KR20160079511A - Device and method for recognizing parking stall - Google Patents

Abstract

The present invention relates to a method for recognizing a parking section around a vehicle when the vehicle is parked and a device for performing the same. According to the present invention, the method for recognizing a parking section comprises the following steps of: obtaining an image around a vehicle; obtaining edge images from the image in accordance with first orientation and second orientation; extracting a plurality of straight lines from the edge images; and specifying a parking section based on the plurality of straight lines.

Description

Technical Field [0001] The present invention relates to a device and method for recognizing a parking zone,

The present invention relates to a method for recognizing a parking zone around the vehicle when the vehicle is parked, and to an apparatus for performing the method.

Inexperienced drivers, such as novice drivers, typically have a great deal of difficulty parking in locations between parked cars. To this end, a number of vehicles employ an ultrasonic parking assist system that warns the driver of the stopping distance so that inexperienced drivers can avoid other cars, people, obstacles, etc. parked.

On the other hand, in order to provide more convenience to the driver, an automatic parking system that is more advanced than the parking assist system is being developed. In the case of the automatic parking system, the recognition of the parking zone should be performed first, and the image-based parking zone recognition method using the camera or the like is mainly used.

However, according to the conventional image-based parking zone recognition technology, it is difficult to accurately recognize parking lot lines due to shadows, rain, or snow.

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above problems of the prior art, and it is an object of the present invention to provide an apparatus and method for detecting an edge in a specific direction and recognizing a parking section based on the detected edge.

The parking lot detecting apparatus according to an embodiment of the present invention includes an image obtaining unit for obtaining an image around a vehicle, an image preprocessing unit for obtaining an edge image from the image according to a first orientation and a second orientation, A straight line extracting unit that extracts a plurality of straight lines from the straight line, and a parking space specifying unit that specifies a parking space based on the plurality of straight lines.

Further, in the parking lot detecting apparatus according to the embodiment of the present invention, the first orientation and the second orientation may have an angle difference of 180 degrees.

Further, in the parking lot detecting apparatus according to the embodiment of the present invention, the straight line extracting unit can extract the plurality of straight lines by applying the Hough transform to the edge image.

Further, in the parking lot detecting apparatus according to the embodiment of the present invention, the parking section specifying unit may obtain a first direction partition line from a pair of straight lines that are parallel to each other and spaced apart from each other by a predetermined width among the plurality of straight lines, And a first direction partition line obtaining section for specifying the parking section based on the one-way partition line.

Further, in the parking lot detecting apparatus according to the embodiment of the present invention, a plurality of the first direction partition lines are obtained, and the parking space specifying section measures the brightness along the plurality of first direction partition lines, Further comprising: a second directional segment obtaining unit for extracting a point that changes in the first directional segment and the second directional segment to obtain a second directional segment line including the point, And specifying the parking zone.

According to an embodiment of the present invention, the second directional segment obtaining unit may obtain the second directional segment when the brightness difference value among the local minimum values of the brightness difference profile graph for the plurality of first directional segment lines is smaller than a predetermined value The location can be extracted.

In the parking lot detection apparatus according to an embodiment of the present invention, the second directional segment obtaining unit may obtain a straight line as a result of applying the residual algorithm to the point as the second directional segment.

According to the parking section recognition method according to various embodiments of the present invention, the edge is detected according to a specific direction and the parking section is recognized based on the detected edge, so that the throughput can be further reduced as compared with the case where edges are detected in all directions. Further, since the edge is detected in a specific direction, the partition line of the parking section can be recognized robustly against various noises.

