JPH041900A - Vehicle monitor method and device - Google Patents

Abstract

PURPOSE:To correctly judge the presence/absence of vehicles in a parking area defined for vehicles by checking the number of edges intersecting with white lines remote from a viewpoint among a group of white lines constituting the parking area in a picture of a parking lot and judging the presence/absence of vehicles in the parking area projected on the picture of the parking lot. CONSTITUTION:In a parking area 111 where a vehicle 107 is parked, a white line 113 among a group of white lines surrounding the parking area 111, which is located at a position remote from a viewpoint is hidden by a contour 112 forming the vehicle 107 as a borderline. In a parking are 108 where vehicle is not parked, a white line 110 among a group of white lines surrounding the parking area 108, which is located at a position remote from the viewpoint is not hidden by the vehicle 107 but seen. When the presence/absence of a vehicle in the parking area 111 is decided, a picture applied to a picture 106 of a parking lot is binarized so that a group of edges 114 are detected. Then, when the contour 112 is detected. It is judged that the parking area 111 has a vehicle. Thus, the presence/absence of a vehicle in each parking area can be correctly judged.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a vehicle monitoring technique that is necessary when notifying a driver of the position of an empty parking lot in a parking lot or the like.

[Conventional technology]

In order to process parking lot images and determine the presence or absence of cars in each parking area, the literature [Makoto Nagao (author): Image Epistemology, Corona Publishing,
P, 105. Two images f (x, y) described in [First edition: February 15, 1988].

a (x, y) difference S S = Σl f (x, y) - a (x, y)
! There is a technology that utilizes (1).

For example, FIG. 2(B) shows an image 103 captured by a TV camera, for example, of a scene in which a car is in a parking area 101 separated by a white line as shown in FIG. 2(A).

In order to process the image 103 and discover the visible car 104, in the prior art, as a first step, the parking area 101 is
A mask image (m (x, y)) 105 shown in FIG. 2(C) is created in advance, with a value of 1 within the range and 0 elsewhere. In the second step, when the image when there is no car in the parking area 101 is fo(x, y), the model image a (X.

y) is synthesized in advance.

a (x, y) = fo (x, y) - m (x, y) (2
) As the third step, when the image 103 is expressed as i (x, y), the parking area image f of the parking area 101 is expressed by the following formula.
Combine (x, y).

f (x, y) = i (x, y) ・m (x, y)
(3) As the fourth step, a (x,
y) and f (x, y) is calculated. In the fifth step, for example, if the difference S is larger than the preset darkness value S0, it is determined that there is a car in the parking area 101.

[Problem to be solved by the invention]

However, using the prior art, for example, FIG.
For example, a parking area 1080 that is largely hidden by cars 107, shown in the parking lot image 106 illustrated in A).
When determining the presence or absence of a car, for example, the parking area image synthesized by formula (3) using the mask image 109 illustrated in FIG. 3(B) representing the parking area 108 synthesized in the same manner as before, Since a part of the image is captured, the difference S calculated in the same way as before is the threshold value S. becomes larger than As a result, there was a problem in that it was incorrectly determined that there was a car in the parking area 108.

SUMMARY OF THE INVENTION An object of the present invention is to provide a vehicle monitoring method and device that correctly determines the presence or absence of a vehicle in a hidden parking area.

[Means to solve the problem]

The vehicle monitoring method of the present invention examines the number of edges that intersect with white lines that are far from the viewpoint among the group of white lines constituting the parking area in the image of the parking lot, and calculates the number of edges that intersect with the white lines that are far away from the viewpoint. It is characterized by determining the presence or absence of.

The vehicle monitoring device of the present invention comprises: a far white line storage means for storing the position of a white line far from a viewpoint among a group of white lines constituting a parking area; and a white line stored by the far white line storage means from an image of the parking lot. A step/edge storage means for storing edges calculated by the edge/edge calculation means for calculating non-parallel edges, and an intersection between the edge stored by the edge storage means and the white line stored by the distant white line storage means. and an intersection detection means for determining the presence or absence of a car in the parking area shown in the image of the parking lot.

