at least one candidate lane for each row of the matrix,

At least one candidate lane defined by at least one lane candidate set constituting the smallest row is obtained as a final lane

.

&Quot; (4) "

,

및

로 정의되는 제2 영역(320) 및 제3 영역(330)이 합산 결과 가장 작은 값을 갖기 때문에, 최종 차로를 (b) 경우와 같이, 제2 영역(320) 및 제3 영역(330)으로 결정할 수 있다.Considering the cases (a) and (b) described above, the lane recognition system is a set of second candidate lanes constituting two rows of the matrix P

,

And

The second area 320 and the third area 330 defined by the first area 320 and the third area 330 have the smallest summed result so that the final area is divided into the second area 320 and the third area 330 You can decide.

As a result, the lane recognizing system can estimate the lane of the vehicle under driving based on the determined final lane and the position of the lane marker in the image in which the Hough transform is performed. For example, the car recognition system can determine that the road of the vehicle underway is a second-degree road, as described above, and that the driving of the image in the Hough transform based on the second- It can be estimated that the car of the moving vehicle is the primary car.

4 is a diagram for explaining a process of updating a score histogram based on an estimated lane according to an embodiment and determining a lane of a running vehicle.

Referring to FIG. 4, the lane identification system according to an embodiment of the present invention estimates the lane difference of the driving lane of the moving vehicle, which is estimated on the score histogram 410, The score histogram 410 can be updated based on the estimated lane by accumulating the number of appearances of the section corresponding to the lane as shown in Equation (5).

Equation (5)

는 추정된 차로를 의미하고,

는 추정된 차로에 해당하는 구간을 의미한다.here,

Is an estimated lane,

Means the section corresponding to the estimated lane.

For example, if it is determined that the last lane is a second lane and the lane of the first lane is the first lane, then the lane recognition system may include a first lane 420 and a second lane 430, The number of occurrences can be accumulated in the first interval 420 on the score histogram 410 having the number of occurrences.

Therefore, the lane recognizing system can determine the last lane corresponding to the section having the highest appearance frequency of the section of the updated score histogram 410 as the lane of the running vehicle.

For example, the process of determining the final lane described above and estimating the lane of the moving vehicle is performed in a unit of a predetermined period or frame, so that the highest occurrence of the first period 420 and the second lane period 430 In the case where the interval having the frequency is the first interval 420, the lane recognition system can determine the first lane corresponding to the first interval 420 as the lane of the running vehicle.

Thus, by using the score histogram 410 that is updated based on the estimated lane, the lane recognition system can lower the probability of determining the lane of the lane on which the lane is running as a wrong lane.

Also, in order to prevent accumulation of values only in a specific section in the score histogram, the lane recognition system may be configured such that when the accumulated interval exceeds a preset threshold value, It can be initialized by multiplying by a forgetting factor.

&Quot; (6) "

는 미리 설정된 임계값을 초과하는 출현도수가 축적된 구간을 의미하고,

는 망각 인자를 의미하며, N은 최종 차로의 개수를 의미한다.
here,

Means a period in which the number of occurrences exceeding a predetermined threshold value is accumulated,

Denotes the forgetting factor, and N denotes the number of the final lane.

FIG. 5 is a view for explaining a process of using a car change detection technique according to an embodiment.

Referring to FIG. 5, the lane recognition system according to an embodiment can determine a lane of a running vehicle by considering a lane change by using a lane change detection technique.

In particular, in the course of updating the score histogram based on the estimated lane described with reference to Fig. 4, the lane recognizing system determines whether the lane of the driving vehicle is changed based on the obtained front image, Technique can be selectively applied.

For example, the lane recognition system changes the departure value calculated by using the vanishing point of the obtained front image, the center point of the vision sensor pre-installed in the running vehicle, and the threshold value associated with the estimated lane, It is possible to judge whether or not the lane of the moving vehicle is changed.

를 먼저 계산할 수 있다.
More specifically, for example, the lane recognition system calculates the deviation value

Can be calculated first.