1 shows a vehicle and a recognition environment on which a parking lot recognition apparatus according to an embodiment of the present invention is mounted.
2 shows a configuration of a parking zone recognition apparatus according to an embodiment of the present invention.
3 shows a flowchart of a method of recognizing a parking zone according to an embodiment of the present invention.
FIG. 4 shows a detailed flowchart of a method of recognizing a parking zone according to an embodiment of the present invention.
5 shows an image acquired by the image acquisition unit of the parking zone recognition apparatus according to an embodiment of the present invention.
6 shows an image of an edge detected by the image preprocessing unit of the parking lot recognition apparatus according to an embodiment of the present invention.
FIG. 7 is a graph illustrating a Huff transform performed by the straight line extracting unit of the parking lot recognizing apparatus according to an embodiment of the present invention. FIG.
8 shows a straight line extracted in the method of recognizing a parking lot according to an embodiment of the present invention.
9 shows a first directional segment line in the method of recognizing a parking zone according to an embodiment of the present invention.
Fig. 10 shows a parking section candidate in the parking section recognition method according to an embodiment of the present invention. Fig.
11 shows a parking section candidate of a closed rectangular parking section in the parking section recognition method according to an embodiment of the present invention.
12 shows a parking section candidate of an open rectangular parking section in the parking section recognition method according to an embodiment of the present invention.
13 shows a parking section candidate of a linear parking section in the parking section recognizing method according to an embodiment of the present invention.
14 shows a parking section candidate of a stepped parking section in the method of recognizing a parking section according to an embodiment of the present invention.
15 is a diagram for explaining conditions for specifying a second directional segment in the method of recognizing a parking lot according to an embodiment of the present invention.
FIG. 16 is a view for explaining a point where a brightness rapidly changes on a first directional partition line in a method of recognizing a parking zone according to an embodiment of the present invention. FIG.
17 is a view for explaining a second directional partition line in the method of recognizing a parking zone according to an embodiment of the present invention.
18 shows an image obtained by obtaining a second directional segment in various types of parking segments according to an embodiment of the present invention.
Figure 19 shows a parking section with further consideration of the parking compartment candidate and the second directional zone in various types of parking zones according to an embodiment of the present invention.
20 is a view for explaining a method of recognizing a stepped parking section according to an embodiment of the present invention.
21 is a diagram for describing a method of recognizing a type of parking zone according to various embodiments of the present invention.
FIG. 22 shows a closed rectangular parking section recognized by type in the parking section recognition method according to an embodiment of the present invention. FIG.
FIG. 23 shows an open rectangular parking section recognized by type in the parking lot recognition method according to an embodiment of the present invention.
FIG. 24 shows a four-line parking zone recognized by type in the parking lot recognition method according to an embodiment of the present invention.
FIG. 25 shows a stepped parking section recognized by type in the parking section recognition method according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, the terms "comprises" or "having" are used to specify that a feature, a number, an element, a component, or a combination thereof is described in the specification and that the term " Elements, parts, or combinations thereof without departing from the spirit and scope of the invention.

1 shows a vehicle and a recognition environment on which a parking lot recognition apparatus according to an embodiment of the present invention is mounted.

Referring to FIG. 1, a parking lot recognition apparatus 1000 according to an embodiment of the present invention may be mounted on a vehicle. The parking zone recognition apparatus 1000 may include an AVM (Around View Monitoring) apparatus having a plurality of cameras, or may be connected at least. The AVM apparatus can acquire a plurality of images around the vehicle as shown in FIG. 1, and can display a plurality of images as a single image through a predetermined process. The parking lot recognition apparatus 1000 can acquire an image through the AVM apparatus and perform the parking lot recognition method according to the present invention.

2 shows a configuration of a parking zone recognition apparatus according to an embodiment of the present invention.

Referring to FIG. 2, a parking lot recognition apparatus 1000 according to an embodiment of the present invention includes an image acquisition unit 101, an image preprocessing unit 102, a straight line extraction unit 103, and a parking lot identification unit 104 ).

The image acquiring unit 101 can acquire images around the vehicle. The image acquisition unit 101 may receive images taken from, for example, a front camera, a rear camera, and side mirror cameras on both sides. In addition, the image acquiring unit 101 may acquire images (so-called Around View images) around the vehicle by combining images transmitted from the plurality of cameras.

The image preprocessing unit 102 may acquire an edge image from the image acquired by the image acquisition unit 101 according to the first orientation and the second orientation. Orientation may mean a predetermined direction for detecting an edge, and may be expressed as an angle. The image preprocessing unit 102 may improve the detection accuracy by reducing a false detection or detection missing by applying a relatively low threshold value (threshold value) after the median filtering at the edge detection.

Meanwhile, in various embodiments of the present invention, the image preprocessing unit 102 performs the image preprocessing in the six directions (-90 degrees / 90 degrees / -45 degrees / 135 degrees / 45 degrees / 135 degrees). That is, the first and second orientations may be any one of the six orientation angles.

