KR102633444B1 - AI-based parking area unmanned surveillance platform - Google Patents

Abstract

The present invention relates to an AI-based unmanned monitoring platform for parking areas, and more specifically, to an AI-based unmanned monitoring platform for parking areas for controlling vehicles that obstruct parking in disabled parking areas while driving autonomously based on AI.
In addition, a driving unit capable of autonomous driving through recognition of the surrounding environment, a photographing unit for photographing a vehicle subject to surveillance by being transported through the driving unit, and a control unit for calculating whether or not a subject is subject to enforcement based on the image captured through the photographing unit. When it is determined that it is subject to enforcement through the control unit, it includes an output unit for outputting a sticker with enforcement information printed based on the recognized image and a multi-joint unit for attaching the sticker output through the output unit to the vehicle subject to enforcement. It is characterized by

Description

AI-based parking area unmanned surveillance platform

The present invention relates to an AI-based unmanned parking area monitoring platform, and more specifically, to an AI-based parking area unmanned monitoring platform for controlling vehicles that obstruct parking in disabled parking spaces while driving autonomously based on AI.

In general, parking areas for the disabled are wider than general parking areas and are installed closer to the main entrance of the building. The number of parking areas for the disabled varies depending on the size of the parking lot, but usually has one to several parking spaces. It is common to have

In addition, in parking lots attached to buildings and automobile-related facilities, parking spaces exclusively for the disabled must be installed at the location closest to the building's entrance or elevating facility for the disabled that allows access for the disabled.

The fact that the width of the parking area exclusively for the disabled is wider than that of the general parking area can be said to be a parking area that emphasizes convenience and accessibility for the disabled, as many users of the parking area exclusively for the disabled are wheelchair users.

However, these advantages of disabled-only parking spaces are being abused by regular car users.

In other words, regular car users often use the disabled-only parking area because it is wider than the general parking area and is closer to the main entrance and the parking time is shorter.

Such violations by regular vehicle users of parking spaces reserved for the disabled make it impossible for the disabled who desperately need the space to use the space.

Parking spaces for the disabled are restricted to only authorized vehicles, and the relevant organizations issue a sign to identify the authorized vehicles, for example, a parking permit for the disabled, thereby identifying vehicles that can use the parking space for the disabled. I am checking.

A parking system for the disabled is in operation for the purpose of protecting the socially vulnerable, but parking spaces for the disabled are located sporadically in numerous locations, making it realistically impossible to crack down on violations of laws related to parking for the disabled.

Accordingly, a system that can automatically crack down on vehicles illegally parked on disabled parking spaces has been developed, such as Korean Patent No. 10-2431894, “Unmanned enforcement method and system for disabled parking spaces,” but each disabled parking space is different. There was a problem with having to install the system.

Korean Patent No. 10-2431894 “Unmanned enforcement method and system for parking spaces for the disabled”

The purpose of the present invention is to solve the problems described above, and to provide an AI-based unmanned monitoring platform for parking spaces to determine and crack down on parking violations in disabled parking spaces located in multiple locations.

In order to achieve the above purpose, the AI-based parking area unmanned monitoring platform according to the present invention has a driving unit capable of autonomous driving through recognition of the surrounding environment, and a photography unit for photographing the vehicle to be monitored by being transported through the driving unit and the photography unit. A control unit for calculating whether a crackdown target is based on a captured image, and an output unit for outputting a sticker with crackdown information printed based on the recognized image when it is determined to be a crackdown target through the control unit. It is characterized by including a multi-joint unit for attaching a sticker to a vehicle subject to enforcement.

In addition, the control unit transmits data for inquiring whether a person is subject to a crackdown to the server through wireless communication, and calculates whether a person is a target of a crackdown based on the received data based on a database stored in the server.

In addition, it further includes a mounting portion to which a plurality of tools disposed on the upper portion of the traveling portion are fixed, and the multi-joint unit is characterized in that one of the tools mounted on the mounting portion can be selectively picked up and exchanged.

