KR20210085614A - System, apparatus and method for vision based parking management - Google Patents

Abstract

The present invention relates to a system, an apparatus, and a method for image-based parking management. According to the present invention, it is possible to detect the movements of a person and a vehicle in a parking lot using a parking lot image, prevent the occurrence of a vehicle accident, and more efficiently manage the parking lot. The system includes: at least one camera photographing a parking lot; and a parking management server using an image taken by the camera to manage the parking lot. The parking management server includes: a person recognition module using the image taken by the camera to calculate movement path information with respect to a person moving in the parking lot; and a parking lot management module managing the parking lot in view of the movement path information on the person.

Description

Image-based parking management system, apparatus and method {System, apparatus and method for vision based parking management}

The present invention relates to an image-based parking management system, apparatus, and method, and more specifically, detects the movement of people and vehicles in a parking lot using an image for a main window, and prevents a vehicle accident by taking this into account, and looks at the parking lot It relates to an image-based parking management system, apparatus, and method that can be efficiently managed.

Recently, various technologies for efficiently managing parking lots have been widely used.

More specifically, in order to provide customers with the convenience and speed of parking and to improve management efficiency, the technology that recognizes the license plate of a vehicle using a camera at the entrance and exit of the parking lot so that they can use the parking lot without issuing a separate parking ticket is spreading. , and furthermore, the number of parking lots applying a technology that allows a driver to easily recognize an empty parking surface by determining the presence or absence of a vehicle on each parking surface using an ultrasonic method and the like and displaying it with a lamp is steadily increasing.

However, in the conventional method as described above, it was difficult to provide and manage comprehensive information about the parking lot, and the utilization compared to the investment cost was not sufficient.

More specifically, one of the inconveniences that drivers often feel in a parking lot is that they feel difficult in identifying the parking area or parking surface where they park their vehicle. There was a limitation in that it was difficult to provide guidance information about the parking area or parking surface where the vehicle of the person was parked.

In contrast, in the prior art, in order to recognize the parking area or parking surface where each driver parked the vehicle and provide guidance information therefor, a plurality of cameras, etc. are installed in each parking area or parking surface to each parking area or parking surface. It should be able to recognize a parked vehicle, but there was a problem that a significant additional cost was required to do this.

Furthermore, if the vehicle number recognition can be performed directly from the camera installed in the parking lot, there will be no need to identify the location by matching the vehicle number when entering the vehicle, but in this case, a number of cameras must be installed in the parking lot and Since the cameras must be high-performance cameras that can recognize the number of a vehicle even from a long distance, the cost of configuring and maintaining the parking management system could increase significantly.

Furthermore, even if the driver moves to the corresponding parking area according to the parking status information of the information board installed at the entrance or passageway of the parking lot, other vehicles are parked in the empty parking area while the vehicle is moving, or the parked vehicle is parked while exiting. The situation may change, which may cause inconvenience to the driver.

As a more specific example, if 1 seat in a parking area in the parking lot is vacant and the vacant seat is located far from the parking lot entrance, the information board at the parking lot entrance will indicate that one seat is vacant, so that the driver entering the parking lot confirms the information on the guide plate and moves in the direction of the empty parking surface. However, in some cases, if another vehicle first parks on an empty parking surface while the vehicle is moving, when the vehicle arrives, the parking surface becomes an occupied parking surface parked by other vehicles, so the vehicle following A difficult situation arises where parking becomes impossible.

According to this problem, in the conventional parking management system, when 85% to 95% of the parking surface is occupied, it is displayed as full on the information board to prevent the above situation. As the state is managed, the problem of inefficient parking lot management follows.

Also, in the parking lot, not only the vehicle but also the person who got off the vehicle or the person who moves to get on the vehicle may exist, and the people in the parking lot may cause a risk of a vehicle accident. Furthermore, by analyzing the movement of the people, it is also possible to extract various information necessary for the management and operation of the parking lot.

Therefore, it is desirable to detect the movement of a person in the parking lot to prevent the risk of a vehicle accident and to reflect it in the more efficient management and operation of the parking lot, but parking management in consideration of this has not yet been made.

Accordingly, it is possible to provide and manage comprehensive information about the parking lot with a small investment cost, and furthermore, it is possible to manage the parking lot more efficiently and safely by using information about the people moving in the parking lot as well as the vehicle. There is a demand for this, but an appropriate solution has not yet been provided.

Republic of Korea Patent Publication No. 10-0670206 (Jan. 10, 2007)

The present invention was created to solve the problems of the prior art as described above, and enables comprehensive information provision and management of parking lots at a small investment cost, and furthermore, information about people and vehicles moving within the main window. An object of the present invention is to provide a parking management system, apparatus and method that can be used to manage a parking lot more efficiently and safely.

Parking management system according to an aspect of the present invention for solving the above problems, one or more cameras for photographing the parking lot; and a parking management server that manages the parking lot by using the image captured by the camera, wherein the parking management server uses the image captured by the camera to move a person moving in the parking lot a human recognition module for calculating path information; and a parking lot management module that manages the parking lot in consideration of the movement route information for the person.

In this case, in the person recognition module, the movement path information for the person is updated by recognizing the movement of the person using the first image frame of the image captured by the camera and the second image frame following the first image frame. can

In addition, the person recognition module may calculate the expected moving path of the person in real time by using the moving path information about the person.

In addition, in the person recognition module, by using the movement path information for the person, it is possible to classify a disembarking mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover moving to a vehicle parked in the parking lot.

Furthermore, in the parking lot management module, the riding mover moves to each vehicle parked in the parking lot, calculates the parking vehicle movement probability to be moved by the parked vehicle, and reflects the parking vehicle movement probability to each parking surface of the parking lot By calculating the probability of the vacant state, it is possible to provide parking guidance information to the vehicle in the parking lot. In addition, in the parking lot management module, the expected movement path of a person moving in the parking lot and the vehicle moving in the parking lot In consideration of the expected movement path, it is possible to provide risk guide information for a person or vehicle in the parking lot.

In addition, the camera, a first camera for photographing the license plate of the vehicle entering the parking lot; and one or more second cameras for photographing the parking lot, wherein the parking management server includes: a license plate recognition module for recognizing a license plate of a vehicle entering the parking lot; a vehicle recognition module for recognizing a vehicle in the parking lot; and a parking surface monitoring module for determining the occupancy state of the plurality of parking surfaces.