The image preprocessing unit 102 can detect the partition line of the parking space in accordance with the first and second orientations separately from the other unnecessary straight lines. Further, since the edge is detected in accordance with the predetermined orientation, the throughput can be reduced and the processing speed can be increased as compared with the case of detecting the edge in all directions. The first orientation and the second orientation may preferably have an angle difference of 180 degrees.

In addition to the edge detection, the image preprocessing unit 102 may convert an image around the vehicle into a monochrome image. However, if the image obtained by the image acquiring unit 101 is a monochrome image, the image preprocessing unit 102 may not convert the monochrome image into the monochrome image.

The straight line extracting unit 103 can extract (acquire) a plurality of straight lines by applying a Hough Transform to an edge image acquired by the image preprocessing unit 102, for example. At this time, the plurality of straight lines extracted by the straight line extracting unit 103 may be a part of a straight line such as a line segment. In this specification, the term "straight line" should not be construed to be limited to a straight line extending at both ends indefinitely. In addition, the technique of extracting a plurality of straight lines by the straight line extracting unit 103 is not limited to the Hough transform and can be extracted by another straight line extraction technique.

The parking section specifying section 104 can specify the parking section based on the plurality of straight lines extracted by the straight line extracting section 103. [ The parking section specifying section 104 can specify the parking section based on the first directional partition line and / or the second directional partition line. The parking section identifying section 104 may include a first direction partition line obtaining section 111 and a second direction partition line obtaining section 112.

The first direction partition line obtaining unit 111 can obtain the first direction partition line based on the straight line pairs that are parallel to each other and spaced apart from each other by a predetermined width among the extracted straight lines. In the present specification, the first direction may mean a longitudinal direction of the parking space or a direction perpendicular to the running direction of the vehicle. The predetermined width between the straight line pairs means the width of the partition line of the parking space, and may have a value of 10 to 20 cm. For example, the first direction partition line obtaining unit 111 may obtain the center line of the straight line pair as the first direction partition line.

The second direction partition line obtaining unit 112 measures brightness along the plurality of first direction partition lines, extracts a point where the brightness suddenly changes, and detects a second direction partition line including a point where the brightness suddenly changes, Can be obtained. In this specification, the second direction may mean the width direction of the parking space or the same direction as the running direction of the vehicle.

For example, the point at which the brightness suddenly changes is obtained by extracting a position where an intensity difference profile value among local minima of the brightness difference profile graph for a plurality of first directional partition lines is smaller than a predetermined value can do. The second direction partition line obtaining unit 112 can obtain a straight line as a result of applying the roughness to the point where the brightness suddenly changes as the second direction partition line. The manner in which the second directional section line obtaining section 112 obtains the second directional section line is not limited to the above-described example, and various methods can be applied to obtain the second directional section line.

The configuration of the parking lot recognition apparatus 1000 according to an embodiment of the present invention is the same as described above. Hereinafter, a method of recognizing a parking lot according to an embodiment of the present invention will be described with reference to FIG. 3 to FIG.

3 shows a flowchart of a method of recognizing a parking zone according to an embodiment of the present invention.

Referring to FIG. 3, the method for recognizing a parking lot according to an embodiment of the present invention includes a step S301 of acquiring an image of a vehicle periphery by an image acquisition unit 101, a step S302 of acquiring an image from the image obtained in step S301, (S302) of extracting a plurality of straight lines from the edge images by the straight line extraction unit 103 and a step S303 of extracting a plurality of straight lines from the edge images based on the edge information 104) of identifying a parking zone based on the plurality of straight lines (S304).

FIG. 4 shows a detailed flowchart of a method of recognizing a parking zone according to an embodiment of the present invention.

Referring to FIG. 4, a method of recognizing a parking zone according to an embodiment of the present invention includes a step S401 of acquiring an image of the surroundings of a vehicle, a step S402 of acquiring an image from the image obtained in the step S401 by the image preprocessing unit 102, A step S403 of obtaining an edge image by performing a Hough transform for each edge image in step S403, a step S404 of obtaining a first directional segment from the straight line extracted in step S403, (S405), a step S406 of measuring the brightness of the first parting lines, a step S407 of extracting a point where the brightness suddenly changes (S407), a step A step S408 of obtaining a second directional segment line, and a step S409 of specifying a parking segment formed by the first and second directional segment lines.

