JPH07114699A - Method and device for detecting body - Google Patents

Abstract

PURPOSE:To detect only a part having height since the part is projected at a different position in a converted image by projecting and converting an image of one camera into an image of the other camera on the basis of specific coordinates. CONSTITUTION:Image signals picked up downward by the 1st television camera 11 and 2nd television camera 12 are inputted to an A/D conversion part 13 and converted into digital signals, which are stored in a 1st multi-valued image memory part 15 and a 2nd multi-valued image memory 16. A stationary body detection area setting part 14 previously set which area of the 1st multi-valued image memory part 15 and 2nd multi-valued image memory part 16 a stationary body is decided. A correspondence coordinate setting part 18 sets which position of a plane part in an image in the multi-valued image memory part 15 and multi-valued image memory part 16 is made to correspond. A projection conversion part 17 performs a projection converting process for the image in the multi-valued image memory part 15 by a conversion expression found from correspondence coordinates set by a correspondence coordinate setting part 18.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting an object using a television camera, an obstacle detecting device for detecting an automobile stopped due to a failure in a railroad crossing using the detection method, and a road. The present invention relates to an obstacle detection device that detects an obstacle above and a parked vehicle detection device that detects a parked vehicle in a parking lot.

[0002]

2. Description of the Related Art FIG. 8 shows the structure of a conventional stationary object detecting device of this kind. In FIG. 8, 81 is a television camera, 82 is an A / D conversion unit, 83 is a multi-valued image memory unit, 8
Reference numeral 4 is a stationary object detection area setting unit, 85 is a background data creation unit, 86 is a data comparison unit, and 87 is a stationary object detection unit.

Next, the operation of the above conventional example will be described. In FIG. 8, the TV camera 81 is installed so as to take an overhead image of the road surface, and the condition of the road surface imaged by the TV camera 81 becomes a two-dimensional imaging condition, which the A / D conversion unit 82 indicates. It is input, converted into a digital signal, and then stored in the multi-valued image memory unit 83.

The stationary object detection area setting unit 84 presets in which area of the multi-valued image memory unit 83 the presence or absence of a stationary object is determined. In the background data creation unit 85,
Of the image data stored in the multi-valued image memory unit 83,
The state in which the object does not exist is selected and the image data is stored.

In the data comparison section 86, the multi-valued image memory section 8 preset in the stationary object detection area setting section 84 is set.
The data in the area 3 is compared with the background data stored in the background data creation unit 85, and when the input data and the background data are different, the information is output to the stationary object detection unit 87. In the stationary object detection unit 87, the data comparison unit 86
If the output from the device is continuous for a certain period of time, it is determined that there is a stationary object in the detection area.

As described above, in the conventional stationary object detecting device, the presence or absence of a stationary object in the detection area is determined by comparing the background data when there is no object with the image data that is sequentially input. Is possible.

[0007]

However, in the above-mentioned conventional stationary object detecting device, even if the background data when no object is present is compared with the image data that is sequentially input, the brightness is compared. Therefore, for example, when trying to detect a stationary object on the road, there are shadows such as buildings and street trees on the side of the road in the detection area. Sometimes it was judged as a stationary object even though it was part of.

Even if such a thing does not exist on the side of the road, the brightness still changes depending on the situation such as the sun hiding in the cloud or the shadow of the cloud reflected on the road surface. I have something to do.

Further, when the background data is created, it is necessary to select a case where no object exists, so that there is a problem that it is difficult to create the background data on a road with a high traffic volume.

The present invention solves such a conventional problem, and is an excellent object detecting method for detecting only a static object having a really high height in a detection area, and an inside of a railroad crossing using this detecting method. An object of the present invention is to provide a detection device for detecting an obstacle, an obstacle on a road, and a parked vehicle.

[0011]

In order to achieve the above object, the object detecting method according to the invention of claim 1 is such that the first television camera and the second television camera installed at a high place so that the ground can be imaged overhead. An image of the ground surface is taken by a TV camera, the images taken by the cameras are stored in an image memory, and the coordinates of points corresponding to the ground portion in the image of each camera are selected in advance, and the correspondence is set. First based on the coordinates of the point
Geometrically transforms the image of the television camera of the second television camera into the image of the second television camera, and compares the image of the second television camera with the image of the second television camera which has been transformed, and detects a portion protruding above the ground. I made it up.