At this time, in the vehicle recognition module, the first vehicle newly entered into the parking lot is recognized from the image taken by the second camera, and the license plate information recognized by the license plate recognition module is stored in correspondence with the first vehicle, and the The movement path information of the second vehicle may be updated by recognizing the movement of the second vehicle using the first image frame of the image and the second image frame following the first image frame.

Also, the vehicle recognition module may calculate the expected moving path of the second vehicle in real time by using the moving path information about the second vehicle.

In addition, in the parking surface monitoring module, by comparing the license plate information and location information for the second vehicle generated by the vehicle recognition module with the area of the plurality of parking surfaces, to determine the vehicle located in the plurality of parking surfaces can

In addition, the parking management server according to another aspect of the present invention, in the parking management server for performing management of the parking lot by using the image taken by one or more cameras for photographing the parking lot, using the image taken by the camera a person recognition module for calculating movement path information for a person moving in the parking lot; and a parking lot management module that manages the parking lot in consideration of the movement route information for the person.

In this case, in the person recognition module, the movement path information for the person is updated by recognizing the movement of the person using the first image frame of the image captured by the camera and the second image frame following the first image frame. can

In addition, the person recognition module may calculate the expected moving path of the person in real time by using the moving path information about the person.

In addition, in the person recognition module, by using the movement path information for the person, it is possible to classify a disembarking mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover moving to a vehicle parked in the parking lot.

In addition, in the parking lot management module, the riding mover moves to each vehicle parked in the parking lot, calculates a parking vehicle movement probability that the parked vehicle will move, and reflects the parking vehicle movement probability to each parking surface of the parking lot It is possible to provide parking guidance information to the vehicle in the parking lot by calculating the probability of the vacant state.

In addition, in the parking lot management module, in consideration of the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, it is possible to provide risk guide information to the person or vehicle in the parking lot.

In addition, in the parking management method according to another aspect of the present invention, in the parking management method for performing management of the parking lot by using the image taken by one or more cameras for photographing the parking lot, using the image taken by the camera a person recognition step of calculating movement path information for a person moving in the parking lot; and a parking lot management step of performing management of the parking lot in consideration of the movement route information for the person.

In this case, in the step of recognizing the person, the movement path information for the person is updated by recognizing the movement of the person using the first image frame of the image captured by the camera and the second image frame following the first image frame. can

In addition, in the step of recognizing the person, an expected moving path of the person may be calculated in real time by using the moving path information about the person.

Also, in the step of recognizing the person, by using the movement route information for the person, it is possible to classify a get-off mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover who moves to a vehicle parked in the parking lot.

In addition, in the parking lot management step, the riding mover moves to each vehicle parked in the parking lot, calculates the parking vehicle movement probability to move the parked vehicle, and reflects the parking vehicle movement probability to each parking surface of the parking lot It is possible to provide parking guidance information to the vehicle in the parking lot by calculating the probability of the vacant state.

In addition, in the parking lot management step, in consideration of the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, it is possible to provide risk guide information to the person or vehicle in the parking lot.

In the parking management system, apparatus and method according to an embodiment of the present invention, it is possible to provide and manage comprehensive information about the parking lot at a small investment cost based on images such as video, and furthermore, to move within the main window An image-based parking management system, apparatus, and method that can manage a parking lot more efficiently and safely using information about people and vehicles are provided.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included as a part of the detailed description to help the understanding of the present invention, provide embodiments of the present invention, and together with the detailed description, explain the technical spirit of the present invention.
1 is a block diagram of a parking management system according to an embodiment of the present invention.
2 is a view for explaining the configuration of a parking management system according to an embodiment of the present invention.
3 is a view for explaining the arrangement structure and operation of the parking lot camera of the parking management system according to an embodiment of the present invention.
4 is a view specifically explaining the operation of the parking management system according to an embodiment of the present invention.
5 is a flowchart of a parking management method according to an embodiment of the present invention.
6 is a diagram illustrating a detailed flowchart of a vehicle recognition step in a parking management method according to an embodiment of the present invention.
7 is a view for explaining the operation of the parking surface monitoring module of the parking management system according to an embodiment of the present invention.
8 is a view for explaining the configuration and operation of a parking management system according to an embodiment of the present invention.
9 and 10 are diagrams for explaining a specific operation of the parking management system according to an embodiment of the present invention.
11 is a block diagram of a parking management server according to an embodiment of the present invention.

The present invention can apply various transformations and can have various embodiments. Hereinafter, specific embodiments will be described in detail based on the accompanying drawings.

The following examples are provided to provide a comprehensive understanding of the methods, apparatus and/or systems described herein. However, this is merely an example and the present invention is not limited thereto.

In describing the embodiments of the present invention, if it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, the terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to intentions or customs of users and operators. Therefore, the definition should be made based on the content throughout this specification. The terminology used in the detailed description is for the purpose of describing embodiments of the present invention only, and should not be limiting in any way. Unless explicitly used otherwise, expressions in the singular include the meaning of the plural. In this description, expressions such as “comprising” or “comprising” are intended to indicate certain features, numbers, steps, acts, elements, some or a combination thereof, and one or more other than those described. It should not be construed to exclude the presence or possibility of other features, numbers, steps, acts, elements, or any part or combination thereof.

In addition, terms such as first and second may be used to describe various components, but the components are not limited by the terms, and the terms are for the purpose of distinguishing one component from other components. used only as

Hereinafter, exemplary embodiments of a parking management system, apparatus and method according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

First, FIG. 1 shows a configuration diagram of a parking management system 100 according to an embodiment of the present invention. As can be seen in Figure 1, the parking management system 100 according to an embodiment of the present invention, one or more cameras 120 for photographing the parking lot, the parking lot using the image taken by the camera 120 It may be configured to include a parking management server 110 for performing management and a network 130 for transmitting an image signal generated by the camera 120 to the parking management server 110 .

At this time, the parking management server 110, as can be seen in FIG. 2, using the image captured by the camera 120, a person recognition module ( 114) and a parking lot management module 115 that manages the parking lot in consideration of the movement route information for the person.

Furthermore, the camera 120 may include a first camera 120a for photographing a license plate of a vehicle entering the parking lot and one or more second cameras 120b for photographing the parking lot, and the parking management server 110 is a license plate recognition module 111 for recognizing a license plate of a vehicle entering the parking lot using the image taken by the first camera 120a, and the parking lot from the image taken by the second camera 120b It may further include a vehicle recognition module 112 for recognizing a vehicle existing in the vehicle and a parking surface monitoring module 113 for recognizing the occupancy state of the plurality of parking surfaces.