&Quot; (7) "

는 1이면 주행 중인 자동차가 우측 차로로 이동한 것을 의미하고, 2이면 좌측 차로로 이동한 것을 의미하며, 0이면 차로가 변경되지 않은 것을 의미한다. 또한,

는 획득된 전방 이미지에 대한 소실점을 의미하고,

는 주행 중인 자동차에 미리 설치된 비전 센서의 중심점을 의미하며,

는 추정된 차로와 관련된 임계값을 의미한다.Here,

1 means that the vehicle under travel has moved to the right lane, 2 means that the vehicle has moved to the left lane, and 0 means that the lane has not been changed. Also,

Represents the vanishing point of the acquired front image,

Means the center point of the vision sensor pre-installed in the running vehicle,

Means a threshold associated with the estimated lane.

By using the departure value thus calculated, the lane recognition system can determine whether the lane of the driving vehicle is changed by determining whether the departure value is changed frame by frame as in Equation (8).

&Quot; (8) "

는 차로 변경 이벤트 값을 의미하고,

는 i번째 프레임에서 이탈값을 의미한다. 따라서, 차로 변경 이벤트 값

가 1이면 주행 중인 자동차가 우측 차로로 이동한 것을 의미하고, 2이면 좌측 차로로 이동한 것을 의미하며, 0이면 차로가 변경되지 않은 것을 의미한다. 즉, 제1 프레임에서의 이탈값이 1이고, 제2 프레임(제1 프레임 이후의 프레임)에서의 이탈값이 2인 경우, 주행 중인 자동차는 우측 차로로 이동한 것을 의미하고, 제1 프레임에서의 이탈값이 2이고, 제2 프레임에서의 이탈값이 1인 경우, 주행 중인 자동차는 좌측 차로로 이동한 것을 의미하며, 제1 프레임에서의 이탈값 및 제2 프레임에서의 이탈값의 변화가 없는 경우, 차로가 변경되지 않은 것을 의미한다.here,

Lt; / RTI >< RTI ID = 0.0 >

Denotes an escape value in the i-th frame. Therefore,

Is 1 means that the vehicle being driven has moved to the right lane, 2 means that the vehicle has moved to the left lane, and 0 means that the lane has not been changed. That is, when the departure value in the first frame is 1 and the departure value in the second frame (frame after the first frame) is 2, it means that the vehicle under travel has moved to the right lane, When the departure value of the first frame is 2 and the departure value of the second frame is 1, it means that the vehicle under travel has moved to the left lane, and the change of the departure value in the first frame and the departure value in the second frame If not, it means that the lane has not been changed.

For example, if the vehicle 510 has a departure value of 1 in the first frame and a departure value of 2 in the second frame, as in the case of (a), the car recognizing system moves to the right lane If the departure value is 2 in the first frame and the departure value is 1 in the first frame as in the case (b), it can be judged that the vehicle under running has moved to the left lane have.

As a result of the determination, if the lane of the moving vehicle is changed, the lane recognition system can initialize the histogram. Thus, the lane recognizing system can determine the lane of the driving vehicle based on the score histogram updated after initialization.

6 is a flowchart showing a lane recognizing method according to an embodiment.

Referring to FIG. 6, a lane recognition method according to an embodiment is performed by a lane recognition system.

First, the lane recognition system obtains a front image from a vision sensor installed in a running vehicle (610).

Next, the car recognition system applies a multi-lane detection method (620) in order to estimate the lane of the moving vehicle based on the obtained front image. A detailed description thereof will be described with reference to FIG.

The car recognition system then updates (630) the score histogram based on the estimated lanes as a result of application. For example, the lane recognizing system can update the score histogram by accumulating the number of occurrences of the section corresponding to the lane estimated on the score histogram having each of the final lanes determined in the process of estimating the lane of the moving vehicle have. In this case, when the lane recognition system accumulates the number of occurrences of the section corresponding to the lane estimated on the score histogram, if the accumulated section exceeds the predetermined threshold value, the number of occurrences of the section is forgotten It can also be initialized by multiplying by a forgetting factor.

Although not shown in a separate drawing, the lane recognition system determines whether or not the lane of the driving vehicle is changed based on the acquired front image in the step 630 of updating the score histogram based on the estimated lane as a result of the application The lane change detection technique is selectively applied to the vehicle, and if the lane of the driving vehicle is changed as a result of the determination, the score histogram can be initialized. At this time, the lane recognition system determines that the departure value calculated by using the vanishing point of the obtained front image, the center point of the vision sensor installed in the running vehicle, and the threshold value associated with the estimated lane is changed in units of frames As a basis, it is possible to selectively apply the change detection technique to the car by determining whether the lane of the car in motion is changed.

