KR20210118275A - System and method for managing diligence of worker - Google Patents

Abstract

An object of the present invention is to provide a vehicle traffic image measurement system that automatically aggregates the amount of rotating traffic for each intersection, vehicle type, vehicle queue, and the like through a deep learning algorithm, analyzes the congestion level of the intersection in real time, and controls an optimal traffic signal, and a measurement method thereof. In order to achieve the above object, a vehicle traffic image measurement system according to the present invention includes a vehicle traffic measurement image sensing device for analyzing an intersection status by detecting and classifying vehicles for each lane of an intersection in real time using AI deep learning; and an AI intersection integrated management platform system that controls or manages the analyzed intersection status.

Description

Vehicle traffic image measurement system and its measurement method {SYSTEM AND METHOD FOR MANAGING DILIGENCE OF WORKER}

The present invention relates to a vehicle traffic image measurement system and a method for measuring the same, and more particularly, to an artificial intelligence (AI)-based vehicle traffic image measurement system and a method for measuring the same.

Recently, many social problems are emerging due to an increase in traffic volume, an increase in vehicles, and improvement in vehicle performance.

Accordingly, various information collection devices for measuring traffic information or monitoring illegal activities of vehicles are being installed in various places on the road.

The traffic volume survey is a survey that measures the number of passing vehicles using the road by type, direction, and time period.

Traffic volume measurement data is widely used in road traffic planning and design, road management operation, etc.

In such a road traffic measurement method, there is a traditional method of measuring the traffic volume using a mechanical counting device while directly checking the vehicle passing by the investigator with the naked eye.

However, this traditional method has disadvantages in that it takes a lot of time and money because it has to manually measure the traffic volume, and in the case of statistical work, all work must be done manually.

Accordingly, the need for a system that automatically performs road traffic measurement and traffic volume statistics is gradually increasing.

Meanwhile, as the number of vehicles gradually increases and traffic problems become more serious, the need to aggregate the traffic volume for each vehicle type on roads in various regions rather than simply counting the traffic volume of vehicles for use in road planning and road operation is also increasing. have.

To this end, it is necessary not only to collect the traffic volume measured for each time period and for each vehicle type, but also a system capable of collecting the measured traffic volume for each road in an effective way is required.

Republic of Korea Patent Publication No. 2018-0107757 (published on October 02, 2018)

It is an object of the present invention to solve the problems of the prior art as described above by automatically counting the traffic volume, vehicle type, queue length, etc. for each intersection approach through a deep learning algorithm, as well as analyzing the congestion level of the intersection in real time, and , to provide a vehicle traffic image measurement system and method for optimally controlling traffic signals.

In order to achieve the above object, the vehicle traffic image measurement system according to the present invention includes a vehicle traffic measurement image sensing device for analyzing the intersection status by detecting and classifying vehicles by lane of the intersection in real time using AI deep learning; and an AI intersection integrated management platform system that controls or manages the analyzed intersection status.

In addition, in the vehicle traffic image measurement system according to the present invention, the vehicle traffic measurement image sensing device includes: an image data input unit for receiving image data of a vehicle passing through an intersection; a vehicle detection unit for detecting a vehicle classified by vehicle type from the image data received using the previously learned deep learning network; a detection vehicle labeling unit for labeling the vehicle type for the detected vehicle; a vehicle re-learning unit for re-learning the labeled and detected vehicle; and a judging unit that determines whether or not vehicle model recognition for the detected vehicle on which re-learning has been made is performed.

In addition, in the vehicle traffic image measurement system according to the present invention, the detection vehicle tracking unit for tracking the detected vehicle re-learned by the vehicle re-learning unit; and an alarm generating unit that generates an alarm when the tracked and detected vehicle is stopped in a specific area for a predetermined time.

In addition, in the vehicle traffic image measurement system according to the present invention, the determination unit may include: a vehicle recognition rate determination unit for determining whether a vehicle model is recognized; a tracked vehicle line passing determination unit which determines whether the tracked vehicle has passed through a predetermined line; and a tracked vehicle stop determining unit that determines whether the tracked vehicle is stopped in a specific area.

In addition, in the vehicle traffic image measurement system according to the present invention, the vehicle traffic image measurement apparatus is characterized in that it acquires data on the number of vehicles passing through the intersection per hour by counting vehicles passing through a predetermined line.

