KR20050065881A - Apparatus for determining type of vehicle for controling parking and method thereof - Google Patents

Abstract

The present invention relates to a vehicle type discrimination apparatus for parking control, a method thereof, and a computer-readable recording medium having recorded thereon a program for realizing the method. Parking control vehicle type determination device for receiving a side image of the vehicle of the present invention, to determine the vehicle type, object extraction means for separating the side image of the vehicle from a pre-stored background image; Morphological skeletal transformation means for performing morphological skeletal transformation using morphemes; Morphological skeletal feature extraction means for performing feature setting for vehicle type recognition; And discriminating means for discriminating a vehicle type by using a feature for vehicle type recognition received from the morphological skeleton feature means. According to the present invention, it is possible to easily determine the type of the vehicle by performing a calculation through hydraulic morphological analysis from the vehicle side image information obtained through the instant capture of the high-speed vehicle.

Description

Apparatus for Determining Type of Vehicle for Controling Parking and Method Thereof}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle type discrimination apparatus for parking control and a method thereof and a computer-readable recording medium recording a program for realizing the method. In particular, the present invention relates to a charge coupling device in an image processing system for parking control for building automation. A parking control vehicle type discrimination apparatus for automatically determining a vehicle type through an image processing process using a coupled device (CCD) camera, and a method thereof and a computer-readable recording medium recording a program for realizing the method. will be.

As buildings are advanced and unmanned management systems are introduced, systems for automatically managing the entry and exit of vehicles in parking lots are being developed and introduced.

The management system mainly recognizes a license plate of a vehicle, receives information about the vehicle, and manages the vehicle. The management system determines the type of the vehicle in response to a more detailed fare system and when the exterior information of the vehicle is digitized and stored. There is a need for a system description.

To this end, in the conventional case, the height information of the vehicle is used to classify the vehicle into a large, medium, and small size rather than the detailed classification of the vehicle, but there is a problem in that it is not possible to accurately recognize a wide variety of vehicle types.

The present invention has been proposed in order to solve the above problems, and first obtains a background image before entering the vehicle, and obtains an image of the vehicle in which the background is separated by comparing with the image after entering the vehicle. It is an object of the present invention to provide a vehicle control apparatus for discriminating a parking vehicle for discriminating the type of a vehicle using a weighted skeleton characteristic unique to each vehicle through morphological skeletal transformation.

In addition, the present invention obtains the background image before the vehicle enters first, and then obtains the image of the vehicle separated from the background by comparing with the image after entering the vehicle, and then weights unique to each vehicle through morphological skeleton transformation It is another object of the present invention to provide a parking control vehicle type discrimination method for determining a type of vehicle using a skeleton feature and a computer-readable recording medium storing a program for realizing the method.

According to an aspect of the present invention, there is provided a vehicle type determination apparatus for controlling a vehicle for receiving a side image of a vehicle and determining a vehicle type, comprising: object extraction means for separating the side image of the vehicle from a pre-stored background image; Morphological skeletal transformation means for performing morphological skeletal transformation using morphemes; Morphological skeletal feature extraction means for performing feature setting for vehicle type recognition; And discriminating means for discriminating a vehicle type by using a feature for vehicle type recognition received from the morphological skeleton feature means.

According to another aspect of the present invention, there is provided a vehicle control method for determining a vehicle type by receiving a side image of a vehicle, the method comprising: a first step of separating a side image of the vehicle from a pre-stored background image; Performing a morphological skeletal transformation using a morpheme; Performing a feature setting for vehicle type recognition; And a fourth step of determining a vehicle type by using the feature for vehicle type recognition received from the morphological skeleton feature means.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. First of all, in adding reference numerals to the components of each drawing, it should be noted that the same components have the same number as much as possible even if displayed on different drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a configuration diagram of an embodiment of a vehicle type determination apparatus for parking control according to the present invention.

As shown in the figure, the vehicle model determining apparatus includes an object extractor 110, a morphological skeleton transforming unit 120, a morphological skeleton feature extracting unit 130, and a vehicle model determining unit 140. .

In addition, it may include a floor buried loop for recognizing the entry of the vehicle, a camera for acquiring the side image of the vehicle, a transmission unit for transmitting the image signal to the vehicle type determination device of the present invention.

The object extracting unit 110 receives a side image of the vehicle from the transmitting unit and is responsible for separating the vehicle image from a pre-stored background image. Hereinafter, the object extractor 110 will be described in detail with reference to FIG. 2.

FIG. 2 is a detailed structural diagram of an object extracting unit of FIG. 1.