The obstacle detecting device in the railroad crossing according to the invention of claim 2 is installed near the railroad crossing and has a first television camera and a second television camera for taking a bird's-eye view of the situation inside the railroad crossing, and each of the cameras. Multi-valued image memory unit that stores the image from, the stationary object detection area setting unit that sets the area for detecting a stationary object in the railroad crossing, and the portion of the image that corresponds to the plane in the railroad crossing in the image from each camera. A coordinate setting unit that sets coordinates; a geometric transformation processing unit that geometrically transforms an image of the first television camera into an image of the second television camera based on the set coordinates; A stationary object detection unit that compares the image with the converted image of the first TV camera to determine the presence or absence of a stationary object in the railroad crossing, and a circuit breaker state detection unit that detects the up and down state of the railroad crossing circuit breaker rod. And the output of this circuit breaker status detector No. from the output signal of the still object detecting unit, and the configuration including the alarm output unit for outputting an alarm when the blocking rod is stationary in railroad crossing is detected have descended.

An on-road obstacle detection device according to a third aspect of the present invention is installed on the side of a road, and a first television camera and a second television camera for taking a bird's-eye view of a situation on the road, and images from the cameras. A multi-valued image memory unit, an obstacle detection region setting unit that sets a region for detecting an obstacle on the road, and whether or not the white line drawn on the road surface in the image from each camera. A coordinate setting unit that sets the coordinates of the portion, a geometric transformation processing unit that geometrically transforms the image of the first television camera into the image of the second television camera based on the set coordinates, and the second television camera. Image and the converted image of the first television camera are compared to determine whether there is an obstacle on the road, and that there is an obstacle on the road from the output signal of the obstacle detection unit. An alarm that outputs an alarm when judged And a configuration with a force unit.

According to another aspect of the present invention, there is provided a parked vehicle detection device, which is installed at a high place in a parking lot and stores a first television camera and a television camera for taking a bird's-eye view of the situation in the parking lot, and images from the cameras. A multi-valued image memory unit, a parked vehicle detection region setting unit that sets a region for detecting a parked vehicle in each parking lot, and parking lots drawn on the road surface in the images from the cameras. A coordinate setting unit that sets the coordinates of the corner of the white line that delimits the line, and a geometric conversion processing unit that geometrically converts the image of the first TV camera into the image of the second TV camera based on the set coordinates. From the output signal of the parked vehicle detection unit that determines the presence or absence of a parked vehicle in each parking lot by comparing the image of the second television camera and the converted image of the first television camera, Number of parked vehicles on the screen And a configuration with a parked vehicle number counting unit for counting.

[0015]

According to the first aspect of the present invention, the corresponding coordinates of the flat plane portion in the detection area in the image of each television camera is determined in advance, and the image of one camera is obtained based on this coordinate. By projecting and converting to the image of the other camera, the part without height is projected at the same position in the image of the other camera and the converted image, but for the part with height such as a stationary object, Since the converted image is projected at different positions, it is possible to detect only a portion having a height by comparing the image of one camera with the converted image. Further, since the stationary object is detected only from the images captured by the two cameras at the same time, it is not necessary to create or update the background image data, and the stationary object can be detected immediately after the processing is started.

According to the second aspect of the present invention, it is possible to reliably detect a broken vehicle or the like that is stopped inside the railroad crossing.

According to the third aspect of the invention, the obstacle on the road can be detected without catching the flow of the vehicle on the road.

According to the fourth aspect of the invention, it is possible to detect a vehicle that is parked in a parking lot without catching a change in the entry and exit of the vehicle into and from the parking lot.

[0019]

【Example】

(Embodiment 1) A first embodiment of the present invention will be described with reference to FIGS.

FIG. 1 is a block diagram showing the configuration of the first embodiment to which the method of the present invention is applied. In FIG. 1, 11
Is a first television camera, 12 is a second television camera, and the first and second television cameras 11 and 12 are installed at a high place where a bird's-eye view of the ground can be taken for a predetermined interval. Thirteen
Is an A / D converter that converts the images captured by the television cameras 11 and 12 into digital signals (multi-valued).
The / D conversion unit 13 includes a first multi-valued image memory unit 15 for storing multi-valued image data from the television camera 11 and a second multi-valued image data unit for storing the multi-valued image data from the television camera 12.
The multi-valued image memory units 16 are connected to each other.

