KR20030084375A - Method and system for detecting a parking space using of image information - Google Patents

Abstract

PURPOSE: A system for recognizing parking state using video information and a method for constructing the system are provided to effectively manage a wide parking lot with a low cost and a small number of equipment. CONSTITUTION: A system for recognizing parking state using video information captured by a CCD camera(10) located in a parking lot includes a capturing unit(102), a memory(104), and a controller(106). The capturing unit converts the video information into a digital pixel value and captures the digital pixel value. The memory sequentially stores pixel values inputted through the capturing unit. The controller detects a variation in the pixel values stored in the memory and, when the variation exceeds a predetermined threshold value, judges that a car is moved into the parking lot or out of the parking lot.

Description

Parking status recognition system using image information and its implementation method {METHOD AND SYSTEM FOR DETECTING A PARKING SPACE USING OF IMAGE INFORMATION}

The present invention relates to a technology for recognizing a parking state of a vehicle, and in particular, a parking state recognition system using image information that is easy to apply to an unmanned parking control system by recognizing whether a vehicle is parked using image information photographed in a parking lot. And a method of implementing the same.

As a vehicle detector for recognizing a parking state of a vehicle, a pressure detector, a magnetic or magnetometer detector, an ultrasonic detector, and a radar detector may be applied. In recent years, an inductive loop detector is the most widely used.

The inductive loop detector is composed of a loop coil, a lead wire, an amplifier, and the like to detect an electrical change of a loop coil embedded / connected to a road to detect the presence and passage of a vehicle.

The reduction of the magnetic inductance of an inductive loop detector increases the frequency so that the loop line can resonate, and the amplifier responds to this frequency by using a feedback circuit that increases its oscillation frequency or a circuit that generates a frequency phase difference. .

Such frequency conversion or phase shift is the basis for obtaining detection data, and the output is driven to output one detection data.

The vehicle detector detects a change in the oscillation frequency or period generated from the oscillator of the amplifier. The change in oscillation frequency or period is due to the reduction of loop inductance when the vehicle is detected. An electric field is generated around the loop detector by the current flowing through the loop line, which can be expressed as an energy magnetic flux. If the vehicle enters the field, the vortex current is induced in the vehicle body to reduce the magnetic field lines. You get

The local control device consisting of a loop coil is composed of a vehicle detection sensor installed on the street of a parking lot, a vehicle detection controller, a communication device, a power supply, and an outdoor standard cabinet installed around the parking lot.

However, the vehicle recognition method using the loop as described above has a disadvantage that the road surface or the parking area must be teared off for the roof installation. The sensitivity of the loop detector is affected by the size of the loop, the length of the lead-in wire, the short-circuit resistance to the ground, and the thickness of the over-lay layer on the roadway. The problem has been raised that the increase in the percentage of change in inductance percentage at the input terminal decreases and eventually the sensitivity decreases.

Therefore, a systematic parking management technique with higher efficiency and higher parking management accuracy is desired.

The present invention has been made in view of the above-described needs, and an object of the present invention is to obtain a parking lot image by a pixel unit through at least one camera and process the image, and compare the pixel value changed over time, thereby parking the vehicle. It is to provide a parking status recognition system using image information to realize the parking management by calculating the parking status and parking time of the area.

In addition, another object of the present invention by using the image information input from the camera installed in the parking lot by comparing the pixel value of the before and after image in the parking area having a time difference to recognize whether the vehicle parking state in the parking area, such natural state The present invention provides a parking state recognition method using image information which prevents an incorrect parking area recognition due to noise by predicting and updating a change in pixel value in a Kalman filter.

According to an embodiment of the present invention for achieving the above object, in the parking state recognition system using the image information captured by the camera disposed in the parking lot, capturing means for converting the image information to a digital pixel value to capture and; Memory means for sequentially storing pixel values input through the capturing unit for each time; It provides a parking state recognition system using the image information, characterized in that it comprises a control means for detecting a change in the pixel value over time stored in the memory means to recognize the change in the entry / exit of the vehicle when a difference occurs more than a predetermined threshold do.

