KR101877001B1 - Virtual line generation method, system and method for detecting reverse driving of vehicle - Google Patents

Abstract

A reverse-run detection system includes a camera for capturing a road image, a parameter setting module for setting an inverse direction in the road image based on a user input, a virtual line generation module for generating a plurality of virtual lines crossing the in- Wherein the traversing object is selected as a traversal candidate when the traversing path of the tracing object intersects with any one of the virtual lines, and when the moving direction and the moving distance of the traversing candidate satisfy a backward running condition And an inverse judgment module for judging that the inverse running situation is present.

Description

TECHNICAL FIELD [0001] The present invention relates to a virtual line generation method, a virtual line generation method,

Embodiments relate to a virtual line generation method, a backpropagation detection system, and a method thereof.

Due to the recent rapid increase in the number of vehicles, accidents on the roads are also increasing. Particularly, reverse driving of a vehicle on the road is one of the main factors causing a large accident.

Therefore, in order to prevent a large accident caused by reverse running of the vehicle, research is being conducted on a technique of warning the driver in advance of the backward running of the vehicle and warn the driver.

However, the backward movement sensing technologies studied so far have problems in that errors or incomplete performance are difficult to achieve when applied to actual road environments due to the characteristics of high noise and complicated road environments.

An object of the present invention is to provide a reverse traverse detection system and a method thereof capable of stably detecting reverse travel even in a complicated road environment.

According to another aspect of the present invention, there is provided an inverse movement detection system including a camera for capturing a road image, a parameter setting module for setting an inverse direction in the road image based on a user input, a plurality of virtual lines crossing the in- An object detection module for detecting a tracking object from the road image, and a tracking object selection unit for selecting the tracking object as a backward candidate if the movement path of the tracking object intersects with one of the virtual lines, And an inverse-movement determining module that determines that the vehicle is in an inverse-running state if the moving direction and the moving distance satisfy the inverse running condition.

According to another aspect of the present invention, there is provided a method of detecting a trajectory of an inverse running detection system, comprising: acquiring a road image through a camera; setting an inverse direction in the road image based on a user input; The method includes the steps of generating a virtual line, detecting a tracking object from the road image, selecting the tracking object as an inversion candidate when the movement path of the tracking object intersects with one of the virtual lines, And if the direction and the moving distance satisfy the backward running condition, determining that the backward traveling situation is determined.

According to another aspect of the present invention, there is provided a method for creating a virtual line of an object detection system, comprising: acquiring an image through a camera; setting an object moving direction in the image based on a user input; Generating a plurality of virtual lines that intersect the object moving direction within the object detection area, wherein the plurality of virtual lines detect a moving path of the object in the image, Lt; / RTI >

According to the embodiment, it is possible to prevent large accidents due to reverse travel by detecting the reverse travel of the vehicle stably even in a complicated road environment.

Fig. 1 schematically shows a reverse-run detection system according to an embodiment.
FIG. 2 and FIG. 3 illustrate examples of setting the inverse detection area in the reverse tracking detection system according to the embodiment.
FIG. 4 shows an example of generating virtual lines in the reverse movement detection system according to the embodiment.
FIG. 5 illustrates an example of generating a vertical virtual line in the reverse-running detection system according to the embodiment.
FIG. 6 is a diagram for explaining a method of selecting an inverse candidate in the inverse movement detection system according to the embodiment.
FIG. 7 is a diagram for explaining a method for determining whether or not a reverse candidate is reversed in the reverse-running detection system according to the embodiment.
8 schematically shows a method of detecting the backward movement according to the embodiment.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art to which the present invention pertains. The present invention may be embodied in many different forms and is not limited to the embodiments described herein.

In order to clearly illustrate the embodiments of the present invention, portions that are not related to the description are omitted, and the same or similar components are denoted by the same reference numerals throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between .

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a system and a method for detecting a reverse movement according to an embodiment will be described in detail with reference to the drawings.

Fig. 1 schematically shows a reverse-run detection system according to an embodiment. FIG. 2 and FIG. 3 illustrate examples of setting an inverse detection area in the reverse movement detection system according to the embodiment, and illustrate examples of setting an inverse travel detection area in a one-way road in a tunnel.

Referring to FIG. 1, the reverse movement detecting system 100 according to the embodiment includes a camera 110, a display 120, a user input unit 130, a memory 140, an output unit 150, a controller 160, . The components shown in FIG. 1 are not essential, so that the reverse tracking detection system 100 according to the embodiment may be implemented to include more or fewer components.

The camera 110 is disposed around the road so as to obtain a road image of the road.

Display 120 may display information processed in inverse tracking system 100. For example, the display 120 may display a road image photographed by the camera 110.

The user input unit 130 may include an input device for receiving a user input from a user. The user input unit 130 may include various types of input devices such as a touch pad, a touch sensor, a keypad, a button, a jog switch, and a mouse. When the user input unit 130 includes a touch sensor that senses a touch operation, the touch sensor may receive a touch input to the display 120 in a mutual layer structure with the display 120.