In addition, the photographing unit is configured to control the direction in which light is irradiated, and captures a plurality of images while controlling the irradiation direction of light, and the control unit merges recognizable parts of the plurality of images to create one overall image. It is characterized by being configured to generate and calculate whether or not it is subject to crackdown based on the generated image.

As described above, according to the AI-based parking area unmanned monitoring platform according to the present invention, the system moves through parking spaces for the disabled located in multiple locations through autonomous driving through recognition of the surrounding environment, from determining parking violations to crackdown using a multi-joint unit. It even has the effect of allowing stickers to be attached.

Figure 1 is a perspective view showing an AI-based unmanned parking area monitoring platform according to the present invention.
Figure 2 is a diagram showing a vehicle identification operation using the AI-based unmanned parking area monitoring platform according to the present invention.
Figure 3 is a diagram showing the operation of identifying disabled people's signs using the AI-based parking area unmanned monitoring platform according to the present invention.
Figure 4 is a diagram showing a sticker output operation using the output unit of the AI-based parking area unmanned monitoring platform according to the present invention.
Figure 5 is a diagram showing a sticker attachment operation using the AI-based unmanned parking area monitoring platform according to the present invention.
Figure 6 is a diagram showing an example of image generation by the AI-based parking area unmanned surveillance platform according to the present invention.
Figure 7 is a diagram showing the lighting unit of the AI-based parking area unmanned monitoring platform according to the present invention.
Figure 8 is a diagram showing an embodiment of an AI learning algorithm through the control unit of the AI-based parking area unmanned monitoring platform according to the present invention.
Figure 9 is a diagram showing an embodiment of an AI-based object recognition algorithm through the control unit of the AI-based parking area unmanned monitoring platform according to the present invention.

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed in this specification are merely illustrative for the purpose of explaining the embodiments according to the concept of the present invention, and the embodiments according to the concept of the present invention are It may be implemented in various forms and is not limited to the embodiments described herein.

Since the embodiments according to the concept of the present invention can make various changes and have various forms, the embodiments will be illustrated in the drawings and described in detail in this specification. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes all changes, equivalents, or substitutes included in the spirit and technical scope of the present invention.

Hereinafter, preferred embodiments of the present invention will be described with reference to the attached drawings.

Figure 1 is a perspective view showing an AI-based parking area unmanned monitoring platform according to the present invention, Figure 2 is a diagram showing a vehicle identification operation using the AI-based parking area unmanned monitoring platform according to the present invention, and Figure 3 is a diagram showing a vehicle identification operation using the AI-based parking area unmanned monitoring platform according to the present invention. It is a diagram showing the operation of identifying disabled signs using an AI-based parking area unmanned monitoring platform according to , and Figure 4 is a diagram showing a sticker output operation using the output unit of the AI-based parking area unmanned monitoring platform according to the present invention, and Figure 5 is a diagram showing a sticker attachment operation using the AI-based parking area unmanned monitoring platform according to the present invention, Figure 6 is a diagram showing an example of image creation by the AI-based parking area unmanned monitoring platform according to the present invention, and Figure 7 is It is a diagram showing the lighting unit of the AI-based parking area unmanned monitoring platform according to the present invention, and Figure 8 is a diagram showing an embodiment of an AI learning algorithm through the control unit of the AI-based parking area unmanned monitoring platform according to the present invention. 9 is a diagram showing an example of an AI-based object recognition algorithm through the control unit of the AI-based parking area unmanned surveillance platform according to the present invention.

