KR101772429B1 - Apparatus for acquiring characteristic information of object and method thereof - Google Patents

Abstract

The present invention relates to an apparatus and a method for acquiring characteristics of an object. According to the present invention, an object characteristics acquiring method by an apparatus for acquiring characteristics information of an object in an image captured by an image capturing means comprises the following steps: dividing an area in an image into a plurality of areas individually corresponding to predetermined optical zoom levels through a learning process using a plurality of optical zoom levels; selecting an area, where the object is detected, among the plurality of areas when the object is detected in the image; and obtaining characteristics information by enlarging the selected area based on the optical zoom level corresponding to the selected area. According to the present invention, the characteristics information of the object can be acquired in a rapid and easy manner even without a cumbersome rotating/zoom process as the image or a relevant area is enlarged and provided based on the optical zoom level corresponding to a sensing area when the object is sensed in the image.

Description

[0001] APPARATUS FOR ACQUIRING CHARACTERISTIC INFORMATION OF OBJECT AND METHOD THEREOF [0002]

The present invention relates to an apparatus and method for acquiring object feature information, and more particularly, to an apparatus and method for acquiring feature information of an object quickly while minimizing an optical zoom operation.

Generally, video surveillance systems using CCTV cameras are installed in places such as buildings, roads, alleys, parks, and indoor spaces and are used for crime prevention, field surveillance, crime prevention, and crime prevention.

The video surveillance system mainly includes a fixed camera that captures wide-area images and a turn-around camera that captures local images. When an object such as a vehicle or a person is detected through a fixed camera, the object is photographed and recognized by controlling the rotation / zoom (PTZ; Pan, Tilt, Zoom) of the rotation type camera based on the position of the detected object.

Since the existing video surveillance system frequently controls the rotation / zooming of the rotating type camera to the detection position in order to magnify and photograph the detected object every time the object is detected, There has been a problem that the lifetime is considerably shortened. Particularly, due to frequent rotation / zooming, the problem of missing the position of the detected object frequently occurs.

In order to solve the problem of the conventional video surveillance system, a method of using only a rotatable camera has been disclosed. However, a system that detects an object using only a rotatable camera, It is very difficult to continuously process a plurality of objects and the rotation / zooming is repeated for each object, so that the life of the motor is shortened.

Therefore, a new system capable of detecting / recognizing objects quickly and accurately is desperately needed while solving the existing problems.

The technology of the background of the present invention is disclosed in Korean Patent Laid-Open No. 10-2006-104606 (published October 10, 2006).

The object of the present invention is to provide an apparatus and method for obtaining feature information of an object that can quickly acquire feature information of an object by minimizing optical zoom adjustment.

The present invention provides an object characteristic information acquiring method of an apparatus for acquiring characteristic information of an object in an image photographed through a photographing means, comprising the steps of: acquiring, at a predetermined optical zoom level through a learning process using a plurality of optical zoom levels The method comprising the steps of: dividing the object into a plurality of corresponding regions; selecting the region of the plurality of regions in which the object is detected when the object is detected in the image; And acquiring feature information of the object by enlarging the selected region.

In addition, the acquiring of the feature information of the object may acquire feature information of the object by analyzing the selected area through a digital zoom having a predetermined level range.

The selecting the region may include calculating a position value of the object in the image, and selecting an area including the calculated position value as a region in which the object is detected.

In the dividing into the plurality of regions, a boundary region may be added and divided between adjacent regions of the plurality of regions, and an average optical zoom level between the adjacent regions may be set to an optical zoom level of the boundary region.

The division into the plurality of regions may divide each of the divided regions into a plurality of sub regions corresponding to a predetermined digital zoom level through a learning process using a plurality of digital zoom levels.

If the predetermined image quality condition is not satisfied as a result of applying the predetermined maximum digital zoom level to the sub region in the boundary between the second region and the sub region in the plurality of divided sub regions in the first region, The first and second areas may be modified to belong to a second area and the second area may be an area where an optical zoom level higher than the first area is set.

The feature information may include at least one of a vehicle type, a vehicle number, a vehicle type, and a vehicle color if the object is a vehicle, and if the object is a person, at least one of sex, age, color, Information.

