KR101822429B1 - Apparatus for acquiring object information and method thereof - Google Patents

Abstract

The present invention relates to an object information obtaining apparatus and a method thereof. According to another aspect of the present invention, there is provided an apparatus for acquiring object information using at least one camera installed at a predetermined position, the apparatus comprising: A control unit for controlling the second camera to estimate the second control value so that the object is located in the reference area of the second camera; and a control unit for calculating an error value between the first control value and the second control value And a correction unit that corrects the first control value based on the error value.
According to the present invention, it is possible to improve object detection and recognition efficiency by automatically correcting a matching error between cameras when object detection is difficult due to positional shift between cameras or when object recognition rate is lowered, There is an advantage that the maintenance and management of the facility is facilitated.

Description

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

The present invention relates to an apparatus and method for acquiring object information, and more particularly, to an apparatus and method for acquiring object information capable of enhancing object recognition efficiency.

In general, an object monitoring system includes a fixed camera and a rotatable camera. When the object such as a vehicle is detected through a fixed camera, the rotation camera (PTZ) Lt; / RTI >

In order to enlarge the detected object through the fixed camera, the object monitoring system recognizes the object by moving the rotating camera to the position of the object in the image. At this time, the object monitoring system recognizes / acquires the object through the matching technique between the fixed camera and the rotating camera.

In general, the object monitoring system controls the rotation / zoom of the rotating type camera based on matching information between the fixed camera and the camera obtained at the initial setting of the rotating type camera. However, if the physical location of the camera is distorted due to factors such as wind and external impact, controlling the rotating camera with existing matching information makes it difficult to shoot the object at the center of the screen, or it may deviate from the screen. As a result, it becomes difficult to search the object and the object recognition rate deteriorates.

The technology of the background of the present invention is disclosed in Korean Patent No. 1596896 (published on Feb. 23, 2013).

The object of the present invention is to provide an apparatus and method for acquiring object information capable of enhancing recognition efficiency of an object by correcting an error of object recognition due to positional deviation between cameras.

The present invention provides an object information acquiring apparatus using at least one camera installed at a predetermined point, the object information acquiring apparatus controlling a second camera based on a first control value previously stored corresponding to a position coordinate of a predetermined object in an image of the first camera A controller for controlling the second camera to estimate the second control value such that the object is located in the reference area of the second camera; and a controller for calculating an error value between the first control value and the second control value And a correction unit that corrects the first control value based on the error value.

The apparatus may further include a recognition unit that extracts feature information by image processing the object from the image of the second camera, and recognizes the object based on the feature information.

The correction unit may correct the first control value using the error value if the recognition rate calculated based on the result of the recognition unit is less than a predetermined value.

The control unit may detect and recognize the object using only the second camera when the error value exceeds the set range by a predetermined number of times or when the first camera has an error.

The object information acquisition apparatus may further include a detection unit that detects the object by analyzing the image of the first camera, wherein when the object is detected from the image of the first camera, Value can be estimated.

The object information obtaining apparatus may further include a storage unit for storing a plurality of first control values respectively corresponding to the plurality of first position coordinates, wherein the correcting unit corrects the plurality of first control values based on the error value, Each value can be corrected.

In addition, the control value may include at least one of a pan, a tilt, and a zoom.

According to another aspect of the present invention, there is provided an object information acquiring method using an object information acquiring apparatus using at least one camera installed at a predetermined point, the object information acquiring method comprising: acquiring first stored control values corresponding to position coordinates of a predetermined object in an image of a first camera Controlling the second camera based on the second control value to control the second camera so that the object is located in the reference area of the second camera, and estimating a second control value by comparing the first control value and the second control value, Calculating an error value between the control values, and correcting the first control value based on the error value.

Here, the object information acquisition method may further include a step of image-processing the object in the image of the second camera to extract the feature information, and recognizing the object based on the feature information.

In addition, the correcting step may correct the first control value using the error value when the recognition rate calculated based on the result of the recognition unit is less than a predetermined value.