In addition, in the vehicle traffic image measurement system according to the present invention, the length of the queue in each direction of the intersection is analyzed in real time through the obtained data, and the signal control of the traffic light is performed to preferentially solve the congestion on the most congested road. characterized in that

In addition, in the vehicle traffic image measurement system according to the present invention, a decision-making system for making a decision based on the analyzed intersection status; and a related institution that receives the analyzed intersection status from the intersection status when an event occurs.

In order to achieve the above object, a vehicle traffic image measurement method according to the present invention includes the steps of receiving image data for a vehicle passing through an intersection by an image data input unit (S1); Detecting, by a vehicle detection unit, a vehicle classified by vehicle type from the input image data using the previously learned deep learning network (S2) (S3); Labeling the vehicle model for the detected vehicle by the detection vehicle labeling unit (S4); re-learning the labeled detected vehicle by the vehicle re-learning unit (S5); a step (S6) of determining, by a determination unit, whether or not vehicle model recognition has been made with respect to the detected vehicle for which re-learning has been made; and if the vehicle model is not recognized by the determination unit, it returns to step S3, and when the vehicle model is recognized by the determination unit, the previously learned deep learning network for vehicle detection is updated using the re-learned deep learning network. It is characterized in that it comprises a step (S7).

In order to achieve the above object, a vehicle traffic image measurement method according to the present invention includes the steps of receiving image data for a vehicle passing through an intersection by an image data input unit (S10); using the deep learning network re-learned by the vehicle re-learning unit (S20) to detect the vehicle classified by vehicle type from the input image data by the vehicle detection unit (S30); tracking the detected vehicle by the geomchi vehicle tracking unit (S40); determining whether or not the tracked vehicle has passed through a predetermined line (S50); And if the passage of the predetermined line of the tracked vehicle is not confirmed, return to step S40, and when the passage of the predetermined line of the tracked vehicle is confirmed, a step (S60) of increasing the statistic of the number of vehicles passing through the intersection per hour (S60). characterized.

In order to achieve the above object, a vehicle traffic image measurement method according to the present invention includes the steps of receiving image data for a vehicle passing through an intersection by an image data input unit (S100); using the deep learning network re-learned by the vehicle re-learning unit (S200) to detect the vehicle classified by vehicle type from the input image data by the vehicle detection unit (S300); tracking the detected vehicle by a geomchi vehicle tracking unit (S400); determining whether the tracked vehicle is stopped in a specific area by the determination unit (S500); And if the stop in the specific area of the tracked vehicle is not confirmed, the process returns to step S400, and when the stop is confirmed in the specific area of the tracked vehicle, an alarm is generated by the alarm generating unit (S600); do it with

Specific details of other embodiments are included in "Details for carrying out the invention" and the accompanying "drawings".

Advantages and/or features of the present invention, and methods for achieving them, will become apparent with reference to the various embodiments described below in detail in conjunction with the accompanying drawings.

However, the present invention is not limited to the configuration of each embodiment disclosed below, but may also be implemented in a variety of different forms, and each embodiment disclosed herein only makes the disclosure of the present invention complete, It is provided to fully inform those of ordinary skill in the art to which the scope of the present invention belongs, and it should be understood that the present invention is only defined by the scope of each of the claims.

According to the present invention, there is an effect of automatically counting the traffic volume, vehicle type, queue length, etc. for each intersection approach through a deep learning algorithm, as well as analyzing the congestion level of the intersection in real time and controlling the optimal traffic signal. .

1 is a system configuration diagram showing the configuration of a vehicle traffic image measurement system according to the present invention.
Figure 2 is a view showing the components of the vehicle traffic image measurement system according to the present invention.
Figure 3 is a view showing the function of the vehicle traffic image measurement system according to the present invention.
4 is a view showing the configuration of a vehicle traffic measurement image sensing device of the vehicle traffic amount image measurement system according to the present invention.
5 is a view showing a vehicle traffic measurement image sensing device of the vehicle traffic image measurement system according to the present invention.
6 is a flowchart showing the flow of AI deep learning learning in the vehicle traffic image measurement method according to the present invention.
7 is a flowchart illustrating a processing flow of vehicle traffic in the method for measuring vehicle traffic image according to the present invention.
8 is a flowchart illustrating a flow of vehicle congestion alarm processing in the vehicle traffic image measurement method according to the present invention.