The present invention provides an object feature information acquiring apparatus for acquiring feature information of an object in an image photographed through a photographing means, the apparatus comprising: A detection unit which detects an object in the plurality of regions and detects the object in the image, and for dividing the plurality of regions into a plurality of corresponding regions; And a control unit for enlarging the selected area to obtain the feature information of the object.

In addition, the controller may analyze the selected area through a digital zoom having a predetermined level range to acquire the feature information of the object.

Also, the detecting unit may calculate a position value of the object in the image, and may select an area including the calculated position value as a region in which the object is detected.

Further, the dividing unit may divide a boundary region between adjoining regions of the plurality of regions, and set an average optical zoom level between the adjacent regions to an optical zoom level of the boundary region.

In addition, the dividing unit may divide each of the divided regions into a plurality of sub regions corresponding to a predetermined digital zoom level through a learning process using a plurality of digital zoom levels.

In addition, when the predetermined image quality condition is not satisfied as a result of applying the predetermined maximum digital zoom level to the sub-area at the boundary with the second area of the divided sub-areas in the first area, The sub region may be assigned to the second region to modify the first and second regions, and the second region may be an area in which an optical zoom level higher than the first region is set.

According to the present invention, when an object is detected in an image, an image or a corresponding area is enlarged and provided based on an optical zoom level and a digital zoom level corresponding to the detected area, Can be obtained quickly and easily.

In addition, according to the present invention, an intra-image area is divided and set in advance through a learning process so that an area in which an object is detected is detected immediately after the object is detected in the image, and the object is enlargedly photographed and recognized .

Thus, it is possible to effectively solve problems such as object continuation processing and camera motor life shortening which have occurred in the past.

1 is a view for explaining an object feature information acquisition system according to an embodiment of the present invention.
2 is a diagram illustrating an apparatus for acquiring object feature information according to an embodiment of the present invention.
3 is a view showing a state in which an image of the photographing means is divided into a plurality of regions corresponding to respective zoom levels in the embodiment of the present invention.
FIG. 4 is a diagram showing a state in which a boundary region is added and divided between a plurality of regions shown in FIG. 3; FIG.
5 is a view for explaining a method for acquiring object feature information using the system shown in FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention.

1 is a view for explaining an object feature information acquisition system according to an embodiment of the present invention. 1, an object feature information acquisition system according to an embodiment of the present invention includes a photographing means 100 and an object feature information acquiring apparatus 200. As shown in FIG.

The photographing means 100 photographs an image of the surveillance region and provides the photographed image to the object feature information obtaining apparatus 200. [ The photographing means 100 includes an optical zoom function and a digital zoom function.

In the case of optical zoom, the zoom level is optically changed by moving the position of the optical lens. As the level (magnification) becomes larger, the resolution increases slightly. In the case of digital zoom, the resolution is not increased by enlarging the image itself through image processing or the like.

The object feature information obtaining apparatus 200 may be connected to the photographing means 100 through a direct communication network or may be connected through a separate field device (not shown).

The object feature information obtaining apparatus 200 detects the object from the image photographed by the photographing means 100 and acquires the feature information of the detected object.

Specifically, the object feature information obtaining apparatus 200 detects an object in an image, enlarges a portion of the detected object or object using the optical zoom level corresponding to the detected region of the object, Thereby acquiring feature information of the object.

Here, the object feature information obtaining apparatus 200 can further increase the analysis and acquisition efficiency of the feature information by additionally applying the digital zoom of the set range to the enlarged image. These additional digital zoom functions may be added or omitted depending on the situation.

Hereinafter, an object feature information obtaining apparatus according to an embodiment of the present invention will be described in detail.

2 is a diagram illustrating an apparatus for acquiring object feature information according to an embodiment of the present invention.

2, the apparatus for acquiring object feature information according to an embodiment of the present invention includes a partition 210, a detection unit 220, a control unit 230, and a storage unit 240.

The division unit 210 receives the image photographed by the photographing means 100. At this time, the received image may correspond to the wide area image photographed through the optical zoom adjustment of the photographing means 100.