In addition, according to another embodiment for achieving the object of the present invention, the parking state recognition using the image information to recognize the presence / absence of the vehicle according to the image pixel value difference of the image information photographed by the camera disposed in the parking lot A method, comprising: storing a basic image pixel value of image information provided from a camera; Storing a comparison image pixel value of the image information provided from the camera after a preset time elapses; Comparing the difference between the base image pixel value and the comparison image pixel value on a pixel-by-pixel basis; Updating the comparison image pixel value to the base image pixel value if the predetermined difference value has not occurred; As a result of the comparison, if a predetermined difference occurs, the present invention provides a parking state recognition method using image information, comprising: displaying image information of the changed parking area to the outside.

1 is a block diagram of a parking state recognition system using image information according to the present invention;

2 is a flowchart illustrating a parking state recognition method using image information according to the present invention;

3A and 3B illustrate a parking lot image information capture screen displayed through the display of FIG. 1;

4 is a diagram illustrating a Kalman filtering technique applied to a parking state recognition technique according to the present invention;

5 is a basic conceptual block diagram of the Kalman filtering technique of FIG.

<Explanation of symbols for the main parts of the drawings>

10: CCD camera

12: interface

100: parking status recognition system

102: capturing unit

104: memory section

106: control unit

108: display unit

110: input unit

200: parking management system

300: unattended payment system

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the present invention.

Prior to the description, the core technical gist of the present invention automatically recognizes whether a vehicle is parked in a parking area by using images input through one or several cameras and automatically calculates parking management and parking fee, that is, parking lot image. By automatically recognizing the parking situation by using, it will be easy to achieve the object of the present invention from this technical idea.

1 is a block diagram of a parking state recognition system 100 using image information according to the present invention, which includes a capturing unit 102, a memory unit 104, a control unit 106, a display unit 108, and an input unit 110. Each includes.

First, the parking state recognition system 100 is connected to the camera 10, the interface 12, the parking management system 200, the unattended payment system 300 as shown.

The camera 10 may employ, for example, a charge coupled device (CCD) camera, and the interface 12 connects the camera 10 to the parking management system 200 according to the present invention.

In this case, the interface 12 serves to provide the parking management system 200 with the video image photographed from the camera 10. A wired cable or a wireless transmitter may be employed as the interface 12. If a radio transmitter is used, it should be used within an appropriate distance that is not affected by noise.

The parking management system 200 recognizes the most suitable parking information to a driver (driver of a vehicle entering) based on the parking state information provided from the parking state recognition system 100 through a display device, and until the vehicle is discharged. Operate an overall unmanned parking program to calculate your time.

The unmanned payment system 300 is a means for imposing a fee that is set according to the time elapsed until the vehicle is discharged, and may be linked with an unmanned card payment or a wireless cellphone payment line.

Meanwhile, an image signal captured by the camera 10 and input through the interface 12 is input to the capturing unit 102 in the parking state recognition system 100 for image processing, and the capturing unit 102 is a control unit. Under the control of 106, the analog video signal input through the camera 10 is converted into a digital pixel value.

The digital pixel values thus converted are recorded in the memory unit 104 in the parking state recognition system 100.

The digital pixel value recorded in the memory unit 104 is read by the controller 106 because the changed pixel value, that is, the first image pixel value and the next image pixel value over time are controlled by the controller 106. For comparison.

The display unit 108 is, for example, a display device on a PC, and is used as a means for outputting a parking lot image, that is, a captured image screen, for the operator to designate each parking block.

The input unit 110 is, for example, an input means such as a mouse, and is used as a means for capturing and specifying an arbitrary parking block or inputting various data related to parking information.

Hereinafter, with reference to the above-described configuration, with reference to the flowchart of Figure 2 attached to the parking state recognition process using the image information according to the present invention will be described in detail.

First, when an arbitrary parking lot video screen is displayed through the display unit 108, the administrator who first sets the system designates each parking block using the input unit 110 based on the video screen (S200). This stores the relative position information of each parking area on the image as shown in FIG. 3A in the memory unit 104.

Thereafter, the image information input to the parking state recognition system 100 is stored in the memory unit 104 as pixel values of each zone image except for a background image by using location information of each previously designated parking zone. The image of each of the parking spaces stored in this way is stored in the memory unit 104 as a basic image (X '(n-1) in FIG. 4 to be described later) (S202).

Thereafter, when the predetermined time elapses, the parking lot image is captured again and the pixel value of each parking area is stored using the previously stored location information. Since this is a target to be compared with a previously stored image, the target image is stored as a comparison target image in the second memory area of the memory unit 104 (S204).