Here, the license plate recognition module 111, the vehicle recognition module 112, the parking surface monitoring module 113, the parking lot management module 115 and the person recognition module 114 of the parking management server 110 are separated from each other. It is also possible to implement it as a hardware device such as a server of a server, and further it is possible to implement to process or analyze an image using cloud computing. Furthermore, the present invention is not limited thereto, and may be implemented in various hardware and software structures, such as implementing a plurality of software modules driven in a server.

In addition, the network 130 connecting the camera 120 and the parking management server 120 may include a wired network and a wireless network, and specifically, a local area network (LAN), a metropolitan area network (MAN: Metropolitan Area Network), a wide area network (WAN: Wide Area Network) may include a variety of communication networks, such as. In addition, the network 130 may include the Internet, such as the well-known World Wide Web (WWW). However, in the present invention, the network 130 is not limited to the networks listed above, and it is also possible to simply transmit the image signal generated by the camera 120 to the parking management server 120 using a conductive wire. In addition, when the network 130 is implemented using a wireless data network, the image signal generated by the camera 120 is efficiently transmitted to the parking management server 120 even in a wide parking lot or a multi-story parking lot. it is possible

3 illustrates the arrangement structure and operation of the camera 120 in the parking lot in the parking management system 100 according to an embodiment of the present invention.

As can be seen in FIG. 3 , in the parking management system 100 according to an embodiment of the present invention, a first camera 120a for photographing the license plate of a vehicle entering the parking lot and one or more second cameras for photographing the parking lot Two cameras 120b may be provided.

More specifically, as can be seen in FIG. 3 , the first camera 120a is disposed in the entry area (B1 of FIG. 3 ) for entering the parking lot to photograph the license plate of the vehicle entering the parking lot. have.

Furthermore, in the parking management system 100 according to an embodiment of the present invention, the first camera 120a is further disposed in the un-parking area (FIG. 3 (B2)) for un-parking from the parking lot. It is also possible to photograph the license plate of the vehicle.

In addition, as can be seen in FIG. 3 , in the parking management system 100 according to an embodiment of the present invention, one or more second cameras 120b for photographing the parking lot may be provided.

At this time, if the entire area of the parking lot can be photographed with one second camera 120b, it is also possible to use only one second camera 120b, but the area of the parking lot is large or it is located on a plurality of floors of the building. In the case where the parking lot is separated, it is preferable to use a plurality of second cameras 120b.

In this case, the plurality of second cameras 120b is preferably arranged at a location that can cover the entire area of the parking lot.

More specifically, as can be seen in FIG. 3 , two second cameras 120b are required to cover the entire area of the parking lot, and in this case, the corner areas where the two second cameras 120b face each other. By arranging at (A1) and (A2) in FIG. 3), the two second cameras 120b can effectively cover the entire area of the parking lot.

In addition, FIG. 4 specifically describes the operation of the parking lot management system 100 according to an embodiment of the present invention.

As can be seen in FIG. 4 , the camera 120 creates an image by photographing the parking lot. In addition, the person recognition module 114 of the parking management server 110 calculates movement path information for a person moving in the parking lot by using the image captured by the camera 120 (eg, in FIG. (C1) and (D1) of 4).

Accordingly, the parking lot management module 115 performs the management of the parking lot in consideration of the movement path information of the person in the parking lot calculated by the person recognition module 114 .

More specifically, the person recognition module 114 recognizes the movement of a person using a first image frame of the image captured by the camera 120 and a second image frame following the first image frame to recognize the movement of the person. The moving path information of the person is calculated while updating the moving path information for the .

In addition, in the person recognition module 114, by using the movement path information for the person, it is possible to classify an alighting mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover who moves to a vehicle parked in the parking lot. have.

As a more specific example, if the movement path of the person starts from a vehicle parked in the parking lot and proceeds toward the exit of the parking lot, the person may be determined to be an alighting mover (eg, (D1) in FIG. 4), and the If a person's moving path starts from the entrance side of the parking lot and proceeds toward the vehicle parked in the parking lot, the person may be determined to be a riding mover (eg, (C1) of FIG. 4 ).

In addition, in the person recognition module 114, in addition to getting off movers who get off the vehicle parked in the parking lot and move to the vehicle parked in the parking lot, and boarding movers who move to the vehicle parked in the parking lot, like parking lot managers, the parking lot It is also possible to classify people staying at the station as non-boarding persons.

For a more specific example, a person who meets predetermined conditions such as moving in a direction irrelevant to the direction of the vehicle parking surface in the parking lot or staying in the parking lot for more than a specific time is classified as an uninvolved getting on and off, and may be filtered and excluded from management, or may be processed in a manner that does not calculate the probability of moving a parked vehicle by moving to each vehicle parked in the parking lot and moving the parked vehicle.

Furthermore, the person recognition module 114 may calculate the expected moving path of the person in real time by using the moving path information about the person (eg, (C2) and (D2) of FIG. 4 ). .

Accordingly, in the parking lot management system 100 according to an embodiment of the present invention, the parking lot management module 115 moves the parking vehicle to which the riding mover moves to each vehicle parked in the parking lot to move the parked vehicle. A probability is calculated (for example, (P1) to (Pn) of FIG. 4), and a probability that each parking surface of the parking lot will be empty by reflecting the moving probability of the parked vehicle is calculated for a vehicle in the parking lot It is possible to provide parking guidance information.

Further, in the parking lot management module 115, the expected movement path of a person moving in the parking lot (eg, (D2) of FIG. 4) and the expected movement path of the vehicle moving in the parking lot (eg, FIG. 4) In consideration of (E)), in the case where the expected movement path of the person and the expected movement path of the vehicle intersect or overlap (for example, in (F) of FIG. 4), with respect to a person or a vehicle in the parking lot Hazard guidance information may also be provided.

In addition, FIG. 5 illustrates a flowchart of a parking lot management method according to an embodiment of the present invention. As can be seen in FIG. 5 , the parking lot management method according to an embodiment of the present invention recognizes a person that calculates movement route information for a person moving in the parking lot by using the image captured by the camera 120 . It may include a parking lot management step (S120) of performing management of the parking lot in consideration of the step (S110) and movement route information for the person.

Hereinafter, with reference to FIGS. 1 to 5 , the parking management system 100 , the apparatus, and the method according to an embodiment of the present invention are divided for each component and looked at in detail.

First, the license plate recognition module 111 recognizes the license plate of the vehicle entering the parking lot by using the image captured by the first camera 120a. In this case, the license plate recognition module 111 may use various technologies capable of recognizing the license plate of the vehicle from the image, such as a conventional license plate recognition algorithm, without any particular limitation.