The division unit 210 divides the area in the received image into a plurality of areas corresponding to the respective optical zoom levels. To this end, the division unit 210 performs a learning process of applying a plurality of optical zoom levels to an image.

Accordingly, the image is divided into a plurality of regions. The division of the area can be achieved by a learning process using a plurality of optical zoom levels. For example, the imaging means 100 divides the intra-image region into a plurality of regions corresponding to the respective levels while individually applying optical zooms of various levels to the captured image. In this case, the size of each divided area may be equal or different.

Whereby a predetermined optical zoom level capable of recognizing an object in the corresponding region is applied to each region in the image. Each area is capable of recognizing an object at the time of applying a predetermined level of optical zoom corresponding thereto and it will be difficult to recognize the object when applying optical zoom to a predetermined level or less.

Generally, the upper region in the photographed image corresponds to a region that is relatively far away from the lower region. In order to enable object recognition for all regions in the image using a learning process, a higher-level optical zoom is set at a farther distance, that is, at a higher level from the image to the upper region.

3 is a view showing a state in which an image of the photographing means is divided into a plurality of regions corresponding to respective zoom levels in the embodiment of the present invention.

In the case of FIG. 3, the image photographed by the photographing means 100 is divided into three regions including the first to third regions, and the present invention is not limited thereto.

Referring to FIG. 3, optical 1x zoom, 2x zoom, and 3x zoom are set for the first, second, and third regions, respectively. In the case of the third region, it is confirmed that the optical zoom of the highest level among the three regions is matched to a position located behind the first and second regions.

As described above, a higher zoom level is applied to the upper region among the divided regions of the image. For example, the first region may apply a first level of optical zoom, the second region may apply a second level of optical zoom, and the third region may apply a third level of optical zoom.

Here, the zoom level difference may be 1 or more between adjacent two divided regions, or the level difference between divided regions may become larger or smaller toward the upper region even in the same image.

In addition, there may exist a boundary region between the two adjoining divided regions in which the zoom value of the average level value of the two divided regions is set. This will be described in detail with reference to FIG. 4 below.

The division unit 210 may provide the storage unit 240 with the area information including the position, size, and the like of each divided area of the photographed image and the optical zoom level information corresponding to each area. In the case of the learning process, automatic learning based on a program, manual learning based on an operator's visual discrimination, etc. can be used to search for a satisfactory optical zoom level such as set image quality and sharpness for each area.

Thereafter, when the detecting unit 220 detects an object in the image, it selects an area in which the object is detected from among the plurality of areas divided by the dividing unit 210. [ Here, the detecting unit 220 calculates the position value of the object in the image and selects the area including the calculated position value as the detected area of the object.

The control unit 230 searches the storage unit 240 for the optical zoom level corresponding to the selected area and controls the photographing unit 100 based on the searched optical zoom level to obtain an enlarged image obtained by enlarging the selected area.

That is, the control unit 230 enlarges the image based on the object based on the optical zoom level corresponding to the selected area, and analyzes the enlarged image to acquire the feature information of the object.

Here, the feature information may include at least one of a vehicle type, a vehicle number, a vehicle type, and a vehicle color when the object is a vehicle. If the object is a person, the feature information may include at least one of sex, age, color, and shape.

The controller 230 may further use the digital zoom when acquiring the feature information of the object. That is, the controller 230 may analyze the image through the digital zoom having a predetermined level range to acquire the feature information. For example, the feature information can be analyzed by applying a digital zoom level corresponding to the position of the object to the image. In addition to the optical zoom function, digital zoom function can be used to acquire feature information more accurately and increase acquisition efficiency.

The storage unit 240 stores information of a plurality of regions divided by the division unit 210 and information of the optical zoom level set corresponding to each region. Of course, the storage unit 240 may store and store at least one of the detected image of the object, the image of the region where the object in the image is located (enlarged image), and the feature information of the object.

In addition, the storage unit 240 may further store information of sub-areas obtained by dividing a plurality of areas into a plurality of areas and information of a digital zoom level set corresponding to each sub-area.