In step S401, the image acquisition unit 101 can acquire an image around the vehicle. As described above, the image of the surroundings of the vehicle can be photographed and received by the front camera, the rear camera, and the side mirror cameras on both sides. Also, the image obtaining unit 101 may combine images obtained from a plurality of cameras to form a so-called Around View image. For example, in step S401, the image acquired by the image acquisition unit 101 may be the image shown in FIG. Referring to FIG. 5, there is shown a combined image of four images of front and rear sides and both sides. The black rectangle in the middle of the image may be the area corresponding to the vehicle itself that can not be obtained from the cameras.

In step S402, the image preprocessing unit 102 may obtain an edge image from the image acquired in step S401 according to the first orientation and the second orientation. For example, a Sobel operator may be used to acquire an edge image. The image preprocessing unit 102 may improve the detection accuracy by reducing a false detection or detection missing by applying a relatively low threshold value (threshold value) after the median filtering at the edge detection.

The image preprocessing unit 102 prepares the edge according to six directions (-90 degrees / +90 degrees / -45 degrees / +135 degrees / +45 degrees / -135 degrees) counterclockwise Can be detected. For example, referring to FIG. 6, the image 610 represents an image obtained by detecting an edge according to all orientations (angles), and the images 601 through 606 represent orientations of -90 degrees / +90 degrees / -45 degrees / + 135 degrees / +45 degrees / -135 degrees, and the edge is detected. In the image 610 of Fig. 6, since edges in all directions are detected, the outline of the parking lot line is mostly shown. On the other hand, in the images 601 and 602, only the edge having the orientation of -90 degrees and + 90 degrees in the counterclockwise direction was detected at the front of the vehicle at zero degree. It can be seen that a significant edge is not detected in the edge images 603 to 606 having the orientations of -45 degrees / + 135 degrees / + 45 degrees / -135 degrees, respectively. According to the parking lot recognition apparatus 1000 and method according to an embodiment of the present invention, as will be described later, in recognizing the parking section, edges having a right angle or a slope of 45 degrees (or 135 degrees) .

Therefore, in the parking lot recognition apparatus 1000 and method according to the embodiment of the present invention, for example, only the images 601 and 602 showing the edge of FIG. 6 can be acquired and used for parking segment recognition. As a result, the partition line of the parking section can be distinguished from the other unnecessary straight lines and can be detected. Further, since the edge is detected according to the predetermined orientation, the throughput can be reduced and the processing speed can be increased, in a case where the edge is detected in all directions (for example, in the case of the image 610). In particular, the first orientation and the second orientation may have an angle difference (= + 90 - (- 90)) of 180 degrees as in the images 601 and 602 of FIG.

In step S403, the straight line extracting unit 103 can perform a Hough transform for each edge image to extract a straight line. For example, the straight line extracting unit 103 can perform the Hough transform on each edge pixel of the images 601 to 606 in Fig. 6, and as a result, transformed graphs such as the parameter spaces 701 to 706 in Fig. 7 Can be obtained. Referring to parameter spaces 701 and 702 in FIG. 7, a graph for each of the edges of images 601 through 606 in FIG. 6 is shown on the parameter space. The straight line extracting unit 103 can specify a point where the graphs of the parameter spaces 701 and 702 intersect each other more than a predetermined number of times (a "+" indication of the parameter spaces 701 and 702). In the parameter spaces 701 and 702, the "+" mark has the same abscissa value because the edges for the parking segment form a parallel straight line (i.e., have the same slope).

For example, the straight line extracting unit 103 can extract a straight line on the acquired image based on the parameter value at the point indicated by "+" in the parameter spaces 701 and 702. On the other hand, since there is no significant edge in the parameter spaces 703 to 706 in the images 603 to 606, there is no significant graph.

