KR102380662B1 - Parking space sharing system based on artificial intelligence - Google Patents

Abstract

The present invention relates to an artificial intelligence-based parking lot sharing system, comprising: a parking information acquisition device installed in each parking lot to acquire parking information regarding a vehicle parking state; a camera 120 installed at the entrance of each parking lot to photograph a vehicle to obtain a vehicle image; In the frame of the vehicle image transmitted in real time from the camera 120, the vehicle including the license plate is recognized through a deep learning model, and the recognized vehicle information is compared with the reference vehicle information stored in advance through learning in the cloud DB or local DB. an IoT device 130 that determines whether or not a vehicle is present according to whether the information matches or not; a circuit breaker 140 installed at the entrance of each parking lot to control vehicle entry; A central control server ( 170). According to the present invention, when the owner of the parking space does not use the parking space, other people can share the parking space, so that the parking space can be used efficiently.

Description

AI-based parking lot sharing system {PARKING SPACE SHARING SYSTEM BASED ON ARTIFICIAL INTELLIGENCE}

The present invention relates to a parking lot sharing system, and more particularly, to an artificial intelligence-based parking lot sharing system that can solve a shortage of parking space.

One of today's serious urban problems is vehicle parking. The parking problem arises because the number of parking spaces cannot support it compared to the exponentially increasing number of vehicles. Accordingly, efforts are being made to solve the parking problem by constructing a parking building, but a fundamental solution has not been presented.

On the other hand, in the case of a multi-family house such as an apartment, a large parking lot is usually secured, but it is used relatively much only at specific times such as at night, and partly during the daytime. Hopefully it can be resolved.

As a prior art related field patent technology, a reservation server configured to provide a map including information of parking lots to parking lot users by registering the location of the parking lot and the available time of the parking lot from the parking lot owner, and installed in each parking area of the parking lots There is proposed a parking lot sharing system using a personal idle space including a circuit breaker configured to control a blocking bar or a leaf spring by receiving from the reservation server whether the parking area has been reserved by the parking member using the parking lot (Patent Document 1) Reference).

In addition, it is installed on each parking surface of the parking lot and functions to block the entry of vehicles, and receives the location of the parking lot and the available time of the parking lot from the parking lot manager, and provides a map containing information about the parking lot to the members. A parking machine configured to communicate with a search server to be provided, comprising: a blocking unit for blocking access to a parking surface; a motor for controlling the posture of the blocking unit; a parking machine control device connected to the Internet for controlling the motor and having a parking reservation function; a parking machine control device communication unit for receiving a parking user who has reserved a use of the parking surface on which the parking machine is installed through the search server and information about the reservation; a smart device access information receiving unit configured to receive information on whether the smart device of the parking customer has approached the reserved parking surface; a light emitting unit configured to emit light to inform the location of the reserved parking surface when the smart device of the parking customer approaches the reserved parking surface; including, wherein the parking machine changes the state of the parking machine to notify whether parking is possible; registering the parking status in the search server; An IoT-based parking machine that directly makes a reservation with parking guests has been proposed (see Patent Document 2).

Domestic Patent Publication 10-2019-0095611 Domestic Patent Publication 10-2019-0117839

An object of the present invention is to provide an artificial intelligence-based parking lot sharing system that can improve the convenience of parking lot users and the efficiency of parking lot management by solving the shortage of parking space.

In order to achieve the above object, an artificial intelligence-based parking lot sharing system according to the present invention is a system for sharing parking lots installed in a plurality of places. device 110; a camera 120 installed at the entrance of each parking lot to photograph a vehicle to obtain a vehicle image; In the frame of the vehicle image transmitted in real time from the camera 120, the vehicle including the license plate is recognized through a deep learning model, and the recognized vehicle information is compared with the reference vehicle information stored in advance through learning in the cloud DB or local DB. an IoT device 130 that determines whether or not a vehicle is present according to whether the information matches or not; a circuit breaker 140 installed at the entrance of each parking lot to control vehicle entry; A central control server ( 170) in an artificial intelligence-based parking lot sharing system, including;
The IoT device 130 recognizes and classifies as a vehicle through vehicle image analysis, and applies an autoencoder that learned the original image without contamination or damage of the classified vehicle, and sends the classified vehicle image to the encoder. After extracting features by encoding and compressing the extracted features, extracting an approximate image that is an approximation of the original image through decoding by a decoder to further collect information on the degree of pollution and damage of the vehicle. It provides an artificial intelligence-based parking lot sharing system.

delete

delete

delete

delete

delete

delete

delete

According to the present invention, when the owner of the parking space does not use the parking space, other people can share the parking space, so that the shortage of parking space can be solved by efficiently using the parking space.