The memory 140 may store a program for the operation of the controller 160. [ The memory 140 may temporarily store data (e.g., road images) processed by the reverse-running detection system 100.

The output unit 150 may include a speaker, a lamp, a light emitting diode (LED), and the like. The output unit 150 may output a reverse warning sound or warn the driver that the vehicle is running in reverse through blinking of a lamp, a light emitting diode, or the like.

The output unit 150 records the corresponding reverse run time in the memory 150 or generates a control signal for requesting the image storage at the corresponding time and outputs the control signal to the external system capable of image recording , A mobile phone, a network video recorder (NVR), a digital video recorder (DVR), a video management system (VMS), a video control system, or the like.

The control unit 160 controls the overall operation of the reverse-running detection system 100. [

The control unit 160 may include a camera calibration module 161, a parameter setting module 162, a virtual line creation module 163, an object tracking module 164, an inversion determination module 165, and the like.

The camera calibration module 161 may generate a distance map including actual distance information based on the camera 110 for each pixel of the road image photographed by the camera 110 through camera calibration.

The camera calibration module 161 can estimate the conversion relationship between the coordinate system (world coordinate system) of the three-dimensional space corresponding to the actual road environment and the coordinate system (image coordinate system) of the two-dimensional road image through camera calibration. That is, each pixel in the road image is mapped to each point in the three-dimensional road environment, and the conversion relation between the world coordinate system corresponding to the actual road environment and the image coordinate system corresponding to the road image can be estimated based on the mapping result.

The camera calibration module 161 can acquire the distance information between the camera 110 and a point corresponding to each pixel in an actual road environment based on the conversion relationship between the coordinate systems. A distance map for pixels in the road image can be generated based on the distance information obtained for each pixel.

The parameter setting module 162 can set the reverse direction based on the user input. The parameter setting module 162 may analyze the touch input to the display 120 to set the reverse direction. The parameter setting module 162 may analyze the user's mouse drag input to set the reverse direction.

2, the parameter setting module 162 can display the graphic object 21 representing the direction of the arrow or the like in an inverse direction on the road image through the display 120 have. The user who has confirmed this can check the reverse direction set by himself / herself through the display 120, and can correct or reverse the reverse direction if necessary.

The parameter setting module 162 can set the range of inversion of travel based on the user input. The range of inversion is a parameter for determining which angle is to be regarded as an inverse of the backward direction. The parameter setting module 162 may analyze the user's touch input to the display 120 to set the range of inversion. The parameter setting module 162 may set the range of inverse travel by analyzing the drag input of the mouse or the like. The parameter setting module 162 may set an inverse range in response to the input of the inverse range through the user input unit 130. [

The parameter setting module 162 can display the graphic objects 22 and 23 indicating the range of inversion of the arrows and the like on the road image through the display 120 when the inverse range is set as shown in FIG. . The user who has confirmed this can check the range of the backward running set by himself / herself through the display 120, and can set or reset the backward running range if necessary.

The parameter setting module 162 may set an inverse running detection area in the road image based on the user input.

When a plurality of points (or pixels) in the road image are selected through the user input unit 130, the parameter setting module 162 can set a virtual area having the selected point (or pixel) as the vertex as the inverse running detection area.

2, the parameter setting module 162 connects the four points to each other as the four points P21, P22, P23, and P24 in the road image 5 are selected through the user input unit 130 And generates imaginary segments S21, S22, S23, and S24. Thus, a virtual reverse-direction detection area 20 having the virtual line segments S21, S22, S23, and S24 as an outline is generated.

The parameter setting module 162 may automatically generate a virtual reverse movement detection area based on the backward direction set by the user.

Taking FIG. 3A as an example, the parameter setting module 162 generates a virtual line segment S35 extending along the reverse direction. The virtual line segments S31 and S33 that are orthogonal to the imaginary line segment S35 extending in the reverse direction and the virtual line segments S32 and S33 that connect the both ends of the two virtual line segments S31 and S33 to each other are connected to each other. S34). Accordingly, the inverse-trend sensing area 30 having the virtual line segments S31, S32, S33, and S34 as an outline is automatically generated.

As shown in FIG. 3, the parameter setting module 162 may modify the reverse-running detection area based on the user input when the reverse-running detection area is automatically generated.

3B, the parameter setting module 162 receives a user input for modifying the virtual segments S31, S32, S33, and S34 through the user input unit 130, (S31, S32, S33, S34). As the virtual line segments S31, S32, S33 and S34 are changed, the inverse movement detection area 30 having the virtual line segments S31, S32, S33 and S34 as an outline is updated. Meanwhile, in the process of modifying the virtual line segments S31, S32, S33, and S34, the two virtual line segments S31 and S33 generated so as to be orthogonal to the inverse direction may be limited to be orthogonal to the inverse direction always .

The parameter setting module 162 may display the set backward movement sensing area over the road image when the backward movement sensing area is set.

2, the parameter setting module 162 draws a line representing the outlines S31, S32, S33, and S34 of the inverse travel sensing area 20 on the road image 5, The reverse direction detection area 20 can be displayed.