8 is an image showing a straight line extracted in the method of recognizing a parking zone according to an embodiment of the present invention. For example, FIG. 8A shows an image obtained by applying a Hough transform to edges of all the rooms shown in the image 610 of FIG. 6 to extract a straight line. If the edge is detected (e.g., image 610) in accordance with all the orientations (angles), and Huff transform is performed on all the detected edges, it becomes difficult to extract a straight line associated with the partition line of the parking segment due to noise or the like. For example, in the straight lines detected in the first and third partition lines in Fig. 8A, false detection (straight line 801) and detection missing (dotted line 802) occur.

On the other hand, FIG. 8 (b) shows an image obtained by applying a Hough transform to the edges of the images 601 and 602 in FIG. 6 and extracting a straight line based on the graphs of the parameter spaces 701 and 702 in FIG. Referring to FIG. 8 (b), it can be confirmed that straight lines associated with the partition lines of all the parking lots are extracted correctly (with high accuracy). The green straight line in Fig. 8 (b) is a straight line extracted based on the edge whose orientation is -90 degrees (i.e., 601 in Fig. 6, 701 in Fig. 7) and the blue straight line is based on the edge whose orientation is + May be an extracted straight line (i.e., 602 in Fig. 6, 702 in Fig. 7).

8 (b), unlike the case of FIG. 8 (a), no edge image is obtained for all orientations, and since Hough transform and straight line extraction are not performed for all edges, 8 (a).

In step S404, the first direction partition line extraction unit 111 of the parking lot identification unit 104 can acquire the first direction partition line from the straight line extracted in step S403. Specifically, the first direction partition line extracting unit 111 can extract a straight line pair having an angle difference of 180 degrees among the straight lines extracted in step S403, which are parallel to each other and spaced apart from each other by a predetermined width. The first direction partition line extraction unit 111 can acquire the intermediate line between the straight line pairs as the first direction partition line. At this time, the predetermined width means the width of the partition line of the parking space as described above, and may have a value between 10 and 20 cm.

For example, in the image of Fig. 8 (b), the green and blue straight lines are straight lines derived from edges having an angle difference of 180 degrees from each other, parallel to each other, and spaced apart by a predetermined width. The first direction partition line extraction unit 111 can extract the middle line of the green and blue straight lines and can obtain the first direction partition line that can constitute the parking section. For example, referring to FIG. 9, the red straight line on the black parking lot partition line corresponds to the first direction partition line.

In step S405, the parking section specifying unit 104 can generate the parking section candidate based on the first directional partition line. For example, the parking section identification unit 104 identifies an area formed by intersecting the first direction partition line extracted in step S404 and a virtual straight line extending from the side of the vehicle (longitudinal direction segment of the black box) As shown in FIG. At this time, it is general that the first direction partitioning line forming the parking division candidate is parallel to other neighboring first direction partitioning lines and is spaced apart from each other by a predetermined parking partition width (2 ~ 3m).

For example, a first directional segment line indicated by a red straight line in FIG. 9 can generate a parking segment candidate along with a side extension line of a black rectangular area (corresponding to a vehicle area not recognized by the camera) in the center of the image. The generated parking section candidate can be grasped from the area surrounded by the dotted line in Fig.

Figs. 11 to 14 show steps of generating a parking section candidate for each type of parking section. Fig. 11, the images 1111 and 1121 that can be obtained in step S401, the images 1112 and 1122 in which the straight lines extracted in steps S402 and S403, and the images 1112 and 1122 in which the first directional division line obtained in step S404 is displayed And 1123, and the images 1114 and 1124 showing the parking section candidates (closed rectangular parking sections) generated based on the first directional partition line in step S405. Fig. 12 shows a case of an open rectangular parking section, Fig. 13 shows a case of a quadruple parking section, and Fig. 14 shows a case of a stepped parking section.

The parking zone candidates can be specified by the steps S401 to S405. In recognition of the parking section, the generated parking section candidate may be satisfied as the parking section. However, if the second directional section line perpendicular to the first directional section line is specified, it is possible to recognize the parking section with higher precision.

Fig. 15 is a diagram for explaining conditions for specifying the second directional segment; Fig.

Referring to Fig. 15, the red straight line represents the first direction partition line, and the green point represents the start point of the so-called partition line for acquiring the second direction partition line. The second directional partition line can be obtained by connecting the starting point of the above-mentioned partition line.