Thereafter, the controller 106 compares each image pixel value in units of pixels for each zone (S206). At this time, the pixel values in the natural state may have a slight error constantly, so that a certain threshold is placed to check the pixels that change more than the threshold. Checking the threshold due to such an error will be described in more detail with reference to FIGS. 4 and 5 to which the Kalman filter is applied.

On the other hand, after performing the pixel value comparison process in step S206, the control unit 106 proceeds to step S208 to determine whether there is a difference between the base image pixel value and the comparison image pixel value.

As a result of the determination in step S208, if no difference occurs in the pixel value, the controller 106 proceeds to step S210 and stores the comparison image as a base image. That is, after comparing the base image and the comparison target image, the comparison target image is updated to the memory unit 104 as the base image. This updated base image will be compared with the newly entered comparison image in the next step.

This process aims to minimize the effects of environmental changes that may occur in long time image differences by comparing images at relatively short time intervals.

That is, when a vehicle is parked in an outdoor parking lot instead of a general indoor parking lot, there are many noise components, for example, a tree shade that grows in size gradually, or a rapid change in light quantity at sunset. Every time the capture is made, the pixel value is changed little by little, and the pixel value changes according to the amount of sun light reflected on the road surface. Thus, a simple comparison can lead to false recognition in this case.

After re-saving the comparison target image as the base image, the parking area situation display part displayed on the display unit 108 is changed according to the image comparison result in the previous step.

The state of the parking area can be divided into the state in which the vehicle is entered and the state in which the vehicle is discharged. If there is no difference in the pixel value, it means that there is no change in the state of the parking area. That is, if the empty parking area in the previous screen means that the empty space still, if it was parked, it means that it is still parking.

If a change occurs in the pixel value, this means that a change in state occurs in the parking area. In other words, if it was an empty parking area on the previous screen, it means that it was parked on this screen, and if it was an area parked on the previous screen, it means that it is an empty area on this screen. Accordingly, the controller 106 proceeds to step S212 and controls to display the image information of the changed parking area on the display unit 108 as shown in FIG. 3B.

If there is little change in the image comparison result, save the comparison image as the base image and maintain the parking area. That is, if it is an empty parking area, it indicates that it is an empty parking area, and if it is a parking area, it keeps counting the parking time that was previously counted. Then, the new comparison image is input and the previous image comparison step is repeated.

If the image change is recognized in the image comparison result, the corresponding time calculation is performed together with the conversion of the parking area status indication (S214). In other words, if there is an empty parking area, the parking area is indicated and the parking time is counted together with the storage of the parking time. If it is a parking area, this means leaving the vehicle, and thus marking an empty parking area, and performing the steps of calculating the parking time, calculating the fare, etc. together with the recording of the leaving time (the leftmost area of FIG. 3B).

Thereafter, after updating the comparison image to the base image and repeating the same process as before (S216).

As described above, in the process of updating to a new image, the pixel values are different depending on the environment. If the difference values accumulate, it becomes a cause of erroneous recognition.

Therefore, in order to solve this problem, it is possible to reduce this natural pixel value change by predicting the pixel value to be changed in the future by using the Kalman filter.

4 is a diagram for explaining such a Kalman filtering technique.

The Kalman filter estimates the present or future by using past data. The Kalman filter is composed of a state equation of Equation 1 and a measurement equation of Equation 2.

x (k + 1) = Φ (k, k + 1) x (k) + ω (k)

z (k) = H (k) x (k) + υ (k)

Where Φ (k, k + 1) is the state transition matrix for unit time, ω (k) is the estimated error, H (k) is the measurement matrix, and υ (k) is the measurement error.

Assume that ω (k) and υ (k) are Gaussian functions with a mean of zero. H (k) is a unit matrix because it corresponds to a simple measurement of pixel values. First, calculate the state variable and state variable covariance at the next measurement time from the current estimate and the covariance value, make a new measurement, and then repeat the process of newly calculating the state variable estimate by mixing the measured value and the state variable estimate obtained above. do.

In the present invention, the value estimated is a pixel value of the parking lot image image pixel, and an equation for the estimated value is shown in Equation 3 below.

here, Is an estimate, Is the mean of the state variables at time k, Denotes the average of the measured values at k time. In addition, g_k may be expressed as Equation 4 below.