Also, the vehicle recognition module 112 recognizes a vehicle in the parking lot using the second camera 120b.

Furthermore, the vehicle recognition module 112 recognizes the first vehicle newly entered into the parking lot from the image taken by the second camera 120b, and uses the license plate information recognized by the license plate recognition module 111 as the first The location information for the second vehicle can be updated by recognizing the movement of the second vehicle using the first image frame of the image and the second image frame following the first image frame. have.

Accordingly, when the vehicle recognition module 112 recognizes a new vehicle (that is, the first vehicle entering the parking lot) from the image captured by the second camera 120b, the license plate recognition module 111 The recognized license plate information is stored in correspondence with the first vehicle.

In addition, the vehicle recognition module 112 recognizes the movement of the second vehicle by using a first image frame of the image captured by the second camera 120b and a second image frame following the first image frame. The location information of the second vehicle is updated.

In this case, in the vehicle recognition module 112, it is possible to select and apply an appropriate algorithm in consideration of the application environment of the present invention among conventionally known tracking algorithms such as Adaboost, multiple instance learning, and kernelized correlation filter. Accordingly, whenever a frame of an image captured by the second camera 120b is input, a tracking algorithm is applied to detect all vehicles existing in the frame.

In addition, the vehicle recognition module 112 may recognize a third vehicle pulled out of the parking lot from the image captured by the second camera 120b and process the third vehicle.

In addition, in the vehicle recognition module 112 , when the second vehicle that was present in the first image frame of the image captured by the second camera 120b is not recognized in the second image frame, it is displayed in the second image frame. Waiting for a subsequent third image frame may retry to recognize the second vehicle.

Accordingly, the vehicle recognition module 112 generates and stores tracking information including the license plate information and location information for the first vehicle entering the parking lot using the image captured by the second camera 120b. In addition, by recognizing movement of a second vehicle recognized as existing in the parking lot and updating location information among tracking information for the second vehicle, tracking of vehicles entering and moving into the parking lot is performed. be able to do

More specifically, FIG. 6 shows a detailed flowchart of the vehicle recognition step S120 in the parking management method according to an embodiment of the present invention.

As can be seen in FIG. 6 , the vehicle recognition module 112 acquires a first image frame of an image, such as a video, from the second camera 120b for photographing the parking lot (S211).

Next, the vehicle recognition module 112 recognizes and extracts the vehicle object C from the first image frame (S212).

Accordingly, the vehicle recognition module 112 determines whether the vehicle object C is already being tracked (S213) with respect to the vehicle object C (S214).

At this time, if the vehicle object (C) is not yet being tracked (that is, when the vehicle object (C) enters a new parking lot, etc.), the vehicle recognition module 112 in the license plate recognition module 111 After determining whether there is newly recognized vehicle identification information (i.e., license plate information) (S218), if there is new vehicle identification information, the vehicle object C is newly included in the tracking target and the license plate information is added to the vehicle object ( C) (S219), and if there is no new vehicle identification information, error processing is performed (S220).

In addition, in the step S214, if the vehicle object C is already being tracked, it is determined whether the vehicle object C has passed the exit area of the parking lot (S215), and if the vehicle object C has passed the exit area, the vehicle object (C) is excluded from the tracking target (S216), and when the exit area is not passed, the location information of the vehicle object (C) is updated (S217).

Next, the vehicle recognition module 112 determines whether processing on all vehicle objects C recognized in the first image frame is completed (S221). If not, returns to step S213 and the remaining vehicle objects C ), and if completed, waits for the next image frame and repeats the above series of processes.

In addition, in the parking surface monitoring module 113 , the parking surface monitoring module 113 uses the image captured by the second camera 120b to determine the state of occupation of the plurality of parking surfaces.

In this regard, FIG. 7 exemplifies the operation of the parking surface monitoring module 113 in the parking management system 100 according to an embodiment of the present invention.

More specifically, as can be seen in FIG. 7 , the parking surface monitoring module 113 can recognize the parking surface of the image by segmenting it, and one parking surface corresponds to one segment. ((a) of FIG. 7).

In this case, the parking surface monitoring module 113 may retain matching information between the pixel area of the image and the physical area of the parking surface.

Accordingly, the parking surface monitoring module 113 performs occupancy detection for each segment. In this case, the parking surface monitoring module 113 may be implemented using a learning technique such as deep learning or machine learning, but the present invention is not necessarily limited thereto.

The parking surface monitoring module 113 compares the location information and license plate information for each vehicle generated by the vehicle recognition module 112 with the areas of a plurality of parking surfaces (segments) of the parking lot, and the vehicle for each parking surface Determination of whether the vehicle is occupied (FIG. 7 (b), (c)) and identification of the occupied vehicle are performed.

At this time, the parking surface monitoring module 113 uses the recognition result of one image first or all the recognition results of the plurality of images for the parking surface that is recognized as overlapping in the plurality of images by the plurality of second cameras 120b. can be taken into account.

Furthermore, in the parking surface monitoring module 113, the possibility (=P) of the vehicle being parked on the first parking surface using the image captured by the second camera 120b is higher than a predetermined first reference value. case is identified as a parked state (eg, P > 0.6), and when the probability is lower than a second predetermined threshold, it is identified as an empty state (eg, P < 0.4), and the probability is greater than or equal to the second threshold. , when it is less than or equal to the first reference value, it may be determined that an object other than a vehicle (eg, when a motorcycle or a cargo box is accumulated on the parking surface) is located in the first parking surface (eg, 0.4 =< P =< 0.6).

In this case, the first reference value and the second reference value may be determined according to a setting of a system administrator, or it is also possible to calculate an optimal value through a learning process such as machine learning in the parking management system 100, further In addition, it is possible to determine the occupancy state of the parking surface by applying various techniques such as deep learning.

In addition, even when the boundary line for the parking surface is not recognized in the image, the parking surface monitoring module 113 provides preset virtual parking surface information (eg, horizontal axis A, B, C, ??, vertical axis 1). , 2, 3, ??) can be used to determine whether the vehicle occupies the parking surface.

Furthermore, the parking surface monitoring module 113 uses the image captured by the second camera 120b to determine whether the vehicle has invaded the parking surface boundary line, and to the parking lot management module 115 . can provide

At this time, in the parking surface monitoring module 113, in consideration of the detection angle for the vehicle according to the location of the second camera 120b and the vehicle and the location of the parking surface boundary line, the parking surface boundary line of the vehicle It can be determined whether there is an invasion.