In various embodiments of the present invention, in order to obtain the second directional line, it is possible to use that the "brightness" at the starting point (green point) changes abruptly and the starting points exist on a straight line. Therefore, in order to specify the parking zones in the various types of parking zones as shown in Figs. 15 (a) to (d), the detection of the green points (the starting points of the zone lines) and the second direction zone lines connecting them something to do.

Returning to Fig. 4, in step S406, the parking section identifying unit 104 may measure the brightness along the first partition line. In this case, "brightness" can refer to a value obtained when pixels in an image are converted to grayscale. Normally, "brightness" has a value from 0 to 255, and a value of 255 is the brightest. For example, in Fig. 16 (a), the line segment ab represents a part of the first direction partition line. Fig. 16 (b) shows a graph of the intensity profile during the time from point a to point b.

In step S407, the second directional partition line obtaining unit 112 of the parking lot identifying unit 104 can extract a point where the brightness measured in step S406 changes rapidly. For example, a position where the intensity difference value among the local minima of the intensity difference profile graph for the first directional section line is smaller than a predetermined value is extracted as a point at which the brightness rapidly changes . At this time, it may be considered to extract the position having the smallest brightness difference value among the local minimum values of the brightness difference profile graph as the point where the measured brightness rapidly changes. In such a case, since a wrong point may be extracted due to external influences such as shadows, a position having a brightness difference value smaller than a predetermined value among the local minimum values of the brightness difference profile graph can be extracted.

For example, referring to FIG. 16 (c), a brightness difference profile graph for the brightness graph of FIG. 16 (b) in the ab section is shown. The parking lot recognizing apparatus 1000 can extract, for example, points of points 1603 and 1604 in Fig. 16 (c) as points where the brightness suddenly changes. Points 1603 and 1604 may correspond to points 1601 and 1602 in the image of Fig. 16 (a).

In step S408, the parking section identifying unit 104 can acquire the second directional segment based on the point where the brightness extracted in step S408 is suddenly changed. For example, the parking lot identifying unit 104 may obtain a straight line derived as a result of applying RANSAC (Random Access Consensus) to the points where the brightness rapidly changes, as the second directional segment. On the other hand, a Least Mean Square (LMS) may be used to derive the second directional segment.

For example, referring to FIG. 17 (a), the parking lot identification unit 104 can apply a lance to the points 1701 to 1708 where the brightness extracted in step S408 rapidly changes. As a result, points 1701, 1703, 1704, and 1706 can be extracted as inliers. The parking section identifying section 104 can obtain the second directional partition line 1710 shown in Fig. 17 (b) from these inlier points. Figs. 18 (a) to 18 (d) show an image obtained by obtaining a second directional segment in various parking segment types.

In step S409, the parking section specifying unit 104 can specify the parking section formed by the first and second directional partition lines. According to the steps S406 to S409, the parking section specifying unit 104 can specify a parking section having a higher precision than the parking section candidate in consideration of the parking section candidate and the second directional partition line. For example, Figures 19 (a) - (d) show a parking section with additional consideration of a parking compartment candidate and a second directional segment for various types of parking compartment. However, referring to Fig. 19 (d) showing the stepped parking section, unlike other parking section types, the specified parking sections are somewhat indefinite overlapping with an angle difference of 90 degrees. Therefore, additional processing is required when recognizing a stepped parking section.

20 is a view for explaining a method of recognizing a stepped parking section according to an embodiment of the present invention.

Referring to Fig. 20, Fig. 20 (a) shows the parking section specified by steps S401 to S409 in Fig. The parking section specifying unit 104 can specify the parking section as shown in Fig. 20 (b) based on the image of Fig. 20 (a) and the edge image of the image of Fig. 20 (a).

Specifically, the parking section specifying unit 104 may count the number of edge pixels on the first direction partition line and the second direction partition line. The relatively small number of edge pixels on a parcel line may mean that the parcel is less likely to be an actual parking lot.

In addition, the parking section identification unit 104 can measure the brightness on the first direction partition line and the second direction partition line of the parking section shown in the image of Fig. 20A, for example, and calculate the average brightness value. The fact that the average brightness is relatively high on the parting line may mean that the parting line is likely to be an actual parking lot line.