The car recognition system then determines 640 the lane of the on-going vehicle based on the updated score histogram. For example, the lane recognizing system can determine the lane of the on-going vehicle by determining the final lane corresponding to the section having the highest number of occurrences of the section of the updated score histogram as the lane of the running vehicle.

FIG. 7 is a flowchart specifically illustrating a step of applying the multi-lane detection technique shown in FIG.

Referring to FIG. 7, the lane recognition system according to the embodiment can apply the multi-lane detection method through the following process.

First, the car recognition system can convert the acquired front image into a top-view image (710).

The car recognition system may then perform edge detection from the top view image (720).

The car recognition system may then perform a hough transform (730) to detect a set of at least one candidate lane based on the image on which the edge detection is performed. At this time, the lane recognizing system performs a Hough transform based on the position of the running vehicle in the image in which the Hough transform is performed, the width of the lane in the image in which the Hough transform is performed, and the number of lanes in the image in which the Hough transform is performed. And detect at least one set of candidate lanes from the performed image.

The car recognition system can then estimate the lane of the moving vehicle based on the set of at least one candidate lane. Specifically, the lane recognition system may form 740 a matrix of one row for each set of at least one candidate lane. The lane recognition system may also compute and add the difference value between the gray level histograms for each of the at least one candidate lane area defined as a set of at least one candidate lane for each row of the matrix 750). Accordingly, the lane recognition system determines at least one candidate lane having the smallest value as the sum result as the final lane for estimating the lane of the moving vehicle (760), and then, based on the determined final lane, (770).

8 is a block diagram illustrating a lane identification system in accordance with one embodiment.

8, a lane recognition system according to an exemplary embodiment includes an acquisition unit 810, a multi-lane detection technique application unit 820, an update unit 830, and a determination unit 840. Referring to FIG.

The acquiring unit 810 acquires the front image from the vision sensor installed in advance in the running vehicle.

The multi-lane detection technique applying unit 820 applies a multi-lane detection technique to estimate the lane of the running vehicle based on the obtained front image.

Specifically, the multi-lane detection technique applying unit 820 may perform edge detection from the top view image after converting the acquired front image into a top-view image.

In addition, the multi-lane detection technique applying section 820 may perform a hough transform to detect a set of at least one candidate lane based on the image on which the edge detection is performed. At this time, the multi-lane road detection applying unit 820 determines the position of the moving vehicle in the image subjected to the Hough transform, the width of the lane in the image subjected to the Hough transform, the number of lanes in the image in which the Hough transform is performed , It is possible to detect a set of at least one candidate lane from the image in which the Hough transform is performed.

Accordingly, the multi-lane detection technique applying section 820 can estimate the lane of the running vehicle based on the set of at least one candidate lane. For example, the multi-lane detection scheme applying unit 820 may form a matrix constituting one row for each set of at least one candidate lane, and then form at least one candidate lane for each row of the matrix, The difference value between the gray level histograms for each of the regions of the candidate lanes of the pixel of interest can be calculated and added. The multi-lane detection technique applying unit 820 determines at least one candidate lane having the smallest value as the sum result as the final lane for estimating the lane of the on-road vehicle, and then, based on the determined final lane, Can be estimated.

The update unit 830 updates the score histogram based on the estimated lane as a result of the application. For example, the update unit 830 updates the score histogram by accumulating the number of occurrences of the section corresponding to the lane estimated on the score histogram having each of the final lanes determined in the process of estimating the lane of the running vehicle can do. At this time, in the process of accumulating the number of occurrences of the section corresponding to the lane estimated on the score histogram, if the interval in which the number of occurrences is accumulated exceeds a predetermined threshold value, the update section 830 updates the number of occurrences May be initialized by multiplying by a forgetting factor.

Also, although not shown in the drawing, the update unit 830 may determine whether the lane of the driving vehicle is changed based on the obtained front image in the process of updating the score histogram based on the estimated lane as a result of the application A change detection technique applying section for selectively applying a lane change detection technique to the vehicle, and an initialization section for initializing a score histogram when the lane of the running vehicle is changed as a result of the determination. In this case, the application of the change detection technique to the vehicle may include a departure value calculated using a vanishing point of the obtained front image, a center point of a vision sensor installed in a running vehicle in advance, and a threshold value associated with the estimated lane, It is possible to selectively apply the change detection technique to the car by determining whether or not the lane of the vehicle under operation is changed based on the change.