In addition, the person recognition module 114 calculates movement path information for a person moving in the parking lot by using the image captured by the camera 120 .

In this case, the person recognition module 114 may calculate movement path information for a person moving in the parking lot, similar to the vehicle recognition module 112 described above.

More specifically, the person recognition module 114 uses a first image frame of the image captured by the second camera 120b and a second image frame following the first image frame to detect the movement of a person in the parking lot. By recognizing it, it is possible to update movement route information for the person.

In this case, in the human recognition module 114, it is possible to select and apply an appropriate algorithm in consideration of the application environment of the present invention among known tracking algorithms such as Adaboost, multiple instance learning, and kernelized correlation filter. Accordingly, whenever a frame of an image captured by the second camera 120b is input, a tracking algorithm is applied to detect movement of a person in the image.

In addition, the person recognition module 114 uses the movement path information for the person to classify a get-off mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover who moves to a vehicle parked in the parking lot. .

For a more specific example, if the movement path of the person starts from the vehicle parked in the parking lot and proceeds toward the exit of the parking lot, the person may be determined to be an alighting mover, and the movement path of the person starts from the entrance side of the parking lot to the parking lot If the person proceeds toward the vehicle parked at , the person may be determined to be a passenger.

In addition, the person recognition module 114 may calculate the expected moving path of the person in real time by using the moving path information about the person.

Accordingly, the person recognition module 114 recognizes the movement of a person in the image captured by the second camera 120b and updates the movement route information for the person, thereby tracking people moving in the parking lot. will be able to perform

Accordingly, in the parking lot management module 115 , the parking lot management module 115 manages the parking lot using the license plate recognition module 111 , the vehicle recognition module 112 , and the parking surface monitoring module 113 . will perform

Accordingly, in the parking lot management module 115, the moving mover moves to each vehicle parked in the parking lot, calculates the parking vehicle movement probability to move the parked vehicle, and reflects the parking vehicle movement probability to the parking lot It is possible to provide parking guidance information for the vehicle in the parking lot by calculating the probability that each parking surface of the vacant state.

In this case, the parking guide information may be implemented in a way that a display lamp is installed on the ceiling of the parking lot, or one indicator lamp is installed for each of a plurality of parking surfaces such as three or four sides, and an empty parking surface in the vehicle passage. It is possible to implement in various structures, such as installing an induction signboard that displays the direction and number of the vehicle, so that the driver can recognize the location of an empty parking surface.

Accordingly, in the parking management system 100, apparatus, and method according to an embodiment of the present invention, it detects the movement of people and vehicles in the parking window, and whether the parking surface becomes empty in consideration of the person who wants to get out of the parked vehicle Alternatively, it is possible to receive more accurate real-time parking information from the perspective of a driver who wants to park by judging whether it will remain occupied by a probability and providing the parking information that reflects this to an entrance information board or a block guide board, etc. By providing real-time parking information with higher accuracy than the management system, it is possible to minimize driver inconvenience and to operate the parking lot more efficiently.

Furthermore, in the parking lot management module 115, in consideration of the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, the risk guide information for the person or vehicle in the parking lot is provided. can

In addition, the parking lot management module 115 receives the occupancy and occupied vehicle identification information (eg, license plate information) for each parking surface from the parking surface monitoring module 113 and stores it in a database to manage it. can

In addition, the parking lot management module 115 may receive information on the vehicle exiting the parking lot from the vehicle recognition module 112 to update the database.

Accordingly, the parking lot management module 115 may manage the parking time for each vehicle by using the information in the database.

In addition, in the parking lot management module 115, when the first vehicle enters the parking lot, the parking success probability is calculated when the vehicle moves to each parking surface, and the parking success probability is taken into consideration for the first vehicle. Route guidance information can be provided.

Furthermore, in the parking lot management module 115, the predicted probability that the first vehicle will be parked on each parking surface is calculated in consideration of the real-time movement path of the first vehicle identified by the vehicle recognition module 112, and the Parking guide information may be provided for the first vehicle in consideration of the predicted probability for each parking surface.

In this case, the parking lot management module 115 may calculate a predicted probability that the first vehicle will be parked on each parking surface in consideration of one or more of the current moving direction and the moving speed of the first vehicle.

In this case, the moving direction of the first vehicle may be calculated in consideration of a detection angle with respect to the first vehicle or a moving path of the first vehicle, and the moving speed of the first vehicle may be determined by a photographing period of the image and the image. The interval may be calculated in consideration of a change in the position of the first vehicle.

Accordingly, in the parking management system 100 according to an embodiment of the present invention, while predicting in real time a movement path of a person and a vehicle moving in the main window using an image, the vehicle has a high probability of successful parking. By determining the surface and the parking area and instructing the driver in the direction of movement using a parking guide indicator, etc., the driver is helped to park more easily and quickly.

In addition, when the vehicle recognition module 112 fails to track the location information for the first vehicle, the parking lot management module 115 calculates the parking prediction probability for each parking surface of the first vehicle. The position of the first vehicle may be estimated by using the .

Accordingly, even when the parking location of the first vehicle is not clearly identified in the parking lot management module 115, the first vehicle is located in a candidate area where the first vehicle can be located based on the predicted probability to the user. information can be provided, and further, by providing an image of the candidate area, it is possible to help the user to identify the location of his/her vehicle.

In addition, FIG. 8 describes the configuration and operation of the parking management system 100 according to an embodiment of the present invention.

As can be seen in FIG. 8 , in the parking management system 100 according to an embodiment of the present invention, the license plate recognition module 111 (License Plate Recognition, LPR) is a first to photograph a vehicle entering the parking lot. The license plate is recognized from the image of the camera 120a.

Accordingly, the vehicle recognition module 112 recognizes the vehicle from the parking lot image of one or more second cameras 120b for photographing the parking lot, and the license plate information recognized by the license plate recognition module 111 and the image of the vehicle. It will match your location information.

In addition, the person recognition module 114 calculates movement path information for a person moving in the parking lot from the parking lot image of one or more second cameras 120b for photographing the parking lot.

In addition, in the parking surface monitoring module 113, using the parking lot image of the second camera 120b and the location information and license plate information for the vehicle recognized by the vehicle recognition module 112, for each parking surface Whether or not the vehicle is occupied and identification information on the occupied vehicle is calculated. At this time, the parking surface monitoring module 113 retains matching information between the pixel area of the image and the physical area of the parking surface to be used to calculate whether a vehicle is occupied for each parking surface and identification information about the occupied vehicle. can

Accordingly, in the parking lot management module 115 , whether the vehicle is occupied for each parking surface provided from the parking surface monitoring module 113 , and identification information on the occupied vehicle, and taking out provided from the vehicle recognition module 112 . In consideration of the information, it is possible to provide information about each parking surface of the main window to the user.