The parameter setting module 162 stores the area information of the reverse-running detection area in the memory 140 when the reverse-direction detection area is set.

The area information of the backward movement detection area may include coordinate information of each vertex constituting the backward movement detection area. 2, the area information of the inverse-traveling detection area may include coordinate information of the points P21, P22, P23, and P24 selected as the vertices of the inverse-

The area information of the reverse movement detection area may include coordinate information of the virtual lines corresponding to the outline of the reverse movement detection area. The coordinate information of the virtual line segment may include coordinate information of points corresponding to both ends of the virtual line segment. 3, the area information of the inverse traveling detection area 30 may include coordinate information of points corresponding to both ends of the virtual line segments S31, S32, S33, and S34 forming the outline of the inverse travel sensing area 30 .

The parameter setting module 162 may store the coordinate information of each vertex constituting the inverse movement detecting area in accordance with the sequence order when storing the area information of the inverse running detection area.

The virtual line generation module 163 uses line segments constituting an inverse running detection area to generate reference line segments or virtual lines. Therefore, the parameter setting module 162 determines whether or not the virtual line generation module 163 has received the area information of the sensing area stored in the memory 140 so that the coordinate information of each line constituting the inverse- The coordinate information of each vertex can be sorted and stored. 2, for example, the parameter setting module 162 may start clockwise or counterclockwise starting from any one of the line segments S21, S22, S23, and S24 (for example, line segment S21) The coordinates of the vertexes may be aligned and stored in the memory 140 so that the vertex coordinates of the line segments are aligned in the counterclockwise order.

The parameter setting module 162 can set parameters necessary for detecting the backward running. Parameters required for detecting the backward movement may include the number of virtual lines, the use of the vertical virtual line, the position of the vertical virtual line, the consistency, the additional detection condition, and the like.

The virtual line and the vertical virtual line are virtual lines that serve as a criterion for determining whether to reverse the tracking object.

The continuity is a parameter indicating how long the tracing object must satisfy the traversing condition to determine that the traversing situation has occurred. The continuity can be set by time, counter value, number of image frames, and the like.

For example, if the continuity is set to time, the tracing object can be determined to be in reverse if it satisfies the back-running conditions for a predetermined time. Also, for example, if continuity is set to a counter value, it can be determined that the tracking object is in a reverse direction if it satisfies the back-running conditions during a predetermined counter value. Also, for example, if the continuity is set to the number of image frames, it can be determined that the tracking object is in reverse when the tracking object satisfies the backward running conditions during a predetermined number of image frames.

The additional detection condition is an additional condition for determining the tracking object as an inverse object, and it is determined whether or not the tracking object moves away from (or approaching) the camera 110, a match degree between the moving direction of the tracking object and the inverse direction . ≪ / RTI >

The parameter setting module 162 can set each parameter necessary for detecting the inverse movement based on the user input.

The parameter setting module 162 may select the sensitivity based on the user input and automatically set the parameters necessary for detecting the inverse movement according to the selected sensitivity. In this case, the setting values of the parameters necessary for detecting the backward movement are stored differently in the memory 140 according to the sensitivity. The parameter setting module 162 reads the parameter setting values corresponding to the selected sensitivity from the memory 140 and automatically sets the parameters based on the sensitivity when the sensitivity is selected by the user input.

Table 1 below shows an example in which the set values of the parameters necessary for detecting the backward movement are set differently according to the sensitivity.

Table 1. Example of parameter setting according to sensitivity

In Table 1, the number of virtual lines is set higher as the sensitivity increases, and the use of vertical virtual lines depends on the sensitivity. Also, the continuity is set to be small as the sensitivity increases.

The virtual line generation module 163 can generate virtual lines for detecting the traversal of the tracking object based on the parameters set by the parameter setting module 162. [

The virtual line generation module 163 can generate a plurality of virtual lines that intersect the inverse direction within the reverse movement detection area.

The virtual line creation module 163 creates a reference line segment in the inverse movement detection area in order to create virtual lines in the inverse movement detection area. The reference line segment is a line segment connecting two line segments that intersect with the backward direction among the virtual line segments constituting the reverse-direction detection region.

The reference line segment may be a line segment traversing the reverse-direction detection region along the reverse direction. The reference line segment may be a line segment connecting the center points of two line segments intersecting with the backward direction.

4, the virtual line generation module 163 includes two line segments S21, S22, S23, and S24 intersecting the inverse direction among a plurality of line segments S21, S22, S23, and S24 constituting the outline of the inverse- S23). Then, the center points P41 and P42 of the selected two line segments S21 and S23 are connected to each other to generate a reference line segment S25.

When the reference line segment is generated, the virtual line generation module 163 can generate a plurality of virtual lines so as to intersect the reference line segment. Virtual line creation module 163 may select a plurality of points on the reference line segment and create virtual lines to intersect selected points.