The parking section identifying unit 104 normalizes the number of the edge pixels and the average brightness value for each of the partition lines and selects a partition line having a relatively high value among the values obtained by adding the two normalized values, ) Can be specified.

20 (c) is a view for explaining another method of recognizing the stepped parking section. The parking section identification unit 104 sets the fitness of each cluster to the best value of the fitness of the parking compartment belonging to the cluster after clustering parking spaces having the same direction and type, unlike the case of FIG. 20 (b) So that comparison between the clusters can be performed. For example, in FIG. 20 (c), the blue region and the green region can be grouped into one cluster, and the one cluster can be compared with the cluster including only the red dotted line region.

According to the parking section recognition method according to various embodiments of the present invention, the edge is detected according to a specific direction and the parking section is recognized based on the detected edge, so that the throughput can be further reduced as compared with the case where edges are detected in all directions. In addition, since the edge is detected according to a specific direction, it is robust against various noise such as deterioration of image quality occurring at night, difference in brightness of a parting line caused by shadows, a damaged parking lot line, stains due to rain or snow, It is possible to recognize the parting line of the image.

In addition, according to the method of recognizing a parking zone according to various embodiments of the present invention, since the second directional segment line is obtained from the point where the brightness of the first directional segment of the parking segment rapidly changes, It becomes possible.

Meanwhile, according to various embodiments, the method of recognizing a parking lot according to various embodiments of the present invention may include a method of recognizing the type of the recognized parking lot.

The method of recognizing the type of the parking section is that the parking section identifying section 104 first determines an angle between the first direction partition line and the front direction (or running direction) of the vehicle, and the step of determining the brightness value of the second direction partition line, And determining the type based on the received information.

The parking section identifying section 104 can determine the specified parking section as a closed or open rectangular parking section when the angle between the sides is about 90 degrees. Subsequently, the parking section identifying section 104 can measure the brightness of the second direction partition line (the green straight line in Fig. 21 (a)) and calculate the average. If the average value of the brightness is equal to or greater than the predetermined threshold value, the parking zone identifying unit 104 can determine that the parking space is a closed rectangular parking space. If the average value is less than the predetermined threshold value,

As another example, the parking section identifying section 104 may determine the type based on the edge of the second directional section after determining the closed or open rectangular parking section. That is, the parking section identifying unit 104 obtains an image as shown in FIG. 21 (c) by applying a distance transform to the edge image as shown in FIG. 21 (b), and obtains the average value of the brightness values of the second directional segment It is possible to judge whether it is a closed rectangular parking zone or an open rectangular parking zone. This makes it possible to cope with a change in the illumination condition that can occur when only the brightness value is considered.

On the other hand, the parking section identifying unit 104 can determine the specified parking zone as a linear or stepped parking zone when the angle between the sides is about 45 degrees (or 135 degrees). The parking section identifying section 104 can measure the brightness of the second directional section line (the green straight line and the blue straight line in Fig. 21 (d)) and calculate the average. The parking section identification unit 104 measures the brightness of the green straight line and the blue straight line shown in FIG. 21 (d), and if the average value of the brightness for the green straight line is larger than the average value of the brightness for the blue straight line, In the opposite case, it can be judged as a stepped parking section.

Likewise, the parking section identifying section 104 may determine the type based on the edge of the second directional section after judging it to be a linear or stepped parking section. That is, the parking section identifying unit 104 obtains the image shown in FIG. 21 (f) by applying the distance conversion to the edge image as shown in FIG. 21 (e), and uses the average value of the brightness values of the second directional segment, Whether it is a parking block or a stepped parking block.

Figures 22-25 illustrate parking segments identified by type according to various embodiments of the present invention. Figures 22-25 illustrate the results of recognizing a closed rectangular parking compartment, an open rectangular parking compartment, a quadric parking compartment, and a stepped parking compartment, respectively.

The method of the present invention as described above can also be implemented by a computer program. And the code and code segments constituting the program can be easily deduced by a computer programmer in the field. In addition, the created program can be stored in a computer-readable recording medium (information storage medium), and can be read and executed by a computer to implement the method of the present invention. And the recording medium includes all types of recording media readable by a computer.