The determination unit 840 determines the lane of the vehicle under driving based on the updated score histogram. For example, the determining unit 840 can determine the lane of the vehicle under driving, by determining the final lane corresponding to the section having the highest frequency of occurrence of the updated score histogram as the lane of the running vehicle.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA) A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

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 exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (10)

Obtaining a front image from a vision sensor installed in advance in a running vehicle;
Applying a multi-lane detection technique to estimate a lane of the driving vehicle based on the obtained front image;
Updating the score histogram based on the estimated lane as a result of the application; And
Determining a lane of the running vehicle based on the updated score histogram
And a lane change detection based lane recognition method. The method according to claim 1,
As a result of the application, updating the score histogram based on the estimated lane
Selectively applying a lane change detection technique to determine whether the lane of the driving vehicle is changed based on the obtained front image; And
As a result of the determination, if the lane of the driving vehicle is changed, initializing the score histogram
Further comprising: a multi-lane detection and lane change detection based lane recognition method. 3. The method of claim 2,
The step of selectively applying the lane change detection technique
A departure value calculated using a vanishing point of the obtained front image, a center point of a vision sensor pre-installed in the running vehicle, and a threshold value associated with the estimated lane is changed in units of frames, A step of judging whether or not the lane of the moving vehicle is changed
And a lane change detection based lane recognition method. The method according to claim 1,
The step of applying the multi-track detection scheme
Converting the acquired front image into a top-view image;
Performing edge detection from the top view image;
Performing a hough transform to detect a set of at least one candidate lane based on the image on which the edge detection is performed; And
Estimating a lane of the running vehicle based on the set of at least one candidate lane
And a lane change detection based lane recognition method. 5. The method of claim 4,
The step of performing the Hough transform
Based on the position of the running vehicle in the image subjected to the Hough transform, the width of the lane in the image in which the Hough transform is performed, and the number of lanes in the image in which the Hough transform is performed, Detecting a set of at least one candidate lane
And a lane change detection based lane recognition method. 5. The method of claim 4,
Wherein estimating the lane of the moving vehicle based on the set of at least one candidate lane
Forming a matrix of one row for each of the at least one candidate lane;
Calculating and summing a difference value between a gray level histogram for each of at least one candidate lane area defined by the set of at least one candidate lane for each row of the matrix;
Determining at least one candidate lane having the smallest value as a final lane for estimating a lane of the running vehicle; And
A step of estimating a driving lane of the running vehicle based on the determined final lane
And a lane change detection based lane recognition method. The method according to claim 1,
As a result of the application, updating the score histogram based on the estimated lane
Accumulating the number of occurrences of the section corresponding to the estimated lane on the score histogram having each of the final lanes determined in the course of estimating the lane of the running vehicle,
Lt; / RTI >
The step of determining the lane of the oncoming vehicle
Determining a final lane corresponding to a section having the highest frequency of occurrence of the section of the updated score histogram as a lane of the running vehicle
And a lane change detection based lane recognition method. 8. The method of claim 7,
Wherein the step of accumulating the number of occurrences of the section corresponding to the estimated lane on the score histogram
If the interval in which the number of appearance frequencies is accumulated exceeds a predetermined threshold value, initializing the frequency by multiplying the number of occurrences of the interval by a forgetting factor
Further comprising: a multi-lane detection and lane change detection based lane recognition method. An acquiring unit for acquiring a forward image from a vision sensor installed in advance in a running vehicle;
A multi-lane detection technique applying a multi-lane detection technique to estimate a lane of the running vehicle based on the obtained front image;
An update unit updating the score histogram based on the estimated lane as a result of the application; And
Determining a lane of the running vehicle based on the updated score histogram,
And a lane change detection based lane recognition system. 10. The method of claim 9,
The update unit
A lane change detection technique applying unit to selectively apply a lane change detection technique to determine whether the lane of the driving vehicle is changed based on the obtained front image; And
As a result of the determination, when the lane of the running vehicle is changed, an initialization unit for initializing the score histogram
Further comprising: a multi-lane detection and lane change detection based lane recognition system.

Images