[Effect]

For example, the present invention determines that there is a car in this parking area when one of the white lines that is located far from the viewpoint is hidden among a group of white lines surrounding a parking area that was determined to have a car in the prior art. However, when one of the white lines is not hidden, it is determined that there are no cars in this parking area.

For example, in the case of the parking area 111 where the car 107 illustrated in FIG. 3(A) is parked, this parking area 111
White line 11 located far from the viewpoint among the group of white lines surrounding
3 is hidden with the outline 112 forming the car 107 as the boundary. On the other hand, in the case of the parking area 108 where no cars are parked, as shown in FIG. There is. Therefore, for example, when determining the presence or absence of a car in the parking area 111, in order to find edges that are not parallel to the white lines 110 and 113 from the parking lot image 106, a differential mask d( x, y; θ) (x2+y") = xexp ' By binarizing the image obtained by applying cos θ to the image 106 of the parking lot, the edge group 114 illustrated in FIG. 3(C) is detected. Next, , the car 10 illustrated in FIG. 3(D) among the edge group 114
By checking the edges that intersect with the white line 115 corresponding to the white line 113 when there is no 7, for example, the contour line 112 is discovered, and as a result, it is determined that there is a car in the parking area 111.

When similar processing is performed on the white line 110, there is no edge that intersects with the white line 110 among the edge group 114, so it is assumed that there is no contour line that hides the white line 110, and as a result, it is determined that there is no car on the parking edge 108. judge.

[Example] Examples of the present invention will be described in detail below.

FIG. 1 is a professional diagram showing one embodiment of the vehicle monitoring device of the present invention.

This vehicle monitoring device includes an image storage unit 1 that stores an image of a parking lot inputted by a TV camera, etc., and a distant white line storage unit that stores the position of a white line that is far from a viewpoint among a group of white lines that make up a parking area. 4, an edge calculation means 2 which calculates edges that are not parallel to the white line stored in the far white line storage means 4 from the image of the parking lot, an edge storage means 3 which stores the calculated edges, and this edge storage means stores the edges. Intersection detection means 5 for determining the intersection of edges and white lines stored in distant white line storage means 4; intersection table storage means 6 for storing an intersection table to be described later; display means 7; edge calculation means 2. The intersection detection means 59 includes a control means 8 for controlling the display means 7. The control means 8 can be realized by a memory and a microprocessor, the image storage means 1, the edge storage means 3, the distant white line storage means 4, and the intersection table storage means 6 can be realized by a memory, and the display means 7 can be realized by a CRT and a memory.

With the vehicle monitoring device having the above configuration, for example, as shown in FIG. 3(A),
The case of determining the presence or absence of cars in the parking areas 111 and 108 illustrated in FIG. 1 will be described in detail.

In this case, the image storage means 1 stores an image 106 of a parking lot illustrated in FIG. 3(A) inputted with a TV camera or the like. The intersection table storage means 6 stores an intersection table in which the edge number column is reset to 0 among the intersection tables 120 in the format illustrated in FIG. 4(A). The far white line storage means 4 is, for example, T
In an image captured by a V camera when there are no cars in parking areas 111 and 108, the pixel group through which the white line 110 and the pixel group through which the white line 115 pass, illustrated in FIG. 3(D), are filled in with, for example, l and 2, respectively. , stores a distant white line image in which 0 is substituted for other pixel values. FIG. 4B shows this distant white line image 116, where 117 indicates a pixel group through which the white line 110 passes, and 118 indicates a pixel group through which the white line 115 passes. The far white line storage means 4 further stores directions θ1. θ
2 is stored in the format illustrated in FIG. 4(C).

Now, the operation of the vehicle monitoring device shown in FIG. 1 begins when the control means 8 starts the edge calculation means 2. As a first step, the activated edge calculation means 2 calculates the direction θ3 of the pixel group stored in the far white line storage means 4. Average value of θ2 = (θ1+θ2)
Calculate (5).