In this regard, the division unit 210 may divide each of the divided regions into a plurality of sub-regions respectively corresponding to a predetermined digital zoom level, and stores the divided sub-region information and the corresponding digital zoom level information (240). ≪ / RTI > In this case, the division unit 210 may analyze the image enlarged at the optical zoom level through the digital zoom level corresponding to the sub-area where the object is detected to obtain the feature information of the object.

The division unit 210 may divide each region into a plurality of sub-regions through a learning process using a plurality of digital zoom levels. 3, each of the first through third regions is divided into three sub-regions. Each of the three sub-regions has a digital 1x zoom Dx1, a 2x zoom Dx2, a 3x zoom Dx3, Is set.

Here, of course, the zoom level difference may be 1 or more, and the size of the sub region may be the same or different between two adjacent sub regions in the same region. Further, the number of division of the sub-areas for each of the first, second, and third areas may be the same or different.

After division of the sub-region, the division unit 210 may perform verification of the digital zoom level for the sub-region located at the boundary between the plurality of regions. This verification can solve the learning error of the digital zoom level and allow more precise definition of the boundary position between the regions. Hereinafter, it is assumed that the second region is a region where a higher optical zoom level is set than the first region.

First, the division unit 210 applies a predetermined maximum digital zoom level to the sub region (uppermost sub region) at the boundary with the second region among the plurality of divided sub regions in the first region, Is satisfied. The maximum digital zoom level means the adjustable maximum zoom level. Of course, Dx3 in FIG. 3 does not necessarily mean the maximum digital zoom level.

If the set image quality condition is not satisfied, the corresponding sub-area (uppermost sub-area) at the boundary is attributed to the second area and information of the first and second areas is modified. Accordingly, the size of the first area will be smaller than that of the existing area, and the size of the second area will be larger than the existing size. By using this verification process, the size or position of the first, second, and third regions can be changed little by little, and if there is no error, the existing state can be maintained without changing.

In addition, as shown in Fig. 4 below, it is possible to control the optical zoom more precisely by adding a boundary region between a plurality of regions obtained by dividing the image of the photographing means.

FIG. 4 is a diagram showing a state in which a boundary region is added and divided between a plurality of regions shown in FIG. 3; FIG.

Referring to FIG. 4, the division unit 210 according to the present invention sets optical 1x, 2x, and 3x for the first, second, and third regions, respectively , The border area between 1.5x zoom (0X1.5) and 2.5x zoom (0x2.5) can be set. The boundary area is an area set in the boundary between the main areas (first to third areas), and a middle (average) zoom value of two adjacent main areas can be set.

Accordingly, the control unit 230 can effectively detect the object with only the optical zoom level operation and acquire the feature information of the detected object without performing a separate digital zoom on the boundary region.

Next, a method of acquiring feature information according to an embodiment of the present invention will be described. 5 is a view for explaining a method for acquiring object feature information using the system shown in FIG.

First, the division unit 210 divides an area in the image acquired by the photographing unit 100 into a plurality of areas corresponding to a predetermined optical zoom level (S510). Accordingly, the received image is divided into a plurality of areas. The information of the plurality of divided regions and the optical zoom level information corresponding to each region may be stored in the storage unit 240.

Thereafter, the detecting unit 220 detects the object in the image (S520), and selects the region where the object is detected (S530). At this time, the detecting unit 220 calculates the position value of the object in the image, and selects the area including the calculated position value as the detected area of the object.

On the other hand, when the boundary region is further divided by the division unit 210, the detection unit 220 determines whether the calculated position value is included in the boundary region, and if the calculated position value is included in the boundary region, And provides it to the control unit 230.

The control unit 230 searches the storage unit 240 for the optical zoom level corresponding to the selected area, and enlarges the selected area based on the searched optical zoom level (S540). Then, the controller 230 analyzes the enlarged image to acquire feature information of the object (S550). Here, the control unit 230 may further utilize the digital zoom in addition to the optical zoom to increase the acquisition efficiency of the feature information. To do this, it is possible to acquire the feature information of the object by analyzing the image through the digital zoom of a predetermined level range.