Reference numeral 14 denotes a stationary object detection area setting unit of the multivalued image memory units 15 and 16, 17 is a projection conversion unit connected to the multivalued image memory unit 15, 18 is a corresponding coordinate setting unit of the projection conversion unit 17, Reference numeral 19 is a comparison processing unit that compares the image output from the multi-valued image memory unit 16 and the image output from the projection conversion unit 17, and 20 is a stationary object detection unit that detects a stationary object from the comparison processing result.

Next, the operation of the first embodiment will be described. In the first embodiment, the respective image signals captured from the bird's-eye view by the first TV camera 11 and the second TV camera 12 are input to the A / D conversion unit 13 and converted into digital signals. It is stored in one multi-valued image memory unit 15 and second multi-valued image memory unit 16, respectively. The stationary object detection area setting unit 14 presets in which area of the first multi-valued image memory unit 15 and the second multi-valued image memory unit 16 the stationary object is determined.

FIG. 2 shows an example of setting a stationary object detection area. The corresponding coordinate setting unit 18 sets which position of the plane portion in the image to correspond to among the multivalued image memory unit 15 and the multivalued image memory unit 16. The projection conversion unit 17 performs a projection conversion process on the image in the multi-valued image memory unit 15 using a conversion formula obtained from the corresponding coordinates set by the corresponding coordinate setting unit 18.

Next, the projection conversion process will be described with reference to FIG. The coordinates (X, Y) on the image of the television camera 11 stored in the multi-valued image memory unit 15 are the coordinates (x, y) on the image of the television camera 12 stored in the multi-valued image memory unit 16. Shall correspond to. in this case,
The conversion formula for projecting and converting the image of the TV camera 11 into the image of the TV camera 12 is as follows.

X = ax + by + c (1) Y = dx + ey + f (2) Coefficients a, b, c, d, e, and f are the corresponding coordinate values on the images of the television camera 11 and the television camera 12 expressed by the above equation ( It is obtained by solving simultaneous equations obtained by substituting 1) and (2), and it is necessary to determine corresponding coordinate values of at least three points or more. In FIG. 2, points P, Q, R, S, T, U on the image of the TV camera 11 are shown.
Are points p, q, r, s, on the image of the TV camera 12.
It corresponds to t and u.

Next, a method of creating a projection conversion image will be described with reference to FIGS. 2 and 3. FIG. 2A is an example of an image of the television camera 11, and FIG. 2B is an example of an image of the television camera 12. Further, FIG. 3 shows a state in which the image of the television camera 11 is projected and converted.

As described above, in the projection-converted image, a portion of the original image having no height component is converted as it is, but a portion having a height component is converted to a slightly inclined shape. .

In the comparison processing unit 19, the multivalued image memory unit 1
5 and the image of the projection conversion unit 17, the comparison processing is performed on the area set by the stationary object detection area setting unit 14. Hereinafter, the comparison process will be described with reference to FIGS.

As shown in FIG. 3, the projection-converted image is an image that is inclined at a portion having a height. Therefore, simply take the difference between FIG. 2B and FIG. This part can be detected by setting a threshold value.
However, in reality, the brightness of an image captured by each TV camera is completely equal to the brightness of an image captured by a TV camera of the same model, even if the TV cameras are of the same model. That is impossible. Therefore, if the difference in image brightness between the two TV cameras is greater than or equal to the threshold value of the difference image, that portion may be detected even if it is not a portion having a height component. Therefore, as shown in FIGS. 4 (a) and 4 (b),
It is equally divided into rectangular block areas having the same number of pixels, and the distance d is defined as in Expression 1 for the corresponding blocks between both images.

[0030]

[Equation 1]

The threshold value of the distance d is set in advance, and blocks having a distance equal to or larger than the threshold value are assumed to have a height component. By doing so, the difference in image brightness between the two cameras does not affect the processing.