4, the virtual line creation module 163 selects a plurality of points P41, P42, P43, and P44 on the reference line segment S45, and selects each of the selected points P41, P42, P43, A plurality of virtual lines L41, L42, L43, L44 and L45 orthogonal to the reference line segment S25 can be generated. The imaginary lines L41 to L42 extend through the respective points P41 to P44 selected on the reference line segment S25 to the line segments S42 and S44 constituting the inverse- ). ≪ / RTI > Accordingly, the virtual lines L41, L42, L43, L44 and L45 extend in the width direction of the road in the road image 5. [

The virtual line generation module 163 can select both end points of the reference line segment as points that intersect with the virtual lines. 4, the virtual line generation module 163 may select both end points P41 and P42 of the reference line segment S25 as points that intersect with the virtual lines L41 and L45.

The virtual line generation module 163 can select points at positions where the reference line segments are equally divided as intersections with the virtual lines.

The virtual line generation module 163 acquires the actual distance between the points corresponding to the start point and the end point of the reference line segment on the road based on the distance map acquired through the camera calibration module 161, Points can be selected as intersections with virtual lines.

The virtual line generation module 163 may calculate the Euclidean length of the reference segment in the road image based on the coordinates of each pixel in the road image and may select the points at the positions to equally divide the Euclidean length as the intersection with the virtual lines . That is, the virtual line generation module 163 may calculate the Euclidean distance between the start point and the end point of the reference line segment, and may select points at positions to equally divide the Euclidean distance.

The virtual line creation module 163 adjusts the number of virtual lines generated according to the number of virtual lines set by the parameter setting module 162 when creating a virtual line.

When the virtual lines are generated, the virtual line generation module 163 sequentially stores the coordinate information of the generated virtual lines in the memory 140 along the reverse direction. In addition, the virtual line creation module 163 may provide identification information such as a line number to each virtual line, and store corresponding identification information corresponding to each line when storing the coordinate information.

The virtual line generation module 163 may further generate vertical virtual lines if the parameter setting module 162 is set to use vertical virtual lines. The vertical virtual lines may extend in directions perpendicular to the virtual lines, i.e., directions perpendicular to the road surface, starting from the end points of the respective virtual lines.

The position of the vertical virtual lines can be set by the user. Parameter setting module 162 allows the user to create a vertical virtual line for at least one of the end points of the virtual line via user input 130, As shown in FIG.

Referring to FIG. 5, the virtual line generation module 163 generates a virtual line by allowing a user to generate a vertical virtual line for the right end point of the virtual lines L41, L42, L43, L44, L52, L53, L54, and L55 starting from the right end point of each of the left-hand end points L41, L42, L43, L44, and L45.

The trajectory detection system 100 tracks the movement of the center point of a region detected by a specific part (e.g., a wheel, a light, a license plate, etc.) of the tracking object or a tracking object when tracking the movement path of the tracking object. In the case of large vehicles such as heavy trucks and large buses, the points to be tracked are relatively high, which may result in a situation that the vehicle is out of the detection area of the reverse direction. The vertical virtual line is set so as to discriminate whether or not the vehicle is reversed, even if a point to be traced in the trajectory is out of the traverse direction detection area.

The length (or height) of the vertical virtual line may correspond to the maximum height value of the size filter for detecting the tracking object. The size filter is a filter for eliminating noise by filtering motion objects excluding a vehicle from a road image. The maximum height value of the size filter may vary depending on the position in the road image to which the size filter is applied. Thus, the length (or height) of the vertical virtual line may vary depending on the position at which the vertical virtual line is created.

When the vertical virtual lines are generated, the virtual line generation module 163 sequentially stores the coordinate information of the vertical virtual lines in the memory 140 along the reverse direction. In addition, the virtual line creation module 163 may provide identification information such as a line number to each vertical virtual line, and store corresponding identification information corresponding to each vertical virtual line when storing the coordinate information.

Based on the learning-based object tracking algorithm and the dynamic background learning-based algorithm, the object tracking module 164 can detect the tracking object corresponding to the vehicle from the road image and acquire the moving path of the tracking object.

The object tracking module 164 analyzes the image frames constituting the road image and detects motion objects from the road image. The object tracking module 164 performs a noise removal process using a size filter and the like on the motion objects detected from the road image, and selects a tracking object corresponding to the tracking object (for example, a vehicle) from among the motion objects .

The size filter is a filter for selecting an object to be tracked among the motion objects detected from the road image. When the motion object is out of the predetermined size (width, height, etc.) range, the object tracking module 164 determines that the motion object is noise and stops the tracking. On the other hand, if the motion object satisfies the predetermined size range, it is determined that the motion object is the vehicle to be tracked and the tracking is started.

The object tracking module 164 tracks movement of a center point of a region detected by a vehicle or a specific part (e.g., a wheel, a light, a license plate, etc.) of the tracking object when the tracking object is detected from the road image, Obtain the movement path.

The object tracking module 164 acquires position information of the tracking object for each image frame, and stores the position information in a buffer of a predetermined size (not shown). When the data amount of the position information obtained by the object tracking module 164 exceeds the buffer size, the position information is deleted in the oldest position order and the new position information is stored in the buffer. Accordingly, the buffer may store the position information of the most recently acquired predetermined number.

The reverse-run determination module 165 can select an in-flight candidate among the track objects detected from the road image.