The method may further include detecting and recognizing the object using only the second camera when the error value exceeds a preset range by a predetermined number of times or an error occurs in the first camera .

The method of acquiring object information may further include the step of analyzing an image of the first camera to detect the object, wherein the step of estimating the second control value comprises: And when it is detected, the second control value can be estimated.

The object information acquiring method may further include storing a plurality of first control values each corresponding to a plurality of first position coordinates, and the step of correcting may further include the step of, based on the error value, Each of the 1 control values can be corrected.

According to the apparatus and method for obtaining object information according to the present invention, it is possible to improve object detection and recognition efficiency by automatically correcting the matching error between cameras when object detection is difficult due to positional deviation between cameras or when object recognition rate is lowered have.

In addition, it is possible to automatically correct the object recognition error due to the positional change between the cameras, thereby providing the advantage that the operator does not need to physically rearrange the camera facility or the position and easily maintain and manage the facility.

1 is a view for explaining an object information acquisition system according to an embodiment of the present invention.
FIG. 2 is a detailed view of the object information obtaining apparatus of FIG. 1. FIG.
Fig. 3 is a diagram for explaining a configuration for correcting inter-camera distortion in the embodiment of the present invention. Fig.
4 is a view for explaining a configuration for calculating an error in the embodiment of the present invention.
5 is a diagram illustrating a method of acquiring object information using the apparatus of 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 information acquisition system according to an embodiment of the present invention.

As shown in FIG. 1, an object information acquisition system according to an embodiment of the present invention includes an imaging device 110 and an object information acquisition device 200, and detects an object at a predetermined point and recognizes the detected object.

The object to be recognized may correspond to at least one of a vehicle, a person, and an animal.

However, for convenience of description, the embodiment of the present invention will be described with reference to an illegal parking intermittent control system having a vehicle as a target object. Of course, the present invention is not limited to vehicle intermittent use.

The photographing apparatus 110 includes a first camera 110 and a second camera 120. The photographing apparatus 110 photographs an image at a predetermined point and provides the photographed image to the object information obtaining apparatus 200. [ At this time, the photographing apparatus 100 can transmit / receive data to / from the object information obtaining apparatus 200 through wired / wireless communication.

The first camera 110 photographs a predetermined area as a fixed camera, and provides the photographed image to the object information obtaining apparatus 200. In the case of FIG. 1, the first cameras 110-1 and 110-2 photograph two front and rear regions, respectively. Of course, at least one first camera 110 can be operated, and the number of installed cameras can be changed.

The second camera 120 is a rotating type camera including a PTZ (Pan, Tilt, Zoom) function for rotating and zooming. The second camera 120 rotates / zooms the camera so that an object detected from the image of the first camera 110 is displayed And provides the photographed image to the object information acquiring apparatus 200. The object information acquiring apparatus 200 acquires the photographed image.

The object information acquisition apparatus 200 receives an image captured by the first camera 110 and the second camera 120, detects an object from the captured image, and recognizes the object based on the detected feature information of the object.

When the object is detected in the image of the first camera 110, the object information obtaining apparatus 200 obtains the pan, tilt, and the like of the second camera 120 using the control value corresponding to the position coordinates of the detected object Tilt, and Zoom.

The object information acquisition apparatus 200 stores control values of the second camera 120 for taking close-up of each coordinate point corresponding to a plurality of coordinate points in the image of the first camera 110, respectively.

At this time, the object information obtaining apparatus 200 stores the control value of the second camera 120 for taking a close-up of the coordinate point corresponding to the coordinate point of the plurality of pixels in the photographed image of the first camera 110, . Of course, at the time of constructing the control value, the construction time is reduced and the data operation efficiency is improved by constructing the pixels selected at the setting interval without targeting all the pixels.