As shown in Figures 1 to 5, the AI-based parking area unmanned monitoring platform according to the present invention includes a driving unit 1 capable of autonomous driving through recognition of the surrounding environment using a plurality of driving cameras 3 and sensors, and A photographing unit (2) for photographing a vehicle subject to surveillance transported through the driving unit (1), a control unit (not shown) for calculating whether or not a vehicle is subject to enforcement based on the image captured through the photographing unit (2), and the If it is determined to be a crackdown target through the control unit, an output unit 4 for outputting a sticker 13 with crackdown information printed based on the information of the recognized image, and a sticker 13 output through the output unit 4. It is configured to include a multi-joint unit (5) for attaching to the vehicle subject to enforcement.

In particular, the driving unit 1 can move autonomously within a preset area of the parking lot, so unlike a fixed control device, there is no limit to the control area, and it is possible to immediately respond to the increase and movement of parking spaces for the disabled. .

In addition, the control unit transmits data to check whether or not a crackdown target exists to the server through wireless communication, and calculates whether or not the crackdown target is based on the received data based on a database stored in the server, or calculates whether or not the crackdown target is subject to the server. And, it may be configured to receive information about whether or not it is subject to crackdown from the server.

In addition, it may be configured to further include a mounting portion 12 to which a plurality of tools 11 disposed on the upper portion of the traveling portion 1 are fixed, and the multi-joint unit 5 is mounted on the mounting portion 12. Any one of the mounted tools 11 can be selectively picked up and exchanged.

At this time, the photographing unit 2 may be fixed to the multi-joint unit 5 or may be implemented as one of the tools 11 attached to the multi-joint unit 5.

More specifically, an adapter 5a is formed on the multi-joint unit 5 so that any one of the plurality of tools 11 mounted on the mounting portion 12 can be coupled, and a camera tool for vehicle recognition ( 11a), a camera tool for checking disabled person signs (11b), and a tool for attaching stickers (11c), etc. depending on the purpose.

For example, when the driving unit 1 arrives at the disabled parking space, the articulated unit 5 picks up the camera tool 11a for vehicle recognition and uses it to check information on the vehicle to be monitored, as shown in FIG. 2. It may be carried out to take an image.

After taking an image to check vehicle information, the multi-joint unit 5 places the camera tool 11a for vehicle recognition on the holder 12 and then replaces it by picking up the camera tool 11b for checking the disabled sign. As shown in Figure 3, the disabled sign of the vehicle subject to surveillance is photographed and an image is taken to confirm the disabled sign information.

At this time, among the tools 11, the camera tool 11a for vehicle recognition and the camera tool 11b for checking disabled signs should be viewed as a capturing unit 2 for taking images, and the camera type (tool) can be replaced according to the purpose. Alternatively, image shooting to check information on the vehicle to be monitored and image shooting to check information on the disabled sign of the vehicle to be monitored may be performed through the photographing unit 2 fixed to the multi-joint unit 5.

In this way, if the crackdown target is confirmed based on the information on the vehicle subject to surveillance and the information on the disabled sign based on the captured image, the sticker 13 with the crackdown information printed through the output unit 4 as shown in FIG. 4 ) is output, and the multi-joint unit 5 picks up the sticker 13 printed through the sticker attachment tool 11c and then attaches the sticker 13 to the vehicle subject to enforcement as shown in FIG. do.

At this time, the sticker attachment tool 11c may be used in a form that holds the sticker 13 or in a form that adsorbs it through vacuum adsorption.

Figure 6 shows that based on the image taken to check the disabled sign information of the vehicle subject to surveillance, the control unit merges the recognizable parts of the plurality of images to create one overall image, and determines whether or not it is subject to crackdown based on the generated image. This illustrates an embodiment of the calculation. In order to solve the problem of unrecognition due to images obscured by reflected light, backlight, tinting, or foreign substances, multiple images are taken from various angles when it is difficult to recognize information required for enforcement at the initial shooting location. Afterwards, the recognizable parts of the plurality of images are merged to create one overall image, and whether or not the target is subject to crackdown is calculated based on the generated image.

As another embodiment, as shown in FIG. 7, the photographing unit 2 may be configured to further include a lighting unit 30 configured to control the direction in which light is irradiated.