The inverse determination module 165 can select the tracking object as the inversion candidate when the intersection of the trajectory of the tracking object and the virtual line occurs for the first time. In this case, the inverse-movement determining module 165 may determine that the movement path of the tracking object crosses the virtual line when the movement path of the tracking object satisfies the following conditions.

The inverse-shift determining module 165 connects the oldest positional information and the latest positional information among a plurality of positional information stored in the buffer by the object tracking module 164, and when the straight line crosses the virtual line, It can be determined that the movement path of the object crosses the virtual line.

If the ratio of the position information detected at the position where the tracking object crosses the virtual line among the plurality of position information stored in the buffer is equal to or more than a preset threshold value (for example, 50% or more), the reverse- It can be determined that the path crosses the virtual line. For example, in FIG. 6, when the position information (P601 to P610) of the tracking object stored in the buffer is 10, 5 or more pieces of position information have been detected through the virtual line L41. The straight line S60 connecting the oldest position information P601 with the position information obtained most recently, that is, points corresponding to the current position information P610 intersects with the virtual line L41. Therefore, the inverse-movement determining module 165 can determine that the movement path of the corresponding tracking object crosses the virtual line.

If the vertical virtual line is used, the inverse-movement determining module 165 may select the backward candidate using the vertical virtual line. The inversed-motion determination module 165 may select the tracking object as a backward candidate when the intersection of the trajectory of the tracking object and the vertical virtual line occurs for the first time. In this case, the inverse-travel determining module 165 may determine that the trajectory of the tracing object crosses the vertical virtual line if the traversing path of the tracing object satisfies the following conditions.

The inverse-shift determining module 165 may determine that the trajectory of the tracked object intersects the vertical virtual line when the current position of the tracked object is included in the inverse-tracking area. In this case, the inverse-movement determining module 165 may determine that the tracking object is located in the reverse-direction sensing area when the object area corresponding to the tracking object is overlapped with the reverse-movement sensing area by a certain amount or more.

In addition, the inverse-travel judging module 165 connects the oldest positional information among the plurality of positional information stored in the buffer by the object tracking module 164 with the latest positional information, and when the straight line crosses the vertical virtual line , It can be determined that the movement path of the tracking object crosses the vertical virtual line.

If the ratio of the position information detected at the position where the tracking object passes the vertical virtual line among the plurality of position information stored in the buffer is equal to or more than a predetermined threshold value (for example, 50% or more) It can be determined that the movement path of the object intersects with the vertical virtual line.

When the backward candidate is selected, the backward validity determination module 165 stores the identification information of the virtual line (or vertical virtual line) that the backward candidate first crosses. Also, any one of the position information of the inverse candidate stored in the buffer is stored as the reference position information. The reference position information is position information of a point that is a reference for calculating the moving direction and the moving distance of the inverse candidate when determining the inverse of the inverse candidate.

The inverse-movement judging module 165 may store the position information obtained immediately before the backward candidate passes the virtual line (or vertical virtual line) as the reference position information. The inverse-shift determining module 165 may store the oldest position information among the position information stored in the buffer at the time of selecting the inverse candidate as the reference position information. The reverse-run determination module 165 may store the most recently acquired location information as the reference location information at the time of selecting the in-flight candidate.

The inverse-shift determining module 165 may store the intersection between the virtual line and the movement path of the inverse-running candidate as the reference position information.

When the backward candidate is selected, the backward judging module 165 continuously tracks the position of the backward candidate in the road image to obtain a moving path including the moving direction and the moving distance of the backward candidate. Based on the moving direction and the moving distance of the inversion candidate, it is possible to finally determine whether the inversion candidate is in the reverse direction.

The inverse-movement judging module 165 can obtain the moving distance and the moving direction of the inverse candidate by connecting the point corresponding to the reference position information and the current position of the inverse-moving candidate in a straight line.

7, the inverse-shift determining module 165 acquires the moving distance and the moving direction of the inverse candidate by connecting the point corresponding to the reference position (P701) of the inverse-traveling candidate and the point corresponding to the current position (P702) .

The moving distance of the inverse candidate can be set to an actual distance in an actual road environment. In this case, based on the distance map, the inverse-movement judging module 165 can estimate the actual distance between the reference position and the current position of the inversion candidate and acquire the actual distance as the movement distance of the inversion candidate.

The traveling distance of the inversion candidate may be set as the Euclidean distance. In this case, the inverse-movement judging module 165 may obtain the Euclidean distance between the projected positions in the backward-traveling direction as the movement distance of the inversion-candidate, after projecting the reference position and the current position of the backward- 7, the inverse-shift determining module 165 projects the reference position P701 and the current position P702 of the inverse candidate on the line segment extending in the inverse direction, (D) between the two candidates.

The inverse-movement judging module 165 may judge that the in-flight candidate is in reverse when the moving direction and the moving distance of the in-flight candidate satisfy a predetermined inverse condition over a predetermined interval.

If the degree of agreement between the moving direction of the inversion candidate and the predetermined inversion direction is equal to or greater than a predetermined level or the moving direction of the inversion candidate is within the predetermined inversion range, the inversion determination module 165 can determine that the inversion candidate is in reverse have.