When the object is detected in the image of the first camera 110, the object information obtaining apparatus 200 controls the rotation / zoom of the second camera 120 based on the control value corresponding to the detected position coordinate of the object in the image do. Then, the object is detected in the image of the second camera 120 controlled, and the object is recognized based on the feature information of the detected object. In the case of a vehicle, the vehicle number is recognized based on the license plate area.

However, when a physical positional shift occurs in at least one of the first and second cameras 110 and 120 due to factors such as wind, weather, external impact, etc., the object in the captured image of the second camera 120, There is a problem that only a part of the object is detected or not detected at all, thereby making it difficult to detect the feature information of the object or the recognition rate of the object.

In this case, existing matching information established between the two cameras 110 and 120 needs to be corrected. In the embodiment of the present invention, the error of the control value due to the camera position error is calculated, and based on the calculated error, And drives the second camera 120 on the basis of the corrected control value, thereby raising the object detection and recognition rate by re-capturing the object.

Hereinafter, an object information obtaining apparatus according to an embodiment of the present invention will be described in detail with reference to FIG. Hereinafter, for convenience of explanation, the case where the object is a vehicle will be described as an example.

FIG. 2 is a detailed view of the object information obtaining apparatus of FIG. 1. FIG. 2, the object information obtaining apparatus 200 according to the embodiment of the present invention includes a storage unit 210, a detection unit 220, a control unit 230, a correction unit 240, and a recognition unit 250, .

The storage unit 210 stores and matches a plurality of first control values (pan / tilt / zoom) corresponding to a plurality of first position coordinates in the image of the first camera 110. The first control value is a value that allows the second camera 120 to take an object close-up in a reference area in an image, and includes at least one of Pan, Tilt, and Zoom.

The detection unit 220 analyzes the image of the first camera 110 to detect the vehicle in the image, and transmits the detected position coordinates of the vehicle to the controller 230 in the image. Here, the position coordinates may correspond to the center point coordinates for the vehicle area or the predetermined minutia point coordinates in the vehicle area.

The control unit 230 extracts a pre-stored first control value corresponding to the position coordinates of the vehicle from the storage unit 210, and controls the second camera 120 based on the extracted first control value. Accordingly, the second camera 120 is controlled to be rotated or zoomed according to the first control value. If the camera 120 does not change much, the second camera 120 focuses the vehicle in a predetermined area in the image.

If there is a slight difference between the cameras, the vehicle may not be properly photographed in the image even if the second camera 120 is controlled based on the first control value. The second camera 120 is controlled to estimate the second control value so that the vehicle is located in the reference area (ex.

For example, when the vehicle is detected from the image of the first camera 110, the control unit 230 controls the rotation or zoom of the second camera 120, And the second control value to be photographed at the time of photographing. At this time, if there is no error in the positional shift between the cameras or is not large, the second control value estimated for the position coordinates of the vehicle will be the same as or similar to the existing first control value.

The reason why the second control value is estimated after the second camera 120 is controlled based on the first control value is that as a result of the control using the previously stored first control value, This is because the vehicle may be photographed outside the reference area or the vehicle may not be photographed outside the image.

In this case, the corrector 240 calculates an error value between the first control value and the second control value, and corrects the first control value based on the error value. In the case of correcting the first control value by reflecting the calculated error to the first control value, it is possible to correct the control error due to the position shift between the cameras, and the object recognition rate through the image of the second camera 120 can be increased.

Here, of course, the corrector 240 may correct all the plurality of first control values based on the error value. When a positional shift occurs between the cameras, a plurality of first control values matched to a plurality of coordinates becomes invalid. The correcting unit 240 corrects each of the plurality of first control values matching each of the plurality of position coordinates based on the calculated error value, and matches each of the coordinates with each of the first control values. Accordingly, all the first control values stored in the storage unit 210, which are matched and stored for each position coordinate, can be updated.

The recognition unit 250 performs image processing of the vehicle in the detected image of the second camera 120, extracts the feature information, and recognizes the vehicle based on the feature information. The recognition unit 250 detects the