When the image 106 of the parking lot stored in the image storage means 1 as the second stage is expressed as f (x, y), f (x, y)
and a differential mask d(x, yl?) in which the average value j shown in equation (5) is substituted for the θΦ value in equation (4).
The following intensity image a(x, y; #
) is calculated.

a (x, y; 1V) -15S f (x-u, y-v)・d (u,
v ; #) dudv l (7) In the third step, the intensity image a(x+yes) is binarized using, for example, an appropriately determined threshold T to calculate a binary image illustrated in FIG. 3(C).

As a fourth step, an edge image 118 in which, for example, the numbers illustrated in FIG. 4(D) are assigned to the pixel groups through which each edge included in the binary image passes, using a labeling technique that can be easily realized with conventional technology. is stored in the edge storage means 3, and the process is terminated.

When the edge calculation means 2 completes its processing, the control means 8 activates the intersection detection means 5. As a first step, the activated intersection detection means 5 represents, for example, the far white line image stored in the far white line storage means 4 as w(x, y), and stores it in the edge storage means 3.
When the edge image 118 stored by is expressed as e (x, y), each pixel value of w (x, y) and e (x, y) is, for example, the X direction is the main scanning direction, and the X direction is the sub scanning direction. For example, the position (x+, )'+)w (x+,
y+) 〉O and e (X+, )'+) ＞0
Find (9) and obtain each pixel value. In the second step, in the cross table 120 stored in the cross table storage means 6, 1 is added to the number of edges of the parking area 111 found as the second row pointed to by the pixel value, and then the pixel value is added to the first edge term. The value 7 is stored.

In the third step, by continuing the same process as described in the first step, the position O1z, )'z) and the pixel value of the pixel P2 illustrated in FIG. 4(E) are obtained. In the fourth step, the edge number that is the same as the pixel value L8 is already in the parking area 11.
If it is not stored in the row 1, the process is performed in the same manner as described in the second step, and 1 is added to the number of edges related to the parking area 111 in the intersection table 120, and then the pixel value is added to the second edge term. , the value 8 is stored, and the above process is repeated until all pixels have been scanned, and then the process is terminated. At the time when the intersection detection means 5 completes the processing, the intersection table 120 has stored the intersection table 121 illustrated in FIG. 4(F).

When the intersection detection means 5 completes its processing, the control means 8 activates the display means 7. The activated display means 7 scans each row of the cross table 121 stored in the cross table storage means 6, and for example, since the number of edges of the parking area 108 represented by the first row is 0, there is no car in the parking area 108. For example, it outputs the character string "Parking area 108...No car" on the CRT, and since the edge number 2 of the parking area 111 represented by the second line is 1 or more, there is no car in the parking area 111. For example, the character string "Parking area 111.
...car available" is output, and all processing ends.

〔Effect of the invention〕

According to the present invention, there is an effect that the presence or absence of a car in each parking area can be accurately determined by processing an image of a parking lot taken from a line-of-sight direction where a portion is hidden by a car in front.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the vehicle monitoring device of the present invention, FIG. 2 is a diagram for explaining the prior art, FIG. 3 is a diagram for explaining the present invention in detail, and FIG. FIG. 2 is a diagram for explaining the operation of the embodiment. 2...Edge calculation means 3...Edge storage means 4...Distant white line storage means 5...Cross detection means 106...Parking lot image 118...・Detected edge image 121 ・
...calculated cross table

Claims (2)

[Claims] (1) In an image of a parking lot, of the group of white lines that make up the parking area, check the number of edges that intersect with white lines that are far away from the viewpoint, and determine whether there are cars in the parking area shown in the image of the parking lot. A vehicle monitoring method characterized by: (2) Distant white line storage means for storing the position of a white line that is far from a viewpoint among a group of white lines constituting a parking area, and calculating edges that are not parallel to the white lines stored in the far white line storage means from an image of the parking lot. an edge calculating means for storing the calculated edge, an edge storing means for storing the calculated edge, and an intersection detecting means for determining the intersection of the edge stored in the edge storing means and the white line stored in the distant white line storing means, A vehicle monitoring device characterized by determining the presence or absence of a car in a parking area shown in an image of a parking lot.