In addition, in the parking lot management module 115, the riding mover moves to each vehicle parked in the parking lot, calculates the parking vehicle movement probability to be moved by the parked vehicle, and reflects the parking vehicle movement probability to each of the parking lot It is possible to provide parking guidance information to the vehicle in the parking lot by calculating the probability that the parking surface is in an empty state.

Furthermore, in the parking lot management module 115, in consideration of the expected movement path of a person moving in the parking lot and the expected movement path of a vehicle moving in the parking lot, risk guide information can be provided for people or vehicles in the parking lot. there will be

In addition, FIGS. 9 and 10 illustrate a specific operation of the parking management system 100 according to an embodiment of the present invention.

More specifically, in the parking management system 100 according to an embodiment of the present invention, the parking lot management module 115 considers the real-time movement path of the first vehicle identified by the vehicle recognition module 112 . It is possible to calculate a predicted probability that the first vehicle will park on each parking surface, and provide parking guidance information to the first vehicle in consideration of the predicted probability for each parking surface.

In this case, the parking guide information may be implemented in a way that a display lamp is installed on the ceiling of the parking lot, or one indicator lamp is installed for each of a plurality of parking surfaces such as three or four sides, and an empty parking surface in the vehicle passage. It is possible to implement in various structures, such as installing an induction signboard that displays the direction and number of the vehicle, so that the driver can recognize the location of an empty parking surface.

In this case, the parking lot management module 115 may calculate a predicted probability that the first vehicle will be parked on each parking surface in consideration of one or more of the current moving direction and the moving speed of the first vehicle.

In addition, in the parking lot management module 115, the riding mover moves to each vehicle parked in the parking lot from the movement path information of people moving in the parking lot identified by the sarim recognition module 114, and the parked vehicle is It is possible to calculate the moving probability of the parking vehicle to be moved, and to calculate the probability that each parking surface of the parking lot is in an empty state by reflecting the parking vehicle movement probability, thereby providing parking guidance information for the vehicle in the parking lot.

In addition, in the parking lot management module 115, in consideration of the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, to provide risk guide information for the person or vehicle in the parking lot. can

Accordingly, in the parking management system 100 according to an embodiment of the present invention, it is possible to provide and manage comprehensive information about the parking lot at a small investment cost based on an image such as a video, and furthermore, in the main window Parking lot can be managed more efficiently and safely by using information on moving people and vehicles.

More specifically, in the parking management system 100 according to an embodiment of the present invention, a method of calculating the expected moving path of the vehicle and the expected moving path of the person is as follows. The method of calculating the expected movement path of the vehicle is mainly described below, but this can also be applied to the case of calculating the expected movement path of a person.

First, in order to calculate the expected moving path of the vehicle and the expected moving path of a person, as can be seen in FIG. 9 , the parking area of the image is divided into n non-overlapping cells of size w (horizontal) x h (vertical). (Here, the cells are referred to as c ₁ , c ₂ , .. c _{n , respectively).}

이 된다. In this case, the movement path of the vehicle and the person may be displayed in the order of cells, and accordingly, the path of the vehicle i passing through k cells may be expressed as [c _i1 , c _i2 ,..., c _ik ], At this time

becomes this

Also, at time t The location of the cell to which the vehicle belongs may be determined as the cell to which the center point of the vehicle area recognized in the image belongs. Vehicles are tracked and managed from the point of entry into the parking lot until they are parked, and for the parked vehicle, tracking of the movement path may be stopped, and the stored route may be deleted.

More specifically, in calculating the center point of the vehicle area from the image, various algorithms may be applied, such as obtaining a center point for each vertex of the vehicle area, and furthermore, using a learned neural network circuit, etc., the vehicle area It is also possible to calculate the position of the cell to which the vehicle belongs by calculating the center point of .

Accordingly, the movement route of the vehicle predicted at time t is received as an input a moving path to the point in time t, is the prediction order m of cells. In this case, a deep learning model based on Recurrent Neural Networks (RNN) may be used for predicting the moving path of the vehicle, but the present invention is not necessarily limited thereto.

The input may be j past cell movement paths from the present, empty parking surface information, and movement direction vectors. In this case, the empty parking surface information may express the occupied state of the parking surface as a one-dimensional vector in the form of a one-hot vector.

In addition, the movement vector may be expressed as a unit vector representing the vehicle movement direction, with an upper left end as (0,0), a positive x-axis from left to right, and a positive y-axis from top to bottom.

In addition, the RNN (Recurrent Neural Networks)-based deep learning model may consist of m- step RNN cells, and when the input movement path is shorter than m- step, zero padding may be performed in front of the path. In addition, when the input movement path is longer than the m- step, the most recent m-step may be used.

Accordingly, the relationship between the RNN model, input, and output for predicting the moving path of the vehicle may be expressed as shown in FIG. 10 .

In this case, in RNN learning for predicting the movement path of the vehicle, actual vehicle movement trajectory data may be used, or virtual trajectory data may be used.

Then, it is calculate the probability (P _o) to be occupied with respect to the parking surface is empty by using the moving path of the predicted vehicle. In this case, the probability P _o may be calculated for each parking surface or parking area, and the occupation probability ( P _o,i ) for the vehicle to occupy the parking surface or parking area i satisfies Equation 1 below. .

[Equation 1]

Then, the occupancy probability of the parking surface or parking area k The calculation of P _o,k may be calculated considering the following parameters.

_{(1) Proximity (Approx k} ) between the vehicle movement path and the parking surface or parking area k: the closer the vehicle, the greater the occupancy probability.

(2) Preference for parking surface or parking area k (Pref _k ): Set the preference for a parking area or parking area that is close to the destination or has high accessibility.

In addition, the occupation probability P _o,k Wow The relationship between these parameters can be expressed by Equation 2 as follows.

[Equation 2]

라 할 때, 이들을 합산한 값을 계산에 고려된 차량 대수 n으로 나눈 값이 아래의 수학식 3과 같이 P_o,k 가 된다.At this time, the occupancy probability for the parking surface or parking area k is calculated for each of all moving vehicles, and the probability by vehicle i is

, the value obtained by dividing the sum of these values by the number of vehicles n _{considered for calculation is equal to P o,k} as in Equation 3 below. becomes

[Equation 3]

In addition, since pedestrians in the parking lot can get out of the parked vehicle, at time t As the vehicle moves out of the parking surface or parking area k, the probability P _{f,k that the parking surface or parking area k becomes available for parking increases as the proximity dist p,k} to the movement path of the pedestrian p becomes smaller. It can be calculated as in Equation (4). Here, g is a constant coefficient.