Before describing the present invention in detail, the terms or words used herein should not be construed as being unconditionally limited to their ordinary or dictionary meanings, and in order for the inventor of the present invention to describe his invention in the best way It should be understood that the concepts of various terms can be appropriately defined and used, and further, these terms or words should be interpreted as meanings and concepts consistent with the technical idea of the present invention.

That is, the terms used herein are only used to describe preferred embodiments of the present invention, and are not used for the purpose of specifically limiting the content of the present invention, and these terms represent various possibilities of the present invention. It should be understood that the term has been defined with consideration in mind.

Also, in the present specification, it should be understood that, unless the context clearly indicates otherwise, the expression in the singular may include a plurality of expressions, and even if it is similarly expressed in plural, it should be understood that the meaning of the singular may be included. .

In the case where it is stated throughout this specification that a component "includes" another component, it does not exclude any other component, but further includes any other component unless otherwise stated. It could mean that you can.

Furthermore, when it is described that a component is "exists in or is connected to" of another component, this component may be directly connected to or installed in contact with another component, and a certain It may be installed spaced apart by a distance, and in the case of being installed spaced apart by a certain distance, a third component or means for fixing or connecting the corresponding component to another component may exist, and now It should be noted that the description of the components or means of 3 may be omitted.

On the other hand, when it is described that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the third element or means does not exist.

Similarly, other expressions describing the relationship between components, such as "between" and "immediately between", or "neighboring to" and "directly adjacent to", have the same meaning. should be interpreted as

In addition, if terms such as "one side", "the other side", "one side", "other side", "first", "second" are used in this specification, with respect to one component, this one component is It is used to be clearly distinguished from other components, and it should be understood that the meaning of the component is not limitedly used by such terms.

In addition, in the present specification, terms related to positions such as "upper", "lower", "left", and "right", if used, should be understood as indicating a relative position in the drawing with respect to the corresponding component, Unless an absolute position is specified with respect to their position, these position-related terms should not be construed as referring to an absolute position.

In addition, in this specification, in specifying the reference numerals for each component of each drawing, the same component has the same reference number even if the component is shown in different drawings, that is, the same reference is made throughout the specification. Symbols indicate identical components.

In the drawings attached to this specification, the size, position, coupling relationship, etc. of each component constituting the present invention are partially exaggerated, reduced, or omitted for convenience of explanation or in order to sufficiently clearly convey the spirit of the present invention. may be described, and therefore the proportion or scale may not be exact.

In addition, in the following, in describing the present invention, a detailed description of a configuration determined that may unnecessarily obscure the subject matter of the present invention, for example, a detailed description of a known technology including the prior art may be omitted.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the related drawings.

1 is a system configuration diagram illustrating the configuration of a vehicle traffic image measurement system according to the present invention, and FIG. 2 is a diagram illustrating components of the vehicle traffic amount image measurement system according to the present invention.

1 and 2 , the vehicle traffic image measurement system 1000 according to the present invention includes a vehicle traffic measurement image detection device 100 , an AI intersection integrated management platform system 200 , and a decision-making system 300 . includes

The vehicle traffic image measurement system 1000 automatically aggregates the rotational traffic for each intersection approach, the vehicle type, and the length of the queue through a deep learning algorithm, and at the same time, based on the Road Capacity Handbook (KHCM) of the Ministry of Land, Infrastructure and Transport, It is a smart traffic signal control system that analyzes congestion levels in real time and controls optimal traffic signals.

The vehicle traffic measurement image sensing apparatus 100 performs a role of analyzing the intersection status by detecting and classifying vehicles by lanes of the intersection in real time using AI deep learning.

A plurality of vehicle traffic measurement image sensing apparatuses 100 may be installed at each intersection.

For example, in the case of a three-way intersection, three are installed, in the case of a crossroad, four are installed, and in the case of a five-way intersection, five are preferably installed.

Thereby, all directions of the intersection can be detected, and all lanes of the intersection can be detected.

In addition, the degree of ±100m (entry direction, passing direction) is detected for each intersection direction.