In addition, an image is acquired by photographing a vehicle approaching with a camera, and the vehicle can be effectively controlled after recognizing the vehicle through a deep learning model from the acquired vehicle image and transmitted in real time through a real-time streaming protocol. .

In addition, based on vehicle status information such as vehicle number, vehicle model, vehicle color, entry/exit time, pollution level, damage level, etc. Related services for the vehicle can be relayed to partner organizations such as

1 is a conceptual diagram of an artificial intelligence-based parking lot sharing system according to the present invention.
2 is a block diagram of an IoT device according to the present invention;
3 to 6 are conceptual diagrams of segmentation according to the present invention.
7 is a conceptual diagram of an autoencoder according to the present invention.
8 is a flowchart of a vehicle pollution degree and damage level recognition method according to the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

An important technical feature of the AI-based parking lot sharing system according to the present invention is that it is configured to greatly improve the convenience of parking lot users and the efficiency of parking lot management by solving the shortage of parking space.

As shown in FIG. 1, the artificial intelligence-based parking lot sharing system according to the present invention is a system for mutually sharing parking lots installed in a plurality of places in various regions, and is largely a parking information acquisition device 110 and a camera 120. , the IoT device 130, the circuit breaker 140, the display device 150, the environmental information collector 160, the central control server 170 and the cooperative organization server 180, each comprising the Internet, LAN ( Data can be exchanged through communication networks such as Local Area Network) and VAN (Value Added Network).

The parking information acquisition device 110 may be composed of an infrared sensor or an ultrasonic sensor, etc., is installed in each parking lot to acquire parking information about the vehicle parking state. That is, to obtain the parking status information of each parking lot.

The camera 120 may include an IP camera, etc., which are installed at the entrance of each parking lot to photograph an entering vehicle in all directions to acquire a vehicle image, and use a real-time streaming protocol (RTSP). protocol) is supported. Therefore, after acquiring the vehicle image, it can be transmitted to the communication network in real time through the real-time streaming protocol.

In the present invention, in order to effectively collect vehicle state information such as vehicle type, vehicle color, entry/exit time, pollution level, and damage level as well as to recognize the vehicle including the vehicle number, a plurality of cameras (not the conventional loop coil method) 120) is newly applied to acquire a vehicle image by photographing the vehicle from the front, rear, both sides, and above the vehicle.

The IoT device 130 determines the presence of a vehicle by recognizing a vehicle including a license plate through a deep learning model in a frame of a vehicle image transmitted in real time from the camera 120 . That is, as shown in FIG. 2 , the recognized vehicle information is compared with reference vehicle information previously stored through learning in a cloud DB or a local DB connected through a communication network, and the presence or absence of a vehicle is determined according to whether the information matches.

The deep learning model may recognize a vehicle from an image by using a one-stage object detection model including SSD or YOLO. That is, vehicle recognition utilizes a one-stage deep learning model (YOLO, SSD), and builds and utilizes this deep learning model with frameworks such as tensorflow and caffe2 for IoT devices. Accordingly, the vehicle can be recognized at a super speed (for example, within 0.5 seconds).

The IoT device 130 may be implemented using a single board computer (eg, Raspberry Pi) and hardware for IoT. frame can be read. An example of a library that can read vehicle images and process every frame is opencv.

In addition, the IoT device 130 has a dual-core or higher CPU, and may include a program for processing an image and a program for driving a deep learning model, wherein the image processing program and the deep learning model driving program have different threads As a multi-threaded program running in , it is possible to minimize the vehicle recognition latency in the image.

The IoT device 130 may comprehensively determine the presence or absence of a vehicle based on the time and location at which the vehicle is recognized. It can be judged that there is

On the other hand, the IoT device 130, like a method of recognizing a vehicle including a license plate through a deep learning model in the frame of the vehicle image to determine whether there is a vehicle, the vehicle model, vehicle color, Vehicle condition information including pollution degree and damage degree may be additionally collected.

The IoT device 130 may pre-process the vehicle image acquired from the camera 120 and extract the resultant image of the vehicle image by substituting the pre-processed vehicle image into an artificial neural network.