[Equation 4]

In this case, the probability P _f is calculated for all moving pedestrians for each parking area or parking area where the vehicle is parked, and dividing the sum by the number of pedestrians considered in the calculation is the probability that the corresponding parking surface or parking area will be available for parking.

Accordingly, the probabilities P _o and P _{f calculated above} Using , the parking availability index P _a for each parking surface or parking area may be calculated as in Equation 5 below. where c _f and c _o are constant coefficients.

[Equation 5]

At this time, as can be seen in Equation 5, the parking availability index increases as the occupancy probability is low and the airspace probability is high.

By using this, in the parking lot management system 100 according to an embodiment of the present invention, _{it is possible to induce a} vehicle newly entering the parking lot in the order of a parking area having a high probability Pa.

Furthermore, in the parking lot management system 100 according to an embodiment of the present invention, the riding mover moves to each vehicle parked in the parking lot and calculates the parking vehicle movement probability to move the parked vehicle, and the parking vehicle movement probability It is possible to provide parking guidance information for a vehicle in the parking lot by calculating a probability that each parking surface of the parking lot is in an empty state by reflecting the .

In addition, in the parking lot management system 100 according to an embodiment of the present invention, considering the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, with respect to the person or vehicle in the parking lot Hazard guidance information may also be provided.

Accordingly, in the parking lot management system 100 according to an embodiment of the present invention, it is also possible to preemptively induce parking so that parking can be dispersed using the prediction of the parking area of the vehicle, and also the movement of people in the parking lot By providing parking guidance information for vehicles or providing risk guidance information in consideration of this, the parking lot can be managed more efficiently and safely.

In addition, the computer program according to another aspect of the present invention is characterized in that it is a computer program stored in a computer readable medium in order to execute each step of the above salpin parking management method on the computer. The computer program may be a computer program including a machine language code generated by a compiler, as well as a computer program including a high-level language code that can be executed in a computer using an interpreter or the like. In this case, the computer is not limited to a personal computer (PC) or notebook computer, and includes a central processing unit (CPU) such as a server, smart phone, tablet PC, PDA, mobile phone, etc. to process any information capable of executing a computer program. It includes a device, and further may include a case of processing information such as processing or analyzing an image using cloud computing or the like. In addition, the computer-readable medium includes an electronic recording medium (eg, ROM, flash memory, etc.), a magnetic storage medium (eg, a floppy disk, a hard disk, etc.), an optically readable medium (eg, CD-ROM). , DVD, etc.) includes any computer-readable storage medium.

In addition, FIG. 11 exemplifies the configuration of the server 120 that provides image search according to an embodiment of the present invention.

As can be seen in FIG. 11 , the parking management server 110 according to an embodiment of the present invention uses an image captured by one or more cameras 120 for photographing a parking lot to perform management of the parking lot. As the management server 110, in consideration of the person recognition module 114 that calculates movement route information for a person moving in the parking lot using the image captured by the camera 120 and movement route information for the person It may be configured to include a parking lot management module 115 that manages the parking lot.

Furthermore, in the present invention, the parking management server 110 includes a license plate recognition module 111 for recognizing a license plate of a vehicle entering the parking lot, a vehicle recognition module 112 for recognizing a vehicle in the parking lot, and the plurality of parking surfaces. It may be configured to further include a parking surface monitoring module 113 to determine the occupancy state of the.

Below, the parking management server 110 according to an embodiment of the present invention will be divided for each component. More detailed information about the parking management server 110 according to an embodiment of the present invention can be inferred from the description of the parking management system and method according to the embodiment of the present invention described above, a more detailed description will be given below. is omitted.

First, the license plate recognition module 111 recognizes the license plate of the vehicle entering the parking lot by using the image captured by the first camera 120a.

In addition, the vehicle recognition module 112 recognizes the vehicle in the parking lot by using the image captured by the second camera 120b.

At this time, in the vehicle recognition module 112, the first vehicle newly entered into the parking lot is recognized from the image taken by the second camera 120b, and the license plate information recognized by the license plate recognition module 111 is the second It is stored in correspondence with one vehicle, and the position information of the second vehicle can be updated by recognizing the movement of the second vehicle using the first image frame of the image and the second image frame following the first image frame. have.

In addition, the parking surface monitoring module 113 determines the occupied state of the plurality of parking surfaces.

In addition, the person recognition module 114 calculates movement path information for a person moving in the parking lot by using the image captured by the camera 120 .

At this time, the person recognition module 114 recognizes the movement of a person using a first image frame of the image captured by the camera 120 and a second image frame following the first image frame to recognize the movement of the person. Movement route information can be updated.

In addition, the person recognition module 114 may calculate the expected moving path of the person in real time by using the moving path information about the person.

In addition, in the person recognition module 114, by using the movement path information for the person, it is possible to classify an alighting mover who gets off the vehicle parked in the parking lot and moves, and a boarding mover who moves to a vehicle parked in the parking lot. have.

Finally, the parking lot management module 115 performs management of the parking lot.

At this time, in the parking lot management module 115, the parking success probability is calculated when the first vehicle moves to each parking surface, and route guidance information is provided to the first vehicle in consideration of the parking success probability. can

In addition, in the parking lot management module 115, in consideration of the real-time movement path of the first vehicle identified by the vehicle recognition module 112, a predicted probability that the first vehicle will be parked on each parking surface is calculated, and each Parking guide information may be provided to the first vehicle in consideration of the predicted probability for the parking surface.

In addition, in the parking lot management module 115, the riding mover moves to each vehicle parked in the parking lot, calculates the parking vehicle movement probability to move the parked vehicle, and reflects the parking vehicle movement probability of the parking lot It is possible to provide parking guidance information for a vehicle in the parking lot by calculating a probability that each parking surface is in an empty state.

In addition, in the parking lot management module 115, in consideration of the expected movement path of the person moving in the parking lot and the expected movement path of the vehicle moving in the parking lot, to provide risk guide information for the person or vehicle in the parking lot. can

Accordingly, in the parking management system 100, apparatus and method according to an embodiment of the present invention, it is possible to provide and manage comprehensive information about the parking lot with a small investment cost based on images such as video, and further To provide an image-based parking management system 100, an apparatus, and a method that can manage a parking lot more efficiently and safely by using information about people and vehicles moving in the main window.