In addition, the vehicle traffic measurement image sensing device 100 is a built-in CPU/GPU that enables high-speed operation through deep learning and operates without a PC.

The vehicle traffic measurement image sensing apparatus 100 may control a traffic light through a traffic signal controller.

That is, it is possible to control signals of traffic lights in real time.

For example, in the intelligent real-time signal control mode, danger detection and congestion detection are possible, real-time control, adaptive signal control, and fully sensitive signal control are possible.

In addition, in the safety signal control mode, TOD (Time Of Day) regular cycle control and Golden Pattern registration are possible.

In more detail, the vehicle traffic measurement image sensing apparatus 100 includes a plurality of vehicle detection AI camera modules for each intersection.

The vehicle detection AI camera module consists of a CCTV camera and an AI EDGE board (with high-performance CPU and GPU built-in, independent operation without a PC).

The vehicle traffic measurement image sensing apparatus 100 analyzes the intersection status by detecting and classifying vehicles by intersection lanes using AI deep learning.

In addition, in a form that can be mounted directly on a traffic light, there is no risk of damage unlike the existing buried loop detector, and the installation is simple and robust.

By installing a plurality of vehicle detection AI camera modules, it detects all lanes on the exit road and on the entrance road, and it is possible to detect a wide distance of ±100m (entry direction, passing direction).

It has a detection range that exceeds that of installing 64 existing buried loop detectors.

In addition, intelligent real-time signal control is possible.

Signal control of traffic lights is possible based on the length of the waiting queue of vehicles in each direction of the intersection.

That is, the most congested road congestion can be solved preferentially by analyzing the length of the queue in each direction in real time.

The AI intersection management platform system 200 serves to control or manage the analyzed intersection status.

That is, the AI intersection management platform system 200 receives intersection traffic statistics information from the vehicle traffic measurement image sensing device 100 to control or manage the intersection status.

In addition, the vehicle traffic image measurement system 1000 according to the present invention includes a decision-making system 300 and a related organization (not shown).

The decision-making system 300 serves to make a decision through the analyzed intersection status.

Relevant organizations such as the disaster situation room, police station, and fire department receive the analyzed intersection status when an event occurs from the intersection status.

In more detail, the AI intersection management platform system 200 collects data from the vehicle traffic measurement image detection device 100, transmits an event to relevant national agencies, and makes decisions about these events. We provide one-stop service.

3 is a view showing the function of the vehicle traffic image measurement system according to the present invention.

Referring to FIG. 3 , the vehicle traffic image measurement system 1000 according to the present invention provides a one-stop service in the order of an interface, message distribution, smart dashboard, situation propagation, and decision making.

In the interface, data is collected by the vehicle traffic measurement image sensing device 100, and in the message distribution, integrated data is distributed and processed to maximize resource efficiency, and in this embodiment, the vehicle traffic measurement image sensing device 100 has its function provides

The smart dashboard informs data in the form of a widget, and in this embodiment, the AI intersection management platform system 200 provides the function.

Circumstance propagation propagates the informational data to relevant organizations for immediate execution.

Decision-making is provided with a plan that can lead to a one-stop decision-making system of the control tower, and in this embodiment, the decision-making system 300 provides the function.

As such, the vehicle traffic image measurement system 1000 according to the present invention is flexibly compatible with the control protocol used by the existing local governments, and can integrate the intersection AI solution without changing the control software, and monitor the overall status and change the user A possible dashboard UI is provided.

4 is a diagram showing the configuration of a vehicle traffic measurement image sensing device of the vehicle traffic image measurement system according to the present invention, and FIG. 5 is a view showing the vehicle traffic measurement image detection device of the vehicle traffic amount image measurement system according to the present invention.

4 and 5 , in the vehicle traffic image measurement system 1000 according to the present invention, the vehicle traffic measurement image sensing apparatus 100 includes a pre-learned deep learning network 110 and an image data input unit 120 . And, the vehicle detection unit 130, the detection vehicle labeling unit 140, the vehicle re-learning unit 150, the determination unit 160, the re-learned deep learning network 170, and the detection vehicle tracking unit ( 180 , and an alarm generating unit 190 .

The image data input unit 120 receives image data of a vehicle passing through an intersection.

The vehicle detection unit 130 detects a vehicle classified by vehicle type from the input image data using the previously learned deep learning network 110 .