As shown in the figure, the object extractor of the present invention includes a difference image calculator 210, a binary image extractor 220, and a noise remover 230.

The difference image calculator 210 is responsible for calculating a difference image of the side image and the background image of the vehicle. Difference image, continuous time Wow Luminance image at Wow Can be obtained using the pixel-by-pixel difference. This is as follows.

The binary image extractor 220 is responsible for extracting a binary image V to which a threshold is applied to the difference image obtained by the difference image calculator 210.

The binary image extractor 220 may extract a binary image of the vehicle object separated from the background, but there are high frequency noise and fine components in the background. In order to remove such fine components, the noise canceller 230 is responsible for obtaining a binary vehicle object image X by applying a morphological opening to the corresponding image.

The morphological operation will be described in detail.

Mathematical morphology based on set theory provides an effective method for digital image interpretation and processing. That is, various features such as shape, size, contrast, and connectivity of an object may be extracted from an image. In addition, since the extraction of the feature for the part that satisfies homogeneity and continuity is effective, image segmentation and shape expression are easy.

The two basic operations of mathematical morphology are dilation and erosion. Is all points, Is a set of A and B When a subset of, the binary dilation of A by B is given by the following equation.

here, Is the translation of A by b. In addition, the binary erosion of A with respect to B is given by the following equation.

Closing and opening by dilation and erosion are defined as follows, respectively.

In addition, the n-th order multiscale closing and opening are as follows.

At this time, to be.

Binary morphological operations can be naturally extended to gray scale images by using minimum or maximum operations. ego Gray scale dilation and erosion are as follows.

Similarly, its closing and opening are defined as follows.

Again, in FIG. 1, the morphological skeletal transform unit 120 is responsible for performing a morphological skeletal transformation on the object extracted by the object extractor 110.

The three-dimensional morphological operations are applied to the process and restoration of the Morphological Skeleton Transform. The morphological framework SK (X) for the discrete binary region can be obtained by the following equation.

here, And "-" means difference operation. Is called the n-th order skeleton subset of X, and the original image can be reconstructed as follows.

If k = 0 then all skeletal elements are used for restoration (Exact Reconstruction) And In this case, the opening of X by mB can be obtained (Smoothed Version). All skeletal subsets The information contained in can be completely expressed as Morphological Skeleton Functions (SKF) defined by the following equation.

The large SKF value represents the main body of the object of pixels converted from the skeletal point, and the small SKF value represents the fine pixel of the contour. Thus, the object can quickly recover from skeletal representations of large SKF values and gradually complete finer details.

3A to 3F illustrate examples of skeletal transformation of the morphological skeletal transformation unit of FIG. 1 using 3 × 3 Rhombus morphemes.

FIG. 3A shows the object X, FIG. 3B shows the skeletal transformation result pixel, and N is 3. FIG. As a result of the skeletal transformation, SKF applying the above equation 14 for each pixel is shown in FIG. 3C. 3d to 3f show the restoration process by the morphemes 3B, 2B, and B, respectively. In other words, the restoration order is changed to 3, 2, 1, 0, and restored to the assumptions of FIGS. 3D, 3E, 3F, and 3A, and it can be seen that the larger n represents the main part of the object.

After extracting the morphological skeleton of the separated vehicle image, setting of a feature for vehicle type recognition is required.

In general, morphological techniques for extracting shape and size information of an image mainly include a pattern spectrum (Morphological Pattern Spectrum) and an algorithm related to skeletalization. The pattern spectrum has a disadvantage in that a large amount of computation is required because it is necessary to perform repeated closing and opening operations for each morpheme using various morphemes, and thus, the framework-related algorithm will be used in the present invention.

The morphological skeletal feature extractor 130 performs a function of performing feature setting for vehicle type recognition on an image obtained by performing morphological skeletal transformation on a separated vehicle object.

At this time, it is not good that there are many features in the image recognition system, and it is ideal to maintain the recognition rate and to consider the load of the system with the appropriate minimum feature selection. Several skeletal matching algorithms have been proposed for this purpose, but simple features are efficient for real-time processing under limited circumstances such as vehicle model recognition.

Accordingly, in an embodiment of the present invention, a skeletal histogram (SKH) and a weighted skeletal feature (WSF) applied to the vehicle model recognition will be proposed.

4 is a detailed structural diagram of an embodiment of the morphological skeletal feature extractor of FIG. 1.

As shown in the figure, the morphological skeletal feature extractor 130 of the present invention includes a skeletal histogram extractor 410 and a weighted skeletal feature extractor 420.