Where σ_x_k is the variance of the state variable at k time, and σ_ω_k is the variance of the measurement error at k time.

The estimated value is stored as the pixel value of the base image to be compared and compared with the newly input comparison image pixel value in the next step.

5 is a basic conceptual block diagram of the Kalman filtering technique of FIG.

As described above, the present invention can find out whether the vehicle is parked in the parking area, so that it can be associated with the parking management system 200. First, since the camera 10 can grasp all the occupied states of the current parking lot by zone, the parking state recognition system 100 can notify the driver who enters the empty area closest to the current entrance through the display unit 108.

The car occupied in this way is immediately recognized according to the above-described procedure through the installed camera 10, and thus the parking state recognition system 100 calculates the time.

At the time of departure, payment can be made using an unmanned card payment system or a wireless mobile phone payment system. When the payment is made, the central computer is notified of the result, and the central computer recognizes the payment for parking status management of the area.

When the driver leaves the car, the central computer recognizes the departure of the area due to the pixel change in the parking area.If the payment status of the area is confirmed, the vehicle is allowed to leave and the departure is confirmed without the fee. If not, the manager will be notified and the barrier at the exit will be blocked.

As a result, the above system can realize the automation of unmanned parking management.

In addition, since the camera applied to the present invention must use the color information of the parking area, it is necessary to employ a color CCD camera. Also, when installing the camera, install it at a suitable height where a wide range of parking lots can be taken. In particular, choose trees or facilities that do not cover the facility so that you will not be able to take pictures of the parking lot.

Indoor car parks usually have low ceilings, making it impossible to photograph a wide range of parking areas, such as outdoor car parks. In this case, by installing a camera having a hemispherical angle of incidence that allows a wide range of shooting even at a relatively low height, it can be implemented to secure a wide range of parking area, that is, more vehicles in the field of view.

Also, if the ceiling is low, installation of multiple cameras is inevitable, but if you adopt a high pixel camera that supports color images and have enough lighting in the indoor parking lot, you will not need to install too many cameras. Will be readily apparent to those of ordinary skill in the art.

As mentioned above, although this invention was concretely demonstrated based on the Example, this invention is not limited to this Example, Of course, various deformation | transformation is possible for it in the range which does not deviate from the summary.

Therefore, the present invention can implement more efficient and accurate systematic parking management.

Claims (8)

In the parking state recognition system using the image information captured by the camera disposed in the parking lot, Capturing means for capturing the image information by converting the image information into digital pixel values; Memory means for sequentially storing pixel values input through the capturing unit for each time; And a control means for detecting a change in a pixel value over time stored in the memory means and recognizing a change in entry / exit of a vehicle when a difference of more than a predetermined threshold occurs. The method of claim 1, The system, A parking state recognition system using image information, characterized in that it is associated with a parking management system that recognizes the most suitable parking information based on the parking state information to the driver and calculates the time until an arbitrary vehicle is released. The method of claim 1, The system, Parking state recognition system using the image information, characterized in that associated with the unattended payment system that charges a fee that is determined according to the time that elapses before any vehicle is released. In the parking state recognition method using the image information to recognize the presence / absence of the vehicle according to the difference of the image pixel value of the image information captured by the camera disposed in the parking lot, Storing basic image pixel values of the image information provided from the camera; Storing a comparison image pixel value of the image information provided from the camera after a preset time elapses; Comparing the difference between the base image pixel value and the comparison image pixel value in units of pixels; Updating the comparison image pixel value to a basic image pixel value if a predetermined difference value has not occurred as a result of the comparison; And displaying the image information of the changed parking area to the outside when a predetermined difference value occurs as a result of the comparison. The method of claim 4, wherein The basic image pixel value corresponds to the first parking lot image information photographed, and the comparison image pixel value corresponds to the parking lot image information currently photographed. The method of claim 4, wherein Comparing the difference of the pixel value in pixel unit, A parking state recognition method using image information, characterized by applying a Kalman filter for minimizing an error that may occur in a pixel value in a natural state. The method of claim 4, wherein The method, A parking state recognition method using image information, characterized in that it is associated with a parking management system that recognizes the most suitable parking information based on the parking state information to the driver and calculates the time until an arbitrary vehicle is released. The method of claim 4, wherein The method, Parking state recognition method using the image information, characterized in that associated with the unattended payment system that charges a fee that is determined according to the time that elapses before any vehicle is released.