The above description is merely illustrative of the technical spirit of the present invention, and various modifications and variations will be possible without departing from the essential characteristics of the present invention by those skilled in the art to which the present invention pertains. Accordingly, the embodiments described in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and are not limited to these embodiments. The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

100: parking management system
110: parking management server
111: license plate recognition module
112: vehicle recognition module
113: parking surface monitoring module
114: human recognition module
115: parking lot management module
120: camera
120a: first camera
120b: second camera
130: network

Claims (22)

one or more cameras to photograph the parking lot; and
and a parking management server that manages the parking lot by using the image captured by the camera; and
The parking management server,
a person recognition module for calculating movement path information for a person moving in the parking lot by using the image captured by the camera; and
A parking management system comprising a; a parking management module that manages the parking lot in consideration of the movement route information for the person. According to claim 1,
In the human recognition module,
Parking management system, characterized in that the movement of the person is recognized by using a first image frame of the image captured by the camera and a second image frame following the first image frame to update movement route information for the person. 3. The method of claim 2,
In the human recognition module,
A parking management system, characterized in that by using the moving path information for the person, calculating the expected moving path of the person in real time. 3. The method of claim 2,
In the human recognition module,
Using the movement route information for the person,
A parking management system, characterized in that it is divided into a disembarking mover who gets off the vehicle parked in the parking lot and moves to a vehicle parked in the parking lot and a boarding mover moving to the vehicle parked in the parking lot. 5. The method of claim 4,
In the parking lot management module,
Calculate the parking vehicle movement probability that the riding mover moves to each vehicle parked in the parking lot and moves the parked vehicle,
A parking management system, characterized in that by reflecting the moving probability of the parked vehicle, calculating the probability that each parking surface of the parking lot is in an empty state, and providing parking guidance information for the vehicle in the parking lot. 4. The method of claim 3,
In the parking lot management module,
Considering the expected movement path of a person moving in the parking lot and the expected movement path of a vehicle moving in the parking lot,
A parking management system, characterized in that it provides risk guide information for a person or a vehicle in the parking lot. According to claim 1,
The camera is
a first camera for photographing a license plate of a vehicle entering the parking lot; and
Including; at least one second camera for photographing the parking lot;
The parking management server,
a license plate recognition module for recognizing a license plate of a vehicle entering the parking lot;
a vehicle recognition module for recognizing a vehicle in the parking lot; and
Parking management system, characterized in that it further comprises; a parking surface monitoring module for identifying the occupied state of the plurality of parking surfaces. 8. The method of claim 7,
In the vehicle recognition module,
Recognizing a first vehicle newly entered into the parking lot from the image taken by the second camera and storing the license plate information recognized by the license plate recognition module in correspondence with the first vehicle,
A parking management system, characterized in that by recognizing the movement of a second vehicle using a first image frame of the image and a second image frame following the first image frame, the movement route information for the second vehicle is updated. 9. The method of claim 8,
In the vehicle recognition module,
A parking management system, characterized in that the predicted moving path of the second vehicle is calculated in real time by using the moving path information of the second vehicle. 9. The method of claim 8,
In the parking surface monitoring module,
By comparing the license plate information and location information for the second vehicle generated by the vehicle recognition module with the areas of the plurality of parking surfaces,
A parking management system, characterized in that the vehicle located in the plurality of parking surfaces is identified. In the parking management server for performing management of the parking lot by using the image taken from one or more cameras for photographing the parking lot,
a person recognition module for calculating movement path information for a person moving in the parking lot by using the image captured by the camera; and
A parking management server comprising a; a parking management module that manages the parking lot in consideration of the movement path information for the person. 12. The method of claim 11,
In the human recognition module,
A parking management server, characterized in that by recognizing the movement of a person using a first image frame of the image captured by the camera and a second image frame following the first image frame, the movement route information for the person is updated. 13. The method of claim 12,
In the human recognition module,
Parking management server, characterized in that for calculating the expected moving path of the person in real time by using the moving path information about the person. 13. The method of claim 12,
In the human recognition module,
Using the movement route information for the person,
A parking management server, characterized in that it is divided into a get-off mover who gets off the vehicle parked in the parking lot and moves to a vehicle parked in the parking lot and a boarding mover who moves to the vehicle parked in the parking lot. 15. The method of claim 14,
In the parking lot management module,
Calculate the parking vehicle movement probability that the riding mover moves to each vehicle parked in the parking lot and moves the parked vehicle,
A parking management server, characterized in that by reflecting the moving probability of the parked vehicle, calculating the probability that each parking surface of the parking lot will be in an empty state, and providing parking guidance information for the vehicle in the parking lot. 14. The method of claim 13,
In the parking lot management module,
Considering the expected movement path of a person moving in the parking lot and the expected movement path of a vehicle moving in the parking lot,
Parking management server, characterized in that for providing risk guide information for people or vehicles in the parking lot. In the parking management method of performing management of the parking lot by using the image taken by one or more cameras for photographing the parking lot,
a person recognition step of calculating movement path information for a person moving in the parking lot by using the image captured by the camera; and
Parking management method comprising a; parking management step of performing management of the parking lot in consideration of the movement route information for the person. 18. The method of claim 17,
In the human recognition step,
A parking management method, characterized in that by recognizing the movement of a person using a first image frame of the image captured by the camera and a second image frame following the first image frame, the movement path information for the person is updated. 19. The method of claim 18,
In the human recognition step,
A parking management method, characterized in that by using the moving path information for the person, calculating the expected moving path of the person in real time. 19. The method of claim 18,
In the human recognition step,
Using the movement route information for the person,
A parking management method, characterized in that divided into a disembarking mover who gets off the vehicle parked in the parking lot and moves to a vehicle parked in the parking lot and a boarding mover moving to the vehicle parked in the parking lot. 21. The method of claim 20,
In the parking lot management step,
Calculate the parking vehicle movement probability that the riding mover moves to each vehicle parked in the parking lot and moves the parked vehicle,
A parking management method, characterized in that by reflecting the moving probability of the parked vehicle, calculating the probability that each parking surface of the parking lot is in an empty state, and providing parking guidance information for the vehicle in the parking lot. 20. The method of claim 19,
In the parking lot management step,
Considering the expected movement path of a person moving in the parking lot and the expected movement path of a vehicle moving in the parking lot,
Parking management method, characterized in that for providing risk guide information for people or vehicles in the parking lot.

Images