The inverse-travel determining module 165 can determine that the in-flight candidate is in reverse if the travel distance of the in-flight candidate is greater than or equal to the predetermined in-travel distance.

Instead of comparing the travel distance of the inverse candidate, the reverse-run determination module 165 may determine that the inverse candidate is the actual distance between the first virtual line (or vertical virtual line) and the last virtual line (or vertical virtual line) Distance) is equal to or greater than the preset in-travel distance, it may be determined that the inverse candidate is in reverse run.

The inverse-movement judging module 165 may determine that the in-flight candidate is in a reverse run when the movement path of the in-motion candidate passes through a predetermined number or more of the virtual lines in the backward direction.

The camera calibration module 161, the parameter setting module 162, the virtual line creation module 163, the object tracking module 164, and the inverse determination module 164 constituting the control unit 160 in the inverse-movement detection system 100 of the above- The processor 165 may be implemented by a processor implemented in one or more central processing units (CPUs) or other chipset, microprocessor, or the like.

8 schematically shows a method of detecting a reverse run in the reverse run detection system according to the embodiment.

Referring to FIG. 8, the inverse-travel detection system 100 performs camera calibration to generate a distance map for each pixel of the road image obtained through the camera 110 (S100).

In step S100, the backward movement detection system 100 estimates the conversion relationship between the coordinate system (world coordinate system) of the three-dimensional space corresponding to the actual road environment and the coordinate system (image coordinate system) of the two- dimensional road image through camera calibration. Based on the conversion relation between the world coordinate system and the image coordinate system, the distance information between the point corresponding to each pixel and the camera 110 in the actual road environment is obtained. Based on the distance information obtained for each pixel, Can be generated.

The generated distance map can then be used to place the virtual lines corresponding to the actual distance at the time of creation of the virtual line. Also, the moving distance of the inverse candidate can be used to calculate the moving distance in the actual road environment.

The backward movement detection system 100 sets the backward travel direction and the backward movement detection area in the road image based on the user input received through the user input unit 130 at step S110. In addition, the backward movement detection system 100 generates a plurality of virtual lines crossing the backward direction in the backward movement detection area (S120).

In step S120, the backbone detection system 100 may further generate a plurality of vertical virtual lines that are perpendicular to each virtual line when the vertical virtual line is set to be used. The vertical virtual line can also be used to determine whether the large-sized bus, large truck, or other large-sized vehicle is reversed even when the movement path is detected in the reverse-direction detection area, that is, off the road.

The backward movement detection system 100 determines the backward running of the objects in the road image based on the backward direction, the backward movement detection area, and the virtual lines through the above-described process.

The inverse-travel detection system 100 detects an object to be tracked from the road image in order to determine whether the vehicle is to be reversed (S130).

In step S130, the backbone detection system 100 detects motion objects from the road image by analyzing image frames constituting the road image. A noise removal process using a size filter or the like is performed on the motion objects detected from the road image, and a tracking object corresponding to the tracking object is selected from the motion objects.

The backward movement detection system 100 selects an inversion candidate among the tracking objects detected from the road image (S140).

In step S140, the backbone detection system 100 continuously obtains the position information of the tracking object, and based on this, the tracking object corresponding to the movement path of the tracking object at the first intersection with the virtual line (or vertical virtual line) Can be selected.

In the backward candidate selection process, the backward movement detection system 100 connects the oldest positional information and the latest positional information among the positional information of the tracked object obtained recently in a straight line in order to remove noise due to a complicated road environment , And when the straight line intersects with the virtual line (or vertical virtual line), it can be determined that the movement path of the tracking object crosses the virtual line (or the vertical virtual line). If the ratio of the position information detected at the position of the tracking object past the virtual line among the plurality of position information acquired recently for the tracking object is equal to or more than a predetermined threshold value (for example, 50% or more) It can be determined that the path crosses the virtual line.

When the backward running candidate is selected, the backward movement detection system 100 stores any one of the positional information obtained up to the current backward candidate as the reference position information.

Thereafter, the inverse-travel detection system 100 continuously obtains the moving direction and the moving distance of the inversion candidate based on the reference position information (S150). Then, it is determined whether the moving direction and the moving distance of the inversion candidate satisfy the predetermined inverse running condition (S160).

If the moving direction of the backward candidate is equal to or greater than a predetermined level with the predetermined backward direction or within a predetermined backward range based on the backward direction, the backward movement detection system 100 determines whether the backward movement candidate is in a backward traveling condition Can be determined to satisfy the following formula.

In step S160, the reverse-travel detecting system 100 may determine that the moving distance of the in-flight candidate meets the back-running condition when the moving distance of the in-flight candidate is greater than or equal to the predetermined in- The reverse traversal detection system 100 may compare the actual distance between the first virtual line (or vertical virtual line) and the last virtual line (or vertical virtual line) Distance) is equal to or greater than the preset in-travel distance, it may be determined that the moving distance of the inverse candidate is satisfied with the backward running condition. The backward movement detection system 100 may determine that the moving distance of the backward candidate is satisfying the backward running condition if the movement path of the backward candidate passes through a predetermined number or more of virtual lines in the backward direction.