According to the apparatus and method for obtaining object feature information according to the present invention, when an object is detected in an image, an image or a corresponding region is enlarged and provided based on an optical zoom level corresponding to the detected region, There is an advantage that information can be acquired quickly and easily.

In addition, according to the present invention, an intra-image area is divided and set in advance through a learning process so that an image is enlarged and controlled at an optimum optical zoom level corresponding to an area where an object is detected immediately after an object is detected in the image .

In addition, according to the present invention, an object can be analyzed by applying a digital zoom to an image enlarged by optical zooming, thereby improving the object recognition efficiency.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

100: photographing means 200: object characteristic information obtaining device
210: division unit 220: detection unit
230: control unit 240:

Claims (14)

An object characteristic information acquiring method of an apparatus for acquiring characteristic information of an object in an image photographed through a photographing means,
Dividing an area in an image into a plurality of areas each corresponding to a predetermined optical zoom level through a learning process using a plurality of optical zoom levels;
Selecting an area in which the object is detected among the plurality of areas when the object is detected in the image; And
And enlarging the selected area based on the optical zoom level corresponding to the selected area to obtain characteristic information of the object,
Wherein the dividing into the plurality of regions comprises:
A boundary region is added and divided between adjacent regions of the plurality of regions, and an average optical zoom level between the adjacent regions is set to an optical zoom level of the boundary region. The method according to claim 1,
Wherein the step of acquiring feature information of the object comprises:
And acquiring feature information of the object by analyzing the selected region through a digital zoom having a predetermined level range. The method according to claim 1,
Wherein the step of selecting the region comprises:
Calculating a position value of the object in the image; And
Selecting an area including the calculated position value as an area in which the object is detected
And an object feature information acquiring step of acquiring object feature information. delete The method according to claim 1,
Wherein the dividing into the plurality of regions comprises:
And dividing each of the divided regions into a plurality of sub regions corresponding to a predetermined digital zoom level through a learning process using a plurality of digital zoom levels. The method of claim 5,
When a predetermined image quality condition is not satisfied as a result of applying a predetermined maximum digital zoom level to a sub region in a border with a second region of a plurality of divided sub regions in the first region, To modify the first and second areas,
Wherein the second region is a region in which an optical zoom level higher than the first region is set. The method according to claim 1,
Wherein the feature information comprises:
If the object is a vehicle, information on at least one of a vehicle type, a vehicle number, a vehicle type, and a vehicle color,
And if the object is a person, information on at least one of sex, age, color, and shape. An object feature information acquiring apparatus for acquiring feature information of an object in an image photographed through a photographing means,
Dividing an area in an image into a plurality of areas each corresponding to a predetermined optical zoom level through a learning process using a plurality of optical zoom levels;
A detecting unit for detecting the object in the plurality of regions when the object is detected in the image; And
And a control unit for enlarging the selected area based on the optical zoom level corresponding to the selected area to obtain characteristic information of the object,
Wherein,
Wherein a boundary region is added and divided between adjacent regions of the plurality of regions, and an average optical zoom level between the adjacent regions is set to an optical zoom level of the boundary region. The method of claim 8,
Wherein,
And acquires feature information of the object by analyzing the selected region through a digital zoom having a predetermined level range. The method of claim 8,
The detecting unit includes:
Wherein the object feature information acquisition unit calculates a position value of the object in the image and selects an area including the calculated position value as an area in which the object is detected. delete The method of claim 8,
Wherein,
And divides each of the divided regions into a plurality of sub regions corresponding to a predetermined digital zoom level through a learning process using a plurality of digital zoom levels. The method of claim 12,
Wherein,
When a predetermined image quality condition is not satisfied as a result of applying a predetermined maximum digital zoom level to a sub region in a border with a second region of a plurality of divided sub regions in the first region, To modify the first and second areas,
Wherein the second area is an area in which an optical zoom level higher than the first area is set. The method of claim 8,
Wherein the feature information comprises:
If the object is a vehicle, information on at least one of a vehicle type, a vehicle number, a vehicle type, and a vehicle color,
And if the object is a person, information on at least one of sex, age, color, and shape.

Images