More specifically, a plurality of light sources are formed in one group (31, 32, 33, ...), and the direction of light irradiation is controlled by blinking any one or more of the plurality of groups to create a plurality of images. While taking pictures, the control unit may merge recognizable parts of a plurality of images to create one overall image and calculate whether or not the subject is subject to enforcement based on the generated image.

Alternatively, the light source constituting the group consists of visible light LEDs (31a, 32a, 33a, ...) and infrared LEDs (31b, 32b, 33b, ...), and is included in one or more of the plurality of groups. After flashing the visible light LEDs (31a, 32a, 33a, ...), the infrared LEDs (31b, 32b, 33b, ...) included in one or more of the plurality of groups are flashed. You can also take images.

In addition, the method of blinking the visible light LEDs (31a, 32a, 33a, ...) and the infrared LEDs (31b, 32b, 33b, ...) of one or more groups among the plurality of groups is set at a preset interval. It is desirable to carry out the method by lighting the next group.

For example, it is preferable to blink the first group of infrared LEDs 31b and then the third group of infrared LEDs 33b.

In addition, the control unit of the AI-based parking area unmanned surveillance platform according to the present invention divides the image captured through the photographing unit into a plurality of layers as shown in A in Figure 8, preprocesses each layer, and then performs deep learning learning for each layer iteratively. Accuracy is improved through the algorithm, and accuracy is improved by verifying the learning results as shown in B.

In addition, external environmental factors are removed from images captured with an object recognition algorithm using the YOLO network structure shown in A of FIG. 9 or the CNN network structure shown in B, and the accuracy of recognition of disabled areas and surveillance target vehicles is improved. It may be configured to increase and more accurately identify disability sign information.

Alternatively, for recognition of disabled areas and vehicles subject to surveillance, the target vehicle is identified using an object recognition algorithm using the YOLO network structure shown in A, and for identification of disabled sign information in a captured image, the CNN network structure shown in B is used. In order to further increase the identification accuracy for a specific location, the algorithm may be selectively optimized and reflected depending on the identification target.

As described above, the present invention has been described focusing on preferred embodiments with reference to the accompanying drawings, but it is clear to those skilled in the art that many obvious modifications can be made without departing from the scope of the present invention from this description. Accordingly, the scope of the present invention should be construed by the appended claims to include examples of many such modifications.

1: Running part
2: Filming Department
3: Driving camera
4: output unit
5: Multi-joint unit
11: Tool
12: Holder
13: Sticker
30: lighting unit
31: 1st group
32: 2nd group
33: 3rd group
34: 4th group

Claims (4)

A driving unit capable of autonomous driving through recognition of the surrounding environment;
a photographing unit transported through the driving unit to photograph a vehicle subject to surveillance;
a control unit for calculating whether or not there is a crackdown target based on the image captured through the photographing unit;
an output unit for outputting a sticker with enforcement information printed based on the recognized image when it is determined that the control unit is subject to enforcement;
Characterized in that it includes a multi-joint unit for attaching the sticker output through the output unit to the vehicle subject to enforcement.
AI-based parking area unmanned surveillance platform.
According to clause 1,
The control unit
Characterized by transmitting data to check whether a crackdown target exists to the server through wireless communication, and calculating whether a crackdown target is based on the received data based on a database stored on the server.
AI-based parking area unmanned surveillance platform.
According to clause 1,
It further includes a mounting portion to which a plurality of tools disposed on an upper portion of the traveling portion are fixed,
The multi-joint unit is characterized in that it can be exchanged by selectively picking up any one of the tools mounted on the holder.
AI-based parking area unmanned surveillance platform.
According to clause 1,
The filming department
It is configured to control the direction of irradiation of light,
Takes multiple images while controlling the direction of light irradiation,
The control unit is configured to merge recognizable parts of a plurality of images to create one overall image and calculate whether or not it is subject to crackdown based on the generated image.
AI-based parking area unmanned surveillance platform.