The inverse-travel detecting system 100 determines that the inverse-running situation has occurred (S180) if the moving direction and the moving distance of the inverse-traveling candidate satisfy the inverse-traveling condition for a preset interval (S170).

In step S170 and step S180, in order to prevent the inaccurate determination error from occurring due to the complexity of the road environment due to the complexity of the trajectory tracking of the backward candidate, the movement path and the moving direction of the backward candidate It is possible to make a final determination that the inversion candidate is in reverse run only if the condition for reverse run exceeding the set interval is satisfied.

The backward movement detection system 100 outputs a warning output (e.g., a warning sound, a lamp blinking, a light emitting diode blinking, or the like) warning that the backward candidate is determined to be running in reverse (S190).

According to the above-described embodiment, the reverse-travel detection system can determine whether the objects in the road image are reversed by using the virtual lines. Therefore, it is possible to determine the reverse direction of the vehicle only by the image acquired through the camera without any additional equipment, There is an advantage of lowering. In addition, it is possible to select the inversion candidate only by intersection between the movement path of the object and the virtual line, and it is possible to determine the inversion by merely comparing the movement distance and the movement direction of the inversion candidate, and the complexity of the inversion determination process is very low. In addition, it is possible to determine the inversion of travel even when the travel path of the vehicle is detected or departed from the road by using the vertical virtual line, and it is possible to stably determine the inversion even in a situation where object tracking accuracy is low due to complicated road environment. In addition, when the backward running candidate satisfies the backward running condition for a predetermined period or more, it is determined that the backward running situation has occurred, and the accuracy of backward determination is increased.

In the above-described embodiment, the vehicle traveling in the reverse direction is detected by using the virtual lines. However, the virtual lines may be used to detect objects moving in a specific direction in another object detection system. In this case, the object detection system sets the movement direction of the object and the sensing area for detecting movement of the object in the image, and generates a plurality of virtual lines based on the movement direction, and can use the object to detect an object moving in a specific direction.

The method of detecting the backward movement according to the embodiment of the present invention can be executed through software. When executed in software, the constituent means of the present invention are code segments that perform the necessary tasks. The program or code segments may be stored on a processor read functional medium or transmitted by a computer data signal coupled with a carrier wave in a transmission medium or network.

A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording device include ROM, RAM, CD-ROM, DVD-ROM, DVD-RAM, magnetic tape, floppy disk, hard disk and optical data storage device. Also, the computer-readable recording medium may be distributed over a network-connected computer device so that computer-readable code can be stored and executed in a distributed manner.

It is to be understood that both the foregoing general description and the following detailed description of the present invention are illustrative and explanatory only and are intended to be illustrative of the invention and are not to be construed as limiting the scope of the invention as defined by the appended claims. It is not. Therefore, those skilled in the art can readily select and substitute it. Those skilled in the art will also appreciate that some of the components described herein can be omitted without degrading performance or adding components to improve performance. In addition, those skilled in the art may change the order of the method steps described herein depending on the process environment or equipment. Therefore, the scope of the present invention should be determined by the appended claims and equivalents thereof, not by the embodiments described.

Claims (26)