Here, the differentiated vehicle type may be a Car, Bus, Truck, or the like.

The detection vehicle labeling unit 140 labels the vehicle model of the detected vehicle.

In this case, the labeling may be processed manually or automatically.

The vehicle re-learning unit 150 re-learns the detected and labeled vehicle.

The determination unit 160 determines whether the vehicle model recognition of the detected vehicle for which the re-learning has been made has been performed.

Here, the determination unit 160 includes a vehicle recognition rate determination unit 161 , a tracking vehicle line passing determination unit 162 , and a tracking vehicle stop determination unit 163 .

The vehicle recognition rate determining unit 161 serves to determine whether a vehicle model is recognized.

The tracked vehicle line passing determination unit 162 determines whether the tracked vehicle has passed through a predetermined line.

The tracked vehicle stop determination unit 163 determines whether the tracked vehicle is stopped in a specific area.

The detected vehicle tracking unit 180 tracks the detected vehicle re-learned by the vehicle re-learning unit 150 .

The alarm generating unit 190 generates an alarm when the detected and tracked vehicle is stopped in a specific area for a predetermined time.

The vehicle traffic image measuring apparatus 100 configured as described above counts vehicles passing through a predetermined line to obtain data on the number of vehicles passing through the intersection per hour.

In addition, by analyzing the length of the queue in each direction of the intersection in real time through the obtained data, the signal control of the traffic light is performed to preferentially solve the congestion on the most congested road.

According to the vehicle traffic image measurement system 1000 according to the present invention as described above, it is possible to increase the traffic volume of the habitually congested intersection by 30% or more.

It can also reduce emissions of ultrafine dust (PM2.5) (30% reduction in vehicle waiting time at intersections).

In addition, it is possible to save daily vehicle energy (194 kL/day to 256 kL/day) by reducing the waiting time for vehicles at each intersection.

On the other hand, it is effective in preventing the main culprits of intersection accidents, such as cutting-in, illegal parking, and blocking of the vehicle in front.

6 is a flowchart showing the flow of AI deep learning learning in the vehicle traffic image measurement method according to the present invention.

Referring to FIG. 6 , in the vehicle traffic image measurement method according to the present invention, AI deep learning learning is performed as follows.

In the first step (S1), image data for a vehicle passing through the intersection is input by the image data input unit 120 .

In the second step (S2), the previously learned deep learning network 110 is used, and in the third step (S3), a vehicle classified by vehicle type is detected by the vehicle detection unit 130 from the input image data.

In the fourth step (S4), the vehicle model of the detected vehicle is labeled by the detection vehicle labeling unit 140 .

In the fifth step (S5), the detected vehicle labeled is relearned by the vehicle re-learning unit 150 .

In the sixth step (S6), it is determined by the determination unit 160 whether or not the vehicle model recognition of the detected vehicle on which the re-learning has been performed.

More preferably, it is determined by the vehicle recognition rate determination unit 161 of the determination unit 160 .

In the seventh step (S7), if the vehicle model recognition is not performed by the determination unit 160, it returns to the third step (S3 step), and when the vehicle model recognition is performed by the determination unit 160, the retrained deep learning network ( 170) is used to update the pre-trained deep learning network 110 for vehicle detection.

In the existing pre-trained deep learning network 110, the image data input by the image data input unit mounted on the upper part of the vehicle was formed into a data set and learned, so in the recognition of vehicle types such as Car, Bus, Truck, etc. There were many problems with errors.

However, in the present invention, the image data input unit is fixedly installed at a traffic light, etc., the vehicle model is detected using the previously learned deep learning network 110, and the detected vehicle is labeled and then relearned by the vehicle re-learning unit 150. Therefore, the acquired re-learned deep learning network 170 increases the recognition rate of vehicle types such as Car, Bus, and Truck.

That is, three types of cars, which are classified as Bus, Truck, and Car, passing through the intersection are recognized.

7 is a flowchart illustrating a processing flow of a vehicle traffic amount in the method for measuring a vehicle traffic amount image according to the present invention.

Referring to FIG. 7 , in the method for measuring vehicle traffic image according to the present invention, the processing of vehicle traffic is performed as follows.

In the first step ( S10 ), image data for a vehicle passing through an intersection is input by the image data input unit 120 .