The specific acts described in the present invention are, by way of example, not intended to limit the scope of the invention in any way. For brevity of description, descriptions of conventional circuit configurations, control systems, software, and other functional aspects of the systems may be omitted. Also, the connections or connecting members of the lines between the components shown in the figures are illustrative of functional connections and / or physical or circuit connections, which may be replaced or additionally provided by a variety of functional connections, physical Connection, or circuit connections. Also, unless stated otherwise such as " essential ", " importantly ", etc., it may not be a necessary component for application of the present invention.

The use of the terms " above " and similar indication words in the specification of the present invention (particularly in the claims) may refer to both singular and plural. In addition, in the present invention, when a range is described, it includes the invention to which the individual values belonging to the above range are applied (unless there is contradiction thereto), and each individual value constituting the above range is described in the detailed description of the invention The same. The use of all examples or exemplary language (e.g., etc.) in this invention is for the purpose of describing the present invention only in detail and is not to be limited by the scope of the claims, It is not. It will also be appreciated by those skilled in the art that various modifications, combinations, and alterations may be made depending on design criteria and factors within the scope of the appended claims or equivalents thereof.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. The present invention is not limited to the drawings.

Claims (14)

A method for detecting a parking zone,
Obtaining an image around the vehicle;
Obtaining an edge image from the image according to a first orientation and a second orientation;
Extracting a plurality of straight lines from the edge images; And
And specifying a parking segment based on the plurality of straight lines. The method according to claim 1,
Wherein the first orientation and the second orientation have an angle difference of 180 degrees. The method according to claim 1,
Wherein the extracting of the plurality of straight lines is performed by applying a Hough Transform to the edge image. The method according to claim 1,
Wherein the step of specifying the parking section comprises: obtaining a first direction partition line from a pair of straight lines that are parallel to each other and spaced apart from each other by a predetermined width among the plurality of straight lines; And
And specifying the parking segment based on the first directional segment. The method of claim 4,
A plurality of the first directional segment lines are obtained,
Wherein the step of specifying the parking zone comprises: measuring brightness along the plurality of first directional lines;
Extracting a point at which the brightness changes abruptly;
Obtaining a second directional partition line including the point;
And specifying the parking section based on the plurality of first direction partition lines and the second direction partition line. The method of claim 5,
The step of extracting a point where the brightness changes abruptly may include extracting a point where a brightness difference value among a local minima of an intensity difference profile graph for the plurality of first directional partitioning lines ≪ / RTI > wherein the method further comprises the step of: The method of claim 5,
Wherein acquiring the second directional section line is a step of obtaining a straight line as the second directional section line as a result of applying a RANSAC (Random Access Consensus) algorithm to the point. An image acquiring unit for acquiring images of the surroundings of the vehicle;
An image preprocessing unit for acquiring an edge image from the image according to a first orientation and a second orientation;
A straight line extracting unit for extracting a plurality of straight lines from the edge images; And
And a parking section specifying section for specifying a parking section based on the plurality of straight lines. The method of claim 8,
Wherein the first orientation and the second orientation have an angle difference of 180 degrees. The method of claim 8,
Wherein the straight line extracting unit extracts the plurality of straight lines by applying a Hough transform to the edge image. The method of claim 8,
The parking section specifying section may acquire a first direction partition line from a pair of straight lines that are parallel to each other and spaced apart from each other by a predetermined width among the plurality of straight lines and acquire a first direction partition line that specifies the parking section based on the first direction partition line And a parking section detecting section for detecting the parking section. The method of claim 11,
A plurality of the first directional segment lines are obtained,
The parking space specifying section may further include a second direction partition line obtaining section for measuring brightness along the plurality of first direction partition lines and extracting a point where the brightness changes abruptly to obtain a second direction partition line including the point Including,
Wherein the parking section specifying section specifies the parking section based on the plurality of first direction partitioning lines and the second direction partitioning line. The method of claim 12,
Wherein the second directional segment obtaining unit extracts a point having a brightness difference value smaller than a predetermined value among a local minimum value of a brightness difference profile graph for the plurality of first directional segment lines. The method of claim 12,
Wherein the second directional segment obtaining unit obtains, as the second directional segment, a straight line obtained by applying a residual algorithm to the point.

Images