In the stationary object detection unit 20, the comparison processing unit 19
From the result of, it is determined whether or not it is a stationary object.

As described above, according to the first embodiment,
This has an effect that only a portion having a height can be detected immediately after the start of processing.

(Embodiment 2) FIG. 5 is a block diagram showing the configuration of the second embodiment of the present invention.

In the second embodiment, the present invention is applied to an obstacle detecting device inside a railroad crossing, and the structure shown in the first embodiment has the same structure as that of the breaker state detecting section for detecting the state of the breaker. 51 and an alarm output unit 52 which is operated by the output signal of the circuit breaker state detection unit 51 and the stationary object detection signal of the stationary object detection unit 20 are added so that a vehicle stopped due to a failure or the like in the railroad crossing can be detected. It was done.

The television cameras 11 and 12 in the second embodiment are installed near the railroad crossing.

Next, the operation of the second embodiment will be described. In the second embodiment, the circuit breaker state detection unit 51 outputs the up / down state of the blocking rod to the alarm output unit 52 in response to the circuit breaker control signal 53. The alarm output unit 52 outputs an output when the stationary object detection unit 20 detects a stationary object in the railroad crossing, and an alarm output 54 when the breaker state detection unit 51 detects that the blocking rod has descended.

As described above, according to the second embodiment, it is possible to surely detect a broken vehicle or the like that is stopped inside the railroad crossing.

(Embodiment 3) FIG. 6 is a block diagram showing the configuration of the third embodiment of the present invention.

In the third embodiment, an alarm output unit 61 which operates by the output from the stationary object detection unit 20 is added to the structure shown in the first embodiment to detect an obstacle dropped on the road. It was made possible.

The television cameras 11 and 12 according to the third embodiment are installed on a roadside lighting pole or the like.

Next, the third embodiment will be described.
In the third embodiment, the alarm output unit 61 operates by an output signal when the stationary object detection unit 20 detects an obstacle on the road that causes a traffic obstacle, and outputs an alarm output 62.

As described above, according to the third embodiment,
Obstacles on the road can be detected without catching changes in the flow of vehicles on the road.

(Fourth Embodiment) FIG. 7 is a block diagram showing the configuration of the fourth embodiment of the present invention.

In the fourth embodiment, the stationary object detecting section 20 having the structure shown in the first embodiment is replaced by the parked vehicle detecting section 71.
In addition, a parked vehicle number unit 72 is further provided to configure a parked vehicle detection device that counts the number of vehicles parked in the parking lot in the parking lot.

The television cameras 11 and 12 in the fourth embodiment are installed at a predetermined high place in the parking lot.

Next, the fourth embodiment will be described.
In the fourth embodiment, the parked vehicle detection unit 71 detects a vehicle being parked in a parking lot, and the result of the detection
The parked vehicle number counting unit 72 in the next stage counts the number of parked vehicles in the screen.

As described above, according to the fourth embodiment,
It is possible to detect a vehicle that is parked in a parking lot without capturing a change in the entry and exit of the vehicle into and from the parking lot.

[0049]

As is apparent from the above embodiment, the present invention uses two television cameras to project and convert the image of one camera into the image of the other camera, thereby forming a flat portion. Is projected at the same position, but parts with height are projected at different positions, so it is possible to avoid false detection of objects due to changes in brightness on the plane. It is possible to reliably detect only a tall object.

Further, in the present invention, since it is possible to detect only the protruding portion, the obstacle detecting device in the railroad crossing,
It can be applied to an obstacle detection device on a road and a parked vehicle number counting device.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of a stationary object detection method according to a first embodiment of the present invention.

FIG. 2A is a diagram showing an example of setting a stationary object detection area and corresponding coordinates of an image of a first TV camera in this embodiment. FIG. 2B is a still object of an image of a second TV camera in this embodiment. The figure which shows an example of setting of a detection area and corresponding coordinates

FIG. 3 is a diagram showing an example of a projection conversion screen of the first television camera in this embodiment.

FIG. 4A is a diagram showing a method of comparison processing of projection conversion screens of the first television camera in the present embodiment. FIG. 4B is a method of comparison processing of projection conversion screens of the second television camera in the present embodiment. Figure

FIG. 5 is a schematic block diagram of an obstacle detection device inside a railroad crossing according to a second embodiment of the present invention.