In the second step (S20), the deep learning network 170 re-learned by the vehicle re-learning unit 150 is used, and in the third step (S30), the vehicle classified by vehicle type from the input image data is detected by the vehicle detection unit ( 130) is detected.

In the fourth step (S40), the detected vehicle is tracked by the geomchi vehicle tracking unit 180.

In the fifth step ( S50 ), it is determined by the determination unit 160 whether the tracked vehicle has passed a predetermined line.

More preferably, it is determined by the tracking vehicle line passing determination unit 162 of the determination unit 160 .

In the sixth step (S60), if the passage of the predetermined line of the tracked vehicle is not confirmed, it returns to the fourth step (step S40), and when the passage of the predetermined line of the tracked vehicle is confirmed, the number of vehicles passing through the intersection per hour statistics will increase.

As such, when it is confirmed that the tracked vehicle passes through a predetermined line, the statistics of the number of vehicles passing the intersection per hour increases, so the vehicle traffic measurement image sensing device 100 measures the number of vehicles driven per hour at the intersection, and AI intersection management It may be managed or controlled by the platform system 200 .

In addition, by analyzing the length of the queue in each direction of the intersection in real time through the obtained data on the statistics of the number of vehicles passing through the intersection per hour, it is possible to perform signal control of the traffic light to preferentially solve the congestion on the most congested road.

8 is a flowchart illustrating a flow of vehicle congestion alarm processing in a method for measuring an image of a vehicle traffic according to the present invention.

Referring to FIG. 8 , in the vehicle traffic image measurement method according to the present invention, vehicle congestion alarm processing is performed as follows.

In a first step ( S100 ), image data of a vehicle passing through an intersection is received by the image data input unit 120 .

In the second step (S200), the deep learning network 170 re-learned by the vehicle re-learning unit 150 is used, and in the third step (S300), the vehicle classified by vehicle type from the input image data is selected as a vehicle. It is detected by the detection unit 130 .

In the fourth step (S400), the detected vehicle is tracked by the geomchi vehicle tracking unit 180.

In the fifth step ( S500 ), the determination unit 160 determines whether the tracked vehicle is stopped in a specific area.

In the sixth step (S600), if the stop in the specific area of the tracked vehicle is not confirmed, it returns to the fourth step (S400), and when the stop is confirmed in the specific area of the tracked vehicle, the alarm generating unit 190 is alarm is generated by

In this way, related organizations such as the disaster situation room, police station, and fire station can respond in real time by receiving the intersection status analyzed along with the alarm by the alarm generating unit 190 when an event occurs from the intersection status.

As described above, according to the present invention, through a deep learning algorithm, the rotational traffic volume for each intersection approach, vehicle type, queue length, etc. are automatically aggregated, and the congestion level of the intersection is analyzed in real time, and the optimal traffic signal is controlled. there is

In addition, there is an effect of increasing the traffic capacity by about 50% at the intersection.

In addition, there is an effect of reducing waiting time for cars at intersections by about 60%.

On the other hand, there is an effect of reducing emissions (VOC emissions) of about 30% per vehicle at the intersection.

In the above, although several preferred embodiments of the present invention have been described with some examples, the descriptions of various various embodiments described in the "Specific Contents for Carrying Out the Invention" item are merely exemplary, and the present invention Those of ordinary skill in the art will understand well that the present invention can be practiced with various modifications or equivalents to the present invention from the above description.

In addition, since the present invention can be implemented in various other forms, the present invention is not limited by the above description, and the above description is intended to complete the disclosure of the present invention, and is usually It should be understood that the present invention is only provided to fully inform those with knowledge of the scope of the present invention, and that the present invention is only defined by each of the claims.