The skeletal histogram extractor 410 is responsible for extracting a skeletal histogram using the following equation for an image obtained by performing a morphological skeletal transformation on the separated vehicle object received from the morphological skeletal transform unit.

At this time, "Cord" means the number of elements (Cardinality). The skeletal histogram is the number of nth skeletal subsets. That is, in the case of FIG. , , And to be.

In addition, the weighted skeletal feature extractor 420 is responsible for extracting a weighted skeletal feature proportional to the area of the morpheme using the histogram of the skeleton received from the skeletal histogram extractor 410 as follows. In other words, the weighted skeletal feature is the only measure that distinguishes the shape and size of an object according to its shape.

Skeletal subset Affects the shape and size of object X in proportion to the square of n. In other words, skeletalization depends on the size of the morphemes. Was given. In addition, the fine component Since should be considered, as the weight in Equation 16 instead Use When the above process is applied to the rectangular object of FIG. 2, the following equation is obtained.

This is a weighted skeletal feature that is the only eigenvalue by the Rhombus morpheme of the object of FIG. 2.

Finally, the vehicle model discrimination unit 140 is responsible for determining a vehicle model using the weighted skeleton feature.

5 is a flowchart illustrating an embodiment of a method for determining a vehicle type for parking control according to the present invention.

As shown in the figure, the vehicle type determination method of the present invention receives a side image of the vehicle, separates the vehicle image from the pre-stored background image, performs object extraction (S501), and determines the shape and size of the extracted object. In order to preserve, morphological skeletal transformation is performed (S502).

Subsequently, in order to set features for vehicle model recognition, morphological skeletal features such as skeletal histograms and weighted skeletal features are extracted (S503), and vehicle type discrimination is performed from the extracted weighted skeletal features (S504).

Equations used in the vehicle type discrimination method of the present invention are the same as those used in the vehicle type discrimination apparatus, and thus detailed description thereof will be omitted.

Hereinafter, with reference to the drawings simulated by using the vehicle type determination device and the method for parking control of the present invention will be described.

In the simulation according to the present invention, an image was obtained by using a fixed camera.

6A is an exemplary view of a background image used in the present invention, and FIGS. 6B and 6C are exemplary views of a car and a truck image disposed in FIG. 6A, respectively.

6D and 6E are diagrams illustrating examples of images separated by the object separating unit, respectively.

If it is applied to a fixed unmanned detection system such as a highway, it will be easier to extract the vehicle object, which is a preprocessing step, because the background is a flat road surface and the distance is constant.

7A to 7C are diagrams illustrating examples of morphemes used in skeletal transformation of the morphological skeletal transform unit of FIG. 1 to illustrate square morphemes, Rhombus morphemes, and circular morphemes, respectively. .

As shown in the figure, the origin of each morpheme is the center.

8A and 8B are obtained by applying the square morpheme of FIG. 7A and skeletal transformation of FIGS. 6D and 6E. One example for showing the distribution of.

Similarly, FIGS. 8C and 8D are obtained by applying the Rhombus morpheme of FIG. 7B and skeletal transformation of FIGS. 6D and 6E. 8E and 8F are obtained by applying circular morphemes of FIG. 7C and skeletal transformation of FIGS. 6D and 6E. One example for showing the distribution of.

9A to 9C are diagrams illustrating examples of skeletal histograms extracted from FIGS. 8A, 8C and 8E, respectively. FIGS. 10A to 10C illustrate skeleton histograms extracted from FIGS. 8B, 8D and 8F, respectively. This is an example for illustration.

The reason why the maximum morpheme size N is relatively small when applying the circle morpheme is that the remaining two morphemes are 3 × 3, 5 × 5, 7 × 7,... However, the size of the circle morpheme is 5 × 5, 9 × 9, 13 × 13,… as n increases. Because it becomes larger.

Table 1 below shows the result of comparing the weighted skeletal feature WSF from the skeletal histogram for five object frames of passenger cars and trucks.

As shown in the above table, there is no problem in vehicle model recognition regardless of which morphemes are applied, and especially when the Rhombus morphemes are applied.

In other words, the ratio of the mean of the weighted skeletal features is 1.21, 1.57, and 1.24 for Square, Rhombus, and Circle morphemes, respectively. It can be seen that the difference is large and it is easy to recognize the vehicle model. In the case of the square morpheme, the weight of the skeletal features is larger for the cars and trucks because the trucks are more angular. This result can also be confirmed in FIGS. 9A and 10A.

The method of the present invention as described above may be implemented as a program and stored in a computer-readable recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.).