FIG. 6 is a schematic block diagram of a road obstacle detection device according to a third embodiment of the present invention.

FIG. 7 is a schematic block diagram of a parked vehicle detection device according to a fourth embodiment of the present invention.

FIG. 8 is a schematic block diagram of a conventional stationary object detection method.

[Explanation of symbols]

 11 First TV Camera 12 Second TV Camera 13 A / D Converter 14 Still Object Detection Area Setting Unit 15 First Multi-Valued Image Memory Unit 16 Second Multi-Valued Image Memory Unit 17 Projection Conversion Unit 18 Corresponding Coordinates Setting unit 19 Comparison processing unit 20 Stationary object detection unit 51 Circuit breaker state detection unit 52 Warning output unit 53 Circuit breaker control signal unit 54 Warning output 61 Warning output unit 62 Warning output 71 Parked vehicle detection unit 72 Parked vehicle number counting unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location G08G 1/00 K 7531-3H 1/09 R 7531-3H H04N 7/18 D

Claims (4)

[Claims] 1. A ground condition is imaged by a first TV camera and a second TV camera installed at a high place so that the ground surface can be imaged in a bird's-eye view, and the images taken by the respective cameras are stored in an image memory. , The coordinates of points corresponding to the ground portion in the images of the respective cameras are selected in advance, and the image of the first television camera is converted into the image of the second television camera based on the coordinates of the corresponding points. The object detection method of geometrically transforming into a second image and comparing the image of the second television camera with the transformed image of the second television camera to detect a portion protruding above the ground. 2. A first television camera and a second television camera, which are installed in the vicinity of a railroad crossing and take a bird's-eye view of the situation inside the railroad crossing, and a multi-valued image memory unit for storing images from the cameras. A stationary object detection area setting unit for setting an area for detecting a stationary object in the railroad crossing, and a coordinate setting unit for setting coordinates of a portion corresponding to a plane in the railroad crossing in the images from the cameras, A geometric transformation processing unit that geometrically transforms the image of the first television camera into the image of the second television camera based on the coordinates, and the image of the first television camera converted from the image of the second television camera. The stationary object detection unit that determines the presence or absence of stationary objects in the railroad crossing, the circuit breaker state detection unit that detects the up and down state of the barrier breaker blocking rod of the railroad crossing, and the output signal of this circuit breaker state detection unit. From the output signal of the stationary object detection unit, An obstacle detection device inside a railroad crossing equipped with an alarm output unit that outputs an alarm when a stationary object is detected inside the railroad crossing when the blocking rod is down. 3. A first television camera and a second television camera installed on the side of the road for taking a bird's-eye view of the situation on the road, a multi-valued image memory unit for storing the images from the cameras, and the on-road An obstacle detection area setting unit that sets an area for detecting an obstacle, and a coordinate setting unit that sets the coordinates of the corner portion of the white line drawn on the road surface in the image from each camera, A geometric transformation processing unit that geometrically transforms the image of the first television camera into the image of the second television camera based on the set coordinates, and the first television camera transformed with the image of the second television camera. An obstacle detection unit that compares images to determine whether there is an obstacle on the road, and an alarm output unit that outputs an alarm when it is determined that there is an obstacle on the road from the output signal of the obstacle detection unit. Road obstacle detection device equipped with. 4. A first TV camera and a TV camera, which are installed at a high place in a parking lot and take a bird's-eye view of the situation in the parking lot, a multi-valued image memory unit for storing images from the cameras, and each parking. A parking vehicle detection area setting unit that sets an area for detecting a parked vehicle in the box, and sets the coordinates of the corner of the white line that divides each parking shuttle drawn on the road surface in the image from each camera. A coordinate setting unit for performing geometric transformation of the image of the first television camera into an image of the second television camera based on the set coordinates;
From the output signal of the parked vehicle detection unit that compares the image of the second television camera with the converted image of the first television camera to determine the presence or absence of the parked vehicle in each parking yard, and the output signal of the parked vehicle detection unit. A parked vehicle detection device including a parked vehicle number counting unit that counts the number of parked vehicles on the screen.