100: vehicle traffic measurement image detection device
200: AI Intersection Integrated Management Platform System
300: decision-making system
110: pre-trained deep learning network
120: image data input unit
130: vehicle detection unit
140: detection vehicle labeling unit
150: vehicle re-learning unit
160: judgment unit
161: vehicle recognition rate determination unit
162: tracking vehicle line passing determination unit
163: tracking vehicle stop determination unit
170: retrained deep learning network
180: index vehicle tracking unit
190: alarm generator
1000: vehicle traffic measurement image detection system

Claims (10)

A vehicle traffic measurement image sensing device that analyzes the status of an intersection by detecting and classifying vehicles by lane of an intersection in real time using AI deep learning; and
AI intersection integrated management platform system that controls or manages the analyzed intersection status;
Vehicle traffic video measurement system.
The method of claim 1,
The vehicle traffic measurement image detection device,
an image data input unit for receiving image data of a vehicle passing through an intersection;
a vehicle detection unit for detecting a vehicle classified by vehicle type from the image data received using the previously learned deep learning network;
a detection vehicle labeling unit for labeling the vehicle type for the detected vehicle;
a vehicle re-learning unit for re-learning the labeled and detected vehicle; and
A judging unit that determines whether or not vehicle model recognition for the detected vehicle on which re-learning has been made;
Vehicle traffic video measurement system.
3. The method of claim 2,
a detection vehicle tracking unit for tracking the detected vehicle re-learned by the vehicle re-learning unit; and
An alarm generating unit that generates an alarm when the tracked and detected vehicle is stopped in a specific area for a predetermined time;
Vehicle traffic video measurement system.
3. The method of claim 2,
The judging unit,
a vehicle recognition rate determining unit for determining whether a vehicle model is recognized;
a tracked vehicle line passing determination unit which determines whether the tracked vehicle has passed through a predetermined line; and
and a tracking vehicle stop determining unit that determines whether the tracked vehicle is stopped in a specific area.
Vehicle traffic video measurement system.
5. The method of claim 4,
The vehicle traffic image measuring device counts vehicles passing a predetermined line to obtain data on the number of vehicles passing through the intersection per hour,
Vehicle traffic video measurement system.
6. The method of claim 5,
Through the obtained data, the length of the queue in each direction of the intersection is analyzed in real time, characterized in that the signal control of the traffic light is performed to preferentially solve the congestion of the most congested road,
Vehicle traffic video measurement system.
The method of claim 1,
a decision-making system for making decisions based on the analyzed intersection status; and
A related organization that receives the analyzed intersection status when an event occurs from the intersection status;
Vehicle traffic video measurement system.
receiving image data for a vehicle passing through an intersection by an image data input unit (S1);
Detecting, by a vehicle detection unit, a vehicle classified by vehicle type from the input image data by using the previously learned deep learning network (S2) (S3);
Labeling the vehicle model for the detected vehicle by the detection vehicle labeling unit (S4);
re-learning the labeled detected vehicle by the vehicle re-learning unit (S5);
a step (S6) of determining, by a determination unit, whether or not vehicle model recognition has been made with respect to the detected vehicle for which re-learning has been made; and
If the vehicle model is not recognized by the determination unit, return to step S3, and when the vehicle model is recognized by the determination unit, the previously learned deep learning network for vehicle detection is updated using the re-learned deep learning network. characterized in that it comprises a step (S7),
A method of measuring vehicle traffic volume.
receiving image data for a vehicle passing through an intersection by an image data input unit (S10);
using the deep learning network re-learned by the vehicle re-learning unit (S20) to detect the vehicle classified by vehicle type from the input image data by the vehicle detection unit (S30);
tracking the detected vehicle by the geomchi vehicle tracking unit (S40);
determining whether or not the tracked vehicle has passed through a predetermined line (S50); and
When the passage of the predetermined line of the tracked vehicle is not confirmed, the process returns to step S40, and when the passage of the predetermined line of the tracked vehicle is confirmed, a step (S60) of increasing the statistic of the number of vehicles passing through the intersection per hour (S60). to do,
A method of measuring vehicle traffic volume.
receiving image data for a vehicle passing through an intersection by an image data input unit (S100);
using the deep learning network re-learned by the vehicle re-learning unit (S200) to detect the vehicle classified by vehicle type from the input image data by the vehicle detection unit (S300);
tracking the detected vehicle by a geomchi vehicle tracking unit (S400);
determining whether the tracked vehicle is stopped in a specific area by the determination unit (S500); and
When the stop in the specific area of the tracked vehicle is not confirmed, the process returns to step S400, and when the stop is confirmed in the specific area of the tracked vehicle, an alarm is generated by the alarm generating unit (S600); characterized in that it comprises a; doing,
A method of measuring vehicle traffic volume.

Images