KR20200025384A - Parking Information Providing Device Using Real-Time Image Processing and System thereof - Google Patents

Abstract

According to one embodiment of the present invention, provided is a parking information providing device utilizing real-time image processing, which comprises: an image analysis unit analyzing a parking surface image based on an image photographed by at least one camera; a parking information detection unit determining whether parking is available using data analyzed by the image analysis unit; and a database storing parking availability data determined by the parking information detection unit.

Description

Parking Information Providing Device Using Real-Time Image Processing and System

The present invention provides a parking information providing apparatus and system using real-time image processing, and more specifically, real-time image processing that provides parking information by recognizing individual parking zones using a camera or an arbitrary sensor of a parking lot and determining whether parking is performed. The present invention relates to a parking information providing apparatus utilized and a system thereof.

In modern society, many people use their cars as a means of transportation due to the increase of vehicles, which is increasing the number of vehicles to be parked stably in the parking lot until the work is completed.

Accordingly, in places such as hospitals, airports, department stores, shopping malls, government offices, large buildings, or parking buildings where many vehicles are accessed, the parking lot is enlarged for parking of many vehicles, and in order to efficiently manage the large parking lot. Parking fees are collected on vehicles to control the entry and exit of vehicles.

That is, the vehicle enters through the entrance to open the breaker when the driver takes the parking ticket to collect the fare so that the vehicle can enter the inside, and when the parked vehicle exits the exit, the fare is appropriate according to the recorded entry time. After the payment is made, the breaker is opened to allow the vehicle to exit the parking lot.

However, in such a conventional parking management system, since the driver must find and park the empty parking surface by himself while the vehicle enters the inside of the parking lot, it takes a lot of time to park and also to find the empty parking surface for parking the vehicle. There is a problem that a lot of accidents occur such as air pollution, time consumption and contact accident of the vehicle in a narrow space by constantly driving the inside of the.

In order to solve this problem, some large parking lots have a small number of parking personnel assigned to share the parking space information between the parking personnel through the radio terminal, but the vehicle entered is directed to the empty parking surface. Considering this, there is a limit to managing a large parking lot by a small number of parking personnel. Therefore, even when the parking personnel are arranged, the above-mentioned problems are solved to some extent, but labor costs are incurred. Is not very efficient.

Therefore, a large parking lot desperately needs a parking guidance system that can determine whether a vehicle is parked on the corresponding parking surface at the central management office (or control room) or at the checkout station at the exit. Recently, research and development have been actively conducted in various ways such as ultrasonic sensor method, RFID (Radio Frequency IDentification) method, parking confirmation equipment method and camera method.

The roof coil method is to embed the roof coil on the concrete floor of the parking surface and then detect the change in the magnetic field that changes depending on whether the vehicle enters or not. Based on the detected information, the manager determines whether the vehicle is on the parking surface. . However, this roof coil method has a disadvantage in that construction is cumbersome and the construction cost is very high, such as embedding the roof coil and recoating the surface after cutting concrete in the case of the existing parking lot. These drawbacks have made the roof coil method a limiting factor for intelligent parking management.

The ultrasonic sensor method attaches an ultrasonic sensor to the ceiling of each parking surface, scans the ultrasonic wave from the ceiling of each parking surface to the floor, and detects the reflected ultrasonic waves, so that the administrator can control the parking surface based on the detected information through the control unit. This is a method of determining the presence of a vehicle. However, since the ultrasonic sensor method is a method in which one ultrasonic sensor checks and manages one parking surface, if there are 100 parking surfaces, the installation cost increases because 100 ultrasonic sensors must be installed. As the probability increases, it is impossible to manage efficiently, and the maintenance cost increases. Also, the installation of a lot of equipment on the parking surface increases the discomfort on the parking surface, and it is impossible to effectively cope with the changing of the parking surface. Furthermore, it is very difficult to apply in outdoor parking lot without the ceiling.

The RFID method is similar to the ultrasonic sensor method, but has an advantage of a long distance that RFID can recognize according to a frequency band. Therefore, up to five parking surfaces can be managed with one RFID, compared to the ultrasonic sensor method. However, since the RFID method must be attached to the vehicle in order to be recognized as RFID, there is no way to recognize the vehicle without the RFID tag. In addition, it is impossible to cope with the change of parking lot, and it is very difficult to apply in outdoor parking lot.