For reference, the pre-processing is to reduce the image to a predetermined size after accessing each individual target object image frame acquired in order to efficiently process the image. The artificial neural network is a machine learning technique implemented by focusing on the neural network of a human or animal brain, and is one of detailed methodologies of machine learning, in the form of a network in which several neurons, which are nerve cells, are connected. It is classified into several types according to its structure and function, and the most common artificial neural network is a multilayer perceptron with a plurality of hidden layers between one input layer and an output layer.

As shown in FIG. 8 , the IoT device 130 compares and analyzes the preprocessed vehicle image and the resultant image of the extracted vehicle image to determine the degree of pollution of the vehicle and the error between the vehicle image and the resultant image. It is possible to determine the degree of damage (S110 ~ S130). That is, by comparing the vehicle image with the resultant image of the vehicle image, when an error of the resultant image exceeds a threshold, it is determined that the vehicle is contaminated or damaged.

To explain further, the error is created by using the pixel-to-pixel difference between the original image and the resulting image. Representative error measurement methods include the l1 norm and l2 norm methods, as well as various other methods.

This is possible. (x is the original image, x' is the result image, the left equation is l1 norm, the right equation is l2 norm)

For example, since the result of decoding is a process of making an approximation of the original image, if an autoencoder with only general vehicle characteristics decodes an image with pollution and damage, it is necessary to properly approximate the degree of pollution and damage to decode. The more difficult it is, the greater the error may be.

The IoT device 130 recognizes and classifies as a vehicle through vehicle image analysis, recognizes a specific location of the classified vehicle, classifies the vehicle into pixel units through segmentation, and extracts the resulting image for the vehicle image. . Alternatively, by applying an autoencoder, the vehicle image is encoded by the encoder to extract features, the extracted features are compressed, and then the approximate result image of the vehicle image can be extracted through decoding by the decoder. there is. Encoding is the process of extracting features from the target object image, and decoding is the process of making an approximation of the original image.

For reference, the autoencoder is a kind of machine learning method and belongs to unsupervised learning. It uses a neural network algorithm to learn with the goal of making the output value of an input through the neural network as similar to the input value as possible. At this time, if the number of neurons in the neural network is greater than or equal to the dimension of the input value, the meaning of learning is lost. Because it is enough to take the input value as it is and export it. That is, to have meaning, the number of neurons must be smaller than the dimension of the input value. As a result of this learning, it is possible to obtain the effect of compression that can restore the original value with a smaller number of values.

The IoT device 130 has a CPU having a dual core or more, and may include a program for processing an image and a program for driving a deep learning model, wherein the image processing program and the deep learning model driving program are each driven in different threads It is a multi-threaded program that can minimize the latency of vehicle recognition in the image.

The circuit breaker 140 is installed at the entrance of each parking lot to which the vehicle enters and controls the vehicle entry.

The display device 150 provides various convenient information to a driver or a parking lot user.

The environmental information collector 160 may be mounted on the circuit breaker 140, and collects environmental information such as temperature, humidity, fine dust and noise level information of the parking lot.

The central control server 170 is a kind of control tower, and the breaker 140 controls the vehicle according to the vehicle state determined by the IoT device 130 in conjunction with the parking information transmitted in real time from the parking information acquisition device 110 . Control to control access. That is, when it is determined that there is a vehicle in the IoT device 130 while checking the parking information of each parking lot in real time, the breaker 140 is controlled so that the vehicle can pass.

In addition, the central control server 170 provides various convenience information that the display device 150 helps drivers or parking lot users or temperature, humidity, fine dust and noise level information collected by the environmental information collector 160 control to do

The central control server 170 includes government offices, gas stations (charging stations), e-commerce related companies, advertisement companies, vehicle manufacturers, sales companies, car washes, repair shops, and insurance companies based on the vehicle status information collected from the IoT device 130 . A service related to the vehicle may be relayed to the server 180 of the cooperative organization.

Specifically, based on the vehicle number, which is the vehicle status information collected from the IoT device 130, and the entry/exit information, the Ministry of Justice, the National Police Agency, the National Tax Service, local governments, related IT companies, traction agencies, etc. to the server 180 A vehicle location information service may be provided.