A camera for photographing a road image,
A parameter setting module for setting an inverse direction in the road image based on a user input,
A virtual line generation module for generating a plurality of virtual lines crossing the backward direction,
An object tracking module for detecting a tracking object from the road image, and
If the movement path of the tracking object intersects with any one of the plurality of virtual lines, the tracking object is selected as the inversion candidate, and if the movement direction and the movement distance of the inversion candidate satisfy the inversion condition, And an inverse run determination module,
Wherein the parameter setting module sets an inverse traverse detection area in the road image based on a user input,
Wherein the virtual line generation module generates reference line segments by connecting the center points of two line segments intersecting with the inverse direction of the plurality of line segments constituting the outline of the inverse running sensing area to each other, And generating the plurality of virtual lines so as to cross each of the plurality of points. The method according to claim 1,
Wherein the object tracking module continuously stores location information of the tracking object in a buffer,
Wherein if the ratio of the positional information detected at a position past any one of the virtual lines among the positional information of the tracked object stored in the buffer is equal to or greater than a predetermined value, Line of the vehicle. The method according to claim 1,
Wherein the virtual line generation module generates a plurality of vertical virtual lines extending in a direction perpendicular to the road surface from the end points of the respective virtual lines. The method of claim 3,
Wherein the traversing determination module selects the tracing object as a traversal candidate when the traversing path of the tracing object intersects with any one of the vertical virtual lines. 5. The method of claim 4,
Wherein the object tracking module continuously stores location information of the tracking object in a buffer,
Wherein if the ratio of the positional information detected at a position past the one of the vertical virtual lines in the positional information of the tracked object stored in the buffer is equal to or greater than a predetermined value, And determines that the virtual line crosses the virtual line. The method according to claim 1,
Wherein the object tracking module continuously stores location information of the tracking object in a buffer,
Wherein the reverse motion determining module sets any one of the position information stored in the buffer as reference position information when the tracking object is selected as the inversion candidate, Back direction detection system. The method according to claim 6,
Wherein the backward movement determining module acquires a direction of a straight line connecting a point corresponding to the reference position information and a current position of the inversion candidate in the road image in a moving direction of the inversion candidate. 8. The method of claim 7,
Wherein the reverse-driving judging module judges that the moving direction of the backward running candidate satisfies the backward running condition if the degree of matching between the moving direction of the backward running candidate and the backward traveling direction is equal to or greater than a predetermined level. The method according to claim 6,
Wherein the inverse-movement determining module acquires a distance between a point corresponding to the reference position information and a current position of the in-flight candidate in the road image, as a travel distance of the in-motion candidate. 10. The method of claim 9,
Wherein the reverse-driving judging module judges that the moving distance of the backward running candidate satisfies the backward running condition when the moving distance of the backward running candidate is not less than a predetermined in-travel distance. The method according to claim 1,
Wherein the inverse-movement determining module determines that the moving distance of the in-motion candidate meets the back-running condition when the moving path of the in-motion candidate crosses a predetermined number or more of virtual lines. The method according to claim 1,
Wherein the reverse-drive judging module judges that the reverse-running state is the reverse-running state if the state in which the moving distance and the moving direction of the reverse-running candidate satisfy the reverse-running condition is maintained for a predetermined period. A method for detecting an inverse movement of an inverse movement detection system,
Obtaining a road image through a camera,
Setting an inverse direction in the road image based on a user input,
Generating a plurality of virtual lines that intersect the inverse direction,
Detecting a tracking object from the road image,
Selecting the tracing object as a traversal candidate when the trajectory of the tracing object crosses one of the plurality of virtual lines;
And determining that the reverse direction is a reverse direction when the moving direction and the moving distance of the backward candidate are satisfied,
Wherein the generating of the plurality of virtual lines comprises:
Setting an inverse traverse detection area in the road image based on a user input,
Generating a reference segment by connecting center points of two line segments intersecting with the inversed direction among a plurality of line segments constituting the outline of the inverse detection area,
Selecting a plurality of points spaced from each other on the reference line segment; and
And generating the plurality of virtual lines to intersect each of the plurality of points. delete 14. The method of claim 13,
Further comprising generating a plurality of vertical virtual lines extending in a direction perpendicular to the road surface from an end point of each virtual line. 16. The method of claim 15,
And selecting the tracing object as the inverse-running candidate if the trajectory of the tracing object crosses any of the vertical virtual lines. 14. The method of claim 13,
Continuously storing the location information of the tracked object in a buffer,
Setting one of the positional information stored in the buffer as the reference positional information when the tracking object is selected as the inverse-running candidate; and
And obtaining a movement path and a movement direction of the backward candidate based on the reference position information. 18. The method of claim 17,
Wherein the step of acquiring the moving path and the moving direction of the inverse-
And obtaining a direction of a straight line connecting a point corresponding to the reference position information and a current position of the inverse candidate in the moving image of the inverse candidate in the road image. 19. The method of claim 18,
Wherein the step of determining the reverse-
And judging that the direction of movement of the inverse-running candidate satisfies the inverse-running condition if the degree of agreement between the direction of movement of the in-flight candidate and the direction of inversion is equal to or greater than a predetermined level. 18. The method of claim 17,
Wherein the step of acquiring the moving path and the moving direction of the inverse-
And obtaining a distance between a point corresponding to the reference position information and a current position of the inversion candidate in the road image as a movement distance of the inversion candidate. 21. The method of claim 20,
Wherein the step of determining the reverse-
And determining that the moving distance of the backward candidate is greater than or equal to the predetermined backward traveling distance, if the moving distance of the backward candidate is equal to or greater than a predetermined in-travel distance. 14. The method of claim 13,
Wherein the step of determining the reverse-
And determining that the travel distance of the backward candidate is satisfying the backward running condition if the travel path of the backward candidate crosses a predetermined number or more of virtual lines. 14. The method of claim 13,
Wherein the step of determining the reverse-
And determining that the backward running situation is the backward running state if the state in which the moving distance and the moving direction of the backward running candidate satisfy the backward running condition is maintained for a predetermined period. A method for generating a virtual line of an object detection system,
Acquiring an image through a camera,
Setting an object moving direction in the image based on a user input,
Setting an object detection area in the image based on user input, and
Generating a plurality of virtual lines intersecting the object moving direction in the object detection area,
Wherein the generating of the plurality of virtual lines comprises:
Generating a reference segment by connecting center points of two line segments intersecting the object movement direction among a plurality of line segments constituting an outline of the object detection region,
Selecting a plurality of points spaced from each other on the reference line segment; and
And generating the plurality of virtual lines to intersect each of the plurality of points,
Wherein the plurality of virtual lines are used to detect a movement path of an object in the image. 25. The method of claim 24,
Further comprising generating a plurality of vertical virtual lines extending from an end point of each virtual line in a direction perpendicular to the floor surface. delete

Images