The camera method has a camera and a marker on each parking surface. In other words, it is difficult to accurately determine whether the parking and the tolerance persist by using the image taken by the camera. (See Patent Publication No. 10-0625201: Parking surface management system using a camera). In addition, the controller determines whether the marker is photographed by the controller to determine whether parking is performed, and recognizes the license plate in the image captured by the camera through the analyzer. However, such a camera method installs cameras and signs on each parking surface. Therefore, if there are 100 parking surfaces, the installation cost increases because 100 cameras and markers need to be installed. There is a disadvantage that it is impossible to efficiently manage and increase the maintenance cost. Also, it is impossible to effectively cope with changing the parking surface, and it is very difficult to apply in outdoor parking lot.

An object of the present invention is to provide a parking information of the individual parking surface by detecting a change in the average color value photographed by the camera.

Another object of the present invention is to provide parking information by extracting the contour of the individual parking surface and quantifying a certain change in the area of the contour.

Another object of the present invention is to provide parking information in a specific color to the interface for outputting the parking surface.

Parking information providing apparatus using real-time image processing according to an embodiment of the present invention comprises an image analyzer for analyzing a parking surface image based on the image taken by at least one camera; A parking information detector determining whether parking is possible using the data analyzed by the image analyzer; And a database for storing parking possibility presence data determined by the parking information detector.

The image analyzer may calculate, for example, an average value or a median value for a color value of the entire inner surface of the parking surface or each of a plurality of predetermined points of the inner surface, and the parking information detector may calculate the image value. It is possible to detect whether parking is possible by determining whether a representative value is out of a predetermined error range.

The image analyzer may recognize at least one display unit displayed on the inner surface of the parking surface, and the parking information detector may not recognize any one or more display units of one or more display units displayed on the parking surface. If not, parking can be determined.

In addition, the image analysis unit performs analysis initialization before analyzing the parking surface image to reset the area value of the boundary surface included in the parking surface and the area value of the inner surface, and the parking information detection unit resets the area of the boundary surface reset by the image analyzer. On the basis of the value and the area value of the inner surface, it is possible to detect whether parking is possible by determining whether it is out of a predetermined error area range.

In addition, the parking information providing apparatus further includes an interface for outputting a parking surface, wherein the interface, the parking surface is not occupied among the parking surface parking available parking surface is the first color, the parking surface is occupied among the parking surface is second parking The parking surface for the disabled can be displayed in the third color among the color and the parking available parking surface. A parking information providing system using real-time image processing according to an embodiment of the present invention includes a plurality of parking information providing apparatuses of claim 1. field; At least one user device interworked with the parking information providing apparatuses in a network; And at least one service providing server interworking with the parking information providing apparatuses in a network, wherein the service providing server has a lot of empty parking spaces among the parking surface information provided to each of the parking information providing apparatuses. Sorting the parking information providing apparatuses, and sequentially performing exposure or alarm on the empty parking space to the at least one user device according to the sorted order, and having payment for a predetermined fee for the exposure or alarm. You can only do this.

The present invention detects a change such as an average value photographed by a camera photographing a parking surface, extracts the contour of an individual parking surface, quantifies a constant change in the area of the contour, and uses the data value or a constant place inside the individual parking surface. By displaying a specific mark and detecting whether or not the specific mark is confirmed by the parking status, the parking of the individual parking surface is quickly provided to the using vehicle, thereby providing convenience in managing the parking surface of the specific place.

1 is a block diagram illustrating an apparatus for providing parking information using camera imaging information according to an exemplary embodiment of the present invention.
2 is a view of an unparked parking surface according to an embodiment of the present invention.
3 is a view of a parked parking surface according to an embodiment of the present invention.
4 is a diagram of a single or a plurality of display units as an example of the display unit of the present invention.
FIG. 5 is a diagram illustrating a business operator that can use the present invention parking information providing apparatus.
FIG. 6 is a diagram for displaying individual identification numbers of individual parking surfaces of a parking lot. FIG.
7 is an exemplary view of a guide plate that can provide private parking information to the driver through the parking information providing apparatus of the present invention.

Hereinafter, a preferred embodiment of the present invention will be described in detail. First, in describing the present invention, a detailed description of related known functions or configurations will be omitted so as not to obscure the subject matter of the present invention.

As used herein, the terms 'about', 'substantially', and the like, are used at, or in close proximity to, numerical values when manufacturing and material tolerances inherent in the meanings indicated are intended to aid the understanding of the invention. To help you correct or

Absolute figures are used to prevent unfair use by unscrupulous intruders.