In addition, based on the vehicle number, which is the vehicle status information collected from the IoT device 130, the vehicle statistical information service is provided to the server 180 of cooperative organizations such as gas stations, electric charging stations, gas charging stations, hydrogen charging stations, etc. can

In addition, the lifestyle information service can be provided to the server 180 of cooperative organizations such as e-commerce related companies and advertisement companies based on vehicle number and entry/exit information, which are vehicle state information collected from the IoT device 130 . .

In addition, it is possible to provide a vehicle sales information providing service to the server 180 of cooperative organizations such as automobile manufacturers and sales companies based on vehicle number, vehicle model, and vehicle color information, which are vehicle state information collected from the IoT device 130 .

In addition, it is possible to provide a car wash service to the server 180 of a cooperative organization such as a business trip car wash and a nearby car wash by comparing the pollution level of the vehicle, which is the vehicle state information collected from the IoT device 130 .

In addition, by comparing the degree of vehicle damage, which is vehicle state information collected by the IoT device 130 , before and after, maintenance services may be provided to the server 180 of a cooperative organization such as on-site maintenance and a nearby repair shop.

In addition, an insurance service may be provided to the server 180 of a cooperative organization such as an insurance company by comparing the degree of vehicle damage, which is vehicle state information collected from the IoT device 130 before and after.

The central control server 170 may determine whether or not the vehicle is in and out based on the time the vehicle has been present for a specific time, so that the breaker 140 can control access to the vehicle. Through this, it is possible to determine whether the vehicle is speeding or not, and it is also possible to perform camera number recognition based on the presence or absence of vehicle license plate recognition.

On the other hand, although the artificial intelligence-based parking lot sharing system according to the present invention has been described according to a limited embodiment, the scope of the present invention is not limited to a specific embodiment, and it is obvious to those of ordinary skill in the art in relation to the present invention. Various alternatives, modifications, and changes can be implemented within the scope.

Accordingly, the embodiments disclosed in the present invention and the accompanying drawings are for explaining, not limiting, the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments and the accompanying drawings . 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.

110: parking information acquisition device 120: camera
130: IoT device 140: breaker
150: display device 160: environmental information collector
170: central control server 180: cooperative organization server

Claims (8)

A system for sharing a parking lot installed in a plurality of places, comprising: a parking information acquisition device 110 installed in each parking lot to acquire parking information related to a vehicle parking state; a camera 120 installed at the entrance of each parking lot to photograph a vehicle to obtain a vehicle image; In the frame of the vehicle image transmitted in real time from the camera 120, the vehicle including the license plate is recognized through a deep learning model, and the recognized vehicle information is compared with the reference vehicle information stored in advance through learning in the cloud DB or local DB. an IoT device 130 that determines whether or not a vehicle is present according to whether the information matches or not; a circuit breaker 140 installed at the entrance of each parking lot to control vehicle entry; A central control server ( 170) in an artificial intelligence-based parking lot sharing system, including;
The IoT device 130 recognizes and classifies as a vehicle through vehicle image analysis, applies an autoencoder that learns an original image without contamination or damage of the classified vehicle, and sends the classified vehicle image to the encoder. After encoding and extracting features and compressing the extracted features, extracting an approximate image that is an approximation of the original image through decoding by a decoder to further collect information on the degree of pollution and damage of the vehicle;
It further includes a display device 150, wherein the central control server 170 controls the display device 150 to provide various convenient information;
Further comprising an environmental information collector 160 for collecting parking lot temperature, humidity, fine dust and noise level information, wherein the central control server 170 is the temperature collected by the display device 150 from the environmental information collector 160 , control to provide information on humidity, fine dust and noise levels;
The central control server 170 compares the pollution level of the vehicle, which is vehicle state information collected from the IoT device 130, before and after, and provides a car wash service to the cooperative organization server 180 including a business trip car wash and a nearby car wash;
The central control server 170 compares the degree of damage of the vehicle, which is the vehicle state information collected from the IoT device 130, before and after, and provides maintenance services to the cooperative organization server 180 including on-site maintenance and nearby repair shops;
The central control server 170 compares the degree of vehicle damage, which is vehicle state information collected from the IoT device 130, before and after, and provides an insurance service to the cooperative organization server 180 including the insurance company;
The central control server 170 determines whether or not the vehicle is in and out through the time the vehicle has been present for a specific time, and the breaker 140 controls the vehicle to access and control the parking lot sharing system. delete delete delete delete delete delete delete

Images