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes are possible in the art without departing from the technical spirit of the present invention. It will be clear to those of ordinary knowledge.

The present invention as described above, by performing the calculation through the hydraulic morphological analysis from the vehicle side image information obtained through the instantaneous capture of the high-speed vehicle, there is an effect that can easily determine the type of the vehicle.

Therefore, the present invention has the effect of contributing to the automation of vehicle model determination by enabling the configuration of the vehicle automatic recognition system essential for parking control or unmanned toll gate system.

1 is a configuration diagram of an embodiment of a vehicle type determination apparatus for parking control according to the present invention;

FIG. 2 is a detailed structural diagram of an object extracting unit of FIG. 1;

3A to 3F are exemplary views illustrating skeletal transformation of the morphological skeletal transformation unit of FIG. 1;

4 is a detailed structural diagram of an embodiment of the morphological skeletal feature extractor of FIG. 1;

5 is a flowchart illustrating an embodiment of a method for determining a vehicle type for parking control according to the present invention;

6A illustrates an example of a background image used in the present invention;

FIG. 6B is an exemplary view of a passenger car image disposed in FIG. 6A;

FIG. 6C is an exemplary view of a truck image disposed in FIG. 6A;

6D and 6E are diagrams illustrating examples of images separated by the object separating unit, respectively;

7A to 7C are diagrams illustrating examples of morphemes used in skeleton conversion of the morphological skeleton conversion unit of FIG. 1;

8A and 8B are obtained by applying the square morpheme of FIG. 7A and skeletal transformation of FIGS. 6D and 6E. One example for showing the distribution of,

8C and 8D are obtained by applying the Rhombus morpheme of FIG. 7B and skeletal transformation of FIGS. 6D and 6E. One example for showing the distribution of,

8E and 8F are obtained by applying skeleton morphemes of FIG. 7C and skeletal transformation of FIGS. 6D and 6E. One example for showing the distribution of,

9A to 9C are exemplary views for illustrating skeletal histograms extracted from FIGS. 8A, 8C, and 8E, respectively.

10A to 10C are exemplary views for illustrating skeleton histograms extracted from FIGS. 8B, 8D, and 8F, respectively.

* Explanation of symbols for the main parts of the drawings

110: object extraction unit 120: morphological skeleton conversion unit

140: morphological skeleton feature extraction unit 140: vehicle model discrimination unit

210: difference image calculator 220: binary image extractor

230: noise canceller 410: skeletal histogram extractor

420: Weighted Skeleton Feature Extractor

Claims (7)

In the parking control vehicle type determination device for receiving a side image of the vehicle to determine the vehicle type, Object extraction means for separating a side image of the vehicle from a pre-stored background image; Morphological skeletal transformation means for performing morphological skeletal transformation using morphemes; Morphological skeletal feature extraction means for performing feature setting for vehicle type recognition; And Discriminating means for discriminating a vehicle type by using a feature for vehicle type recognition received from said morphological skeleton feature means Vehicle type determination device for parking control comprising a. The method of claim 1, The object extraction means, Difference image calculating means for calculating a difference image between the side image of the vehicle and the background image; Binary image extracting means for extracting a binary image applying a threshold to the difference image; And Noise removing means for removing noise present in the binary image Parking control vehicle type determination device comprising a. The method of claim 1, The morphological skeletal transformation means, Performing morphological skeletal transformation by the following equation Vehicle type determination device for parking control, characterized in that. (only, Is the nth skeleton subset of X) The method of claim 1, The morphological skeletal feature extraction means, Skeletal histogram extracting means for extracting a skeletal histogram with respect to the image received from the morphological skeletal transforming means; And Weighted skeletal feature extraction means for extracting weighted skeletal features proportional to the area of morphemes using the skeletal histogram Parking control vehicle type determination device comprising a. The method of claim 4, wherein The skeletal histogram extracting means, Performing skeletal histogram extraction by the following equation Vehicle type determination device for parking control, characterized in that. (Where "Cord" is the number of elements (Cardinality)) The method of claim 4, wherein The weighted skeleton feature extraction means, Performing weighted skeletal feature extraction by the following equation Vehicle type determination device for parking control, characterized in that. In the parking control vehicle type determination method for receiving a side image of the vehicle, to determine the vehicle type, A first step of separating a side image of the vehicle from a pre-stored background image; Performing a morphological skeletal transformation using a morpheme; Performing a feature setting for vehicle type recognition; And A fourth step of discriminating the vehicle type by using the feature for vehicle type recognition received from the morphological skeleton feature means Parking control vehicle type determination method comprising a.