1 is a block diagram showing a parking information providing apparatus using a camera according to an embodiment of the present invention, Figure 2 is a view of an unparked parking surface according to an embodiment of the present invention, Figure 3 is an embodiment of the present invention A drawing of a towed parking surface.

The parking information providing apparatus 200 may include an image analyzer 220 analyzing a parking surface image based on images captured by the plurality of cameras 100; A parking information detector 225 included in the image analyzer 220 and configured to determine whether the vehicle is parked using the data of the image analyzer 220; And a database 250 for storing data regarding whether parking is possible and determined by the parking information detector.

The parking surface is preferably one or two or more clustered in a certain area, there is no limitation of the number of the camera 100 according to the number of parking surfaces, but according to the consideration factors such as outdoors or indoors or other features. For example, dozens or hundreds of cameras per camera or 5-7 cameras per camera can be secured, and it is preferable to install the cameras from various angles so that there are no blind spots on the parking surface. In other words, when monitoring a large outdoor parking lot, a camera installed in a high place can acquire images of hundreds (or more) parking lots (slots), whereas in a low-ceiling indoor parking lot, even if a wide-angle camera is used, several parking lots are available. You will be in charge of the noodles.

In addition, the camera 100 may be suitable for the shooting angle of 30 ~ 60 ° angle, even more preferably may be 40 ~ 50 °, but in the open outdoors using a wide angle camera to shoot the angle within 180 degrees to the corresponding data Can be used. At this angle, one camera can efficiently photograph several parking surfaces.

Each parking surface is preferably installed according to the standard, but may be irregular in the case of outdoors, and each parking surface may be divided into a parking boundary 10 and an inner surface 30 as shown in FIG. 2. Optionally, the display unit 50 may be present inside the parking surface.

Each of the three features to be described below, or two or more in combination, can extract the boundary surface 10 from the image captured by the camera, for example, by the contour automatic extraction method, and the area of the boundary surface 10 and the display unit 50. Can be calculated automatically. In addition, the boundary surface 10 of the parking surface may be preset.

Setting the boundary surface 10 of the parking surface in advance may be utilized under the premise that the arrangement of the parking surface is maintained for a long time as long as the camera is fixed well without being shaken by wind or due to external force. This is because, by omitting the processing of extracting the boundary surface 10 by software from the captured image data, the available computing resources can be focused on determining whether the parking surface is occupied. In this case, the user (for example, the manager of the parking lot or the person in charge of maintenance of the device) may designate the boundary surface 10 by hand by drawing a certain line on the captured image / image provided through any interface. If the interface 10 can be manually designated in this manner, it will be useful when the paint constituting the interface 10 is damaged due to corrosion, weathering, friction, etc., to become faint or broken in good faith, and the division is not clear.

On the other hand, the automatic contour extraction technique is, for example, 1) using the edge detection, 2) using the image enhancement and line information, 3) using the color and aspect ratio, 4) Hough Transform The method using may be used, but is not limited to the above-listed methods, but various methods may be used as it is or may be modified.

In addition, by adding some statistical techniques specific to contour extraction during machine learning, the accuracy can be improved through learning.

The image analyzer 220 may automatically calculate the total length and area of the parking boundary surface by extracting the contour of the parking boundary 10 of each parking surface using the automatic contour extraction method, and also inside each parking surface. One or more of the surface 30 color value and area may be automatically calculated.

The color value is not a color value determined according to the paint color of the parking surface (e.g., but not limited to YCbCr or RGB), but the color value of the actual image captured by the camera due to the color and brightness of the ambient light. It is a standard. It has a constant value in underground parking lots or in parking lots that are blocked from outside, rather than outside parking lots affected by the weather.

The display unit 50 may be located in the interior of the parking surface, but may be located in the center, and the number, size, or shape is not limited, but the display unit 50 may be hidden by the vehicle at the camera angle when the vehicle is parked. And size.

In addition, the display unit 50 can be easily extracted and the error can be reduced because the difference in the color value is displayed in the complementary color to the color of the parking surface so that the location and size can be easily extracted to the camera.

Similarly, the color of the boundary surface is preferably represented by the color of the parking surface and complementary colors.

The parking information detector 225 may determine whether the parking state or the parking is possible when the change in the data value of the image analyzer 220 before and after the vehicle parking on each parking surface is out of a predetermined range.

In the present invention, the elements for determining whether each parking surface is parked may be divided into three, and even if only one of the three is used, it may be used as a basis for determining parking.

In addition, at least one camera photographing a specific parking surface may be used based on the parking decision even if one of the data values photographed by one of the cameras is used.

First, as shown in FIGS. 2 and 3, when the vehicle is parked, the boundary surface 10 can not see more than one surface of the rectangular boundary surface 10 according to the angle of the camera. When a certain area of the boundary surface 10 is covered, the parking state can be displayed by detecting the difference between the beginning and the area.

In addition, the boundary surface 10 may be partially covered by the parking of the side parking surface. In this case, the initial value of the boundary surface 10 should be corrected to the hidden value, and it should be determined whether parking is based on this value.

Next, the inner surface 30, the inner surface 30 can quantify the change in the color value of the inner surface 30, rather than the numerical value of the change in area, it is possible to use the change in the color value as a parking decision. For example, a representative value such as a mean value or a median value of color values (YCbCr, RGB, or values based on various color coordinate systems) of the entire inner surface or several points within the inner surface of the captured image may be calculated. This representative value can be fixed where there is little change in lighting 24 hours, such as underground parking, if the parking space is empty. However, it may be adjusted dynamically or manually when sunlight is incident or there is a change in artificial lighting even outdoors or indoors. Using this representative value, the representative value of the color at the at least one or more points of the individual boundary surface 10 or inside the boundary surface 10 of the current captured image is compared with the value when it is determined that the parking space is empty. Therefore, if it is within a certain error range, the parking space is determined to be vacant.

Specifically, color values may be extracted from six points at the top and bottom, for example, two at the left and two at one boundary surface 10, but is not limited thereto. For example, it is assumed that the color value representative value at all or part of the interface 10 or at certain points is YCbCr (100, 80, 215). One way is that the sum of each number of YCbCr in the representative value is 100 + 85 + 215 = 395. If the color value representative value of the boundary surface 10 measured in the current image is YCbCr (90, 81, 220), the sum of each component is 90 + 81 + 220 = 391. If the threshold for judging that the parking surface is occupied is set to ± 20%, the parking lot is not occupied because 391 is within the range of (316, 474), which is an error range of 395 ± 20%. Instead of the summation method, it is also possible to compare the representative values for each component of Y, Cb, Cr, R, G, and B, which are each component of color values such as YCbCr or RGB, and to determine whether the threshold is out of at least one component Do.

On the other hand, in the case of outdoors, representative values by color values may change with time. It may also be affected by shadows, weather or seasons. Therefore, the reference value for determining that the color value of the inner surface 30 of the parking surface is empty can be set dynamically or by user setting.

On the contrary, if the comparison value is out of a certain error range, the parking space may be determined as occupy, and as soon as such a state change occurs, the parking space may be immediately determined whether or not the parking space is occupied. Can only change state if it lasts 3-10 seconds. For example, the predetermined time may be shortened to about 3 seconds in the downtown area and as long as 10 seconds in the outer area. The availability of the parking surface determined as described above is stored or updated in each parking slot record of the database 250 for each parking slot.

Finally, the display unit 50 may be located inside the parking surface, preferably in the center or at two or more points inside the parking surface.

4 illustrates that the display unit 50 may be positioned as a single or plural in the inner surface 30 by using 'P' as an example. The display unit 50 should be adjusted in size or position so as not to be influenced by identification in consideration of the camera angle and the occlusion of the side car, and the colors of the parking surface and complementary colors are advantageous for the efficiency of identification. Meanwhile, the display unit may have an arbitrary figure as shown in FIGS. 2 to 3 or may have a simple alphanumeric symbol as shown in FIG. 4. In the case of the alphanumeric symbol, it is possible to determine whether the display unit 50 is recognized using conventional optical character recognition (OCR) technology.

The presence or absence of parking may be determined by whether the display unit 50 is covered or not, and may be used independently or complementarily to the above two parking whether identification elements.

The information determined by the parking information detector 225 is stored in the database 250, and the database 250 is stored inside or outside the apparatus as shown in FIG. Through the storage and retrieval of data can be made. In addition, the information stored in the database 250 is transmitted to the external terminal after being processed by the service providing server 400 through a communication network so that the user of the external terminal can know whether the individual parking space is occupied through an appropriate application and interface. The database 250 stores the number of individual parking lots and stores whether the parking space is occupied as a separate record in the form of a flag.

The operation of the database 250 will be described with reference to FIG. 6. For example, if there are 100 cars in the 4th basement of Hyundai Department Store in Samsung-dong, if the site number of Samsung-dong Hyundai Department Store is 13 and the floor number is 4, each parking lot in the 4th basement of Hyundai Department Store in Samsung-dong will be 13-4-1, 13 Numbered as -4-2, 13-4-3, ... 13-4-100.

As shown in FIG. 5, the service providing server 400 may perform various transactions in cooperation with a parking lot owner, a local government, or a police officer, based on a business operator using the apparatus. For example, a parking lot owner with a lot of empty parking spaces can pay a fee and preferentially expose or advertise through an interface or application provided to the user. In addition, vehicle-related service providers, such as vehicle supplies dealers or car wash may pay the advertising fee and implement the advertisement of the vehicle goods or vehicle-related services to the user of the terminal 500. Local governments and police can provide information such as traffic campaigns and lost children, and pay a fee to the operator of the device.

7 shows a guide plate that can display private use information of individual parking surfaces. Currently used parking lot guidance is guidance by parking personnel or lighting of green and red lamps attached to the top of individual parking surfaces. However, there is a disadvantage in that parking spaces that can be parked extensively by columns, walls, or other vehicles are not identified in a wide parking lot. In contrast, FIG. 7 illustrating an exemplary embodiment of the present invention describes an interface that can be intuitively informed of which space is occupied, for example, at an entrance or a place of each level in an indoor parking lot.

For example, in the interface of FIG. 7, the red compartment is an occupied parking surface, the green compartment is an available parking surface, and the blue compartment is a disabled parking surface. These colors may vary depending on customs, laws, and circumstances of each country, region, and parking lot operator. When the lower portion of Figure 7 is the entrance of the floor or outdoor parking lot, the driver who enters the entrance immediately turn left and then park left or right turn and then turn right to proceed to the end to reach the empty green car and proceed to the end. Intuitive identification of vacant parking surfaces can be accomplished by quickly and efficiently than circulating. In addition, if the disabled vehicle can go straight to the top of the disabled parking lot.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and alterations are possible within the scope without departing from the technical spirit of the present invention. It will be evident to those who have knowledge of.

10: boundary surface 30: interior surface
50: display unit 100: camera
150: parking sensor 200: parking information providing device
220: image analysis unit 225: parking information detection unit
250: database 300: communication network
400: service providing server 500: terminal

Claims (6)

An image analyzer configured to analyze a parking surface image based on an image captured by at least one camera;
A parking information detector determining whether parking is possible using the data analyzed by the image analyzer; And
Parking information providing device using a real-time image processing including a database for storing the parking availability data determined by the parking information detector. The method of claim 1,
The image analyzer calculates an average value or a median value for the color value of the entire inner surface of the parking surface or each of several predetermined points of the inner surface,
The parking information detecting unit utilizes real-time image processing to detect the possibility of parking by determining whether the average value or the median value calculated by the image analyzer is outside a predetermined error color range. The method of claim 1,
The image analyzer recognizes at least one display unit displayed on the inner surface of the parking surface,
And the parking information detecting unit utilizes real-time image processing to determine whether the at least one display unit of the one or more display units displayed on the parking surface is parked when the image analysis unit is not recognized. The method of claim 1,
The image analyzer resets the area value of the boundary surface included in the parking surface and the area value of the inner surface by performing analysis initialization before analyzing the parking surface image.
The parking information detector determines whether the vehicle is out of a predetermined error area range based on the area value of the boundary surface and the area value of the inner surface reset by the image analyzer and detects whether parking is possible. Providing device. The method of claim 1,
Further comprising an interface for outputting a parking surface,
The interface is,
Real-time image processing, which displays the parking surface that is not occupied among the parking surfaces as the first color, the parking surface that is occupied among the parking surfaces is not available for parking, and the second color, and the parking surface for disabled persons among the available parking surfaces as the third color. Parking information providing device utilizing. Parking information providing apparatus of claim 1;
At least one user device interworked with the parking information providing apparatuses in a network; And
At least one service providing server interworking with the parking information providing devices in a network;
The service providing server,
The parking information providing apparatuses are sorted in order of increasing number of empty parking spaces among the parking surface information provided by the parking information providing apparatuses, and the exposure or alarm for the empty parking spaces is sequentially determined according to the sorted order. On one user device,
Parking information providing system using real-time image processing to be performed only when there is a payment for a predetermined fee for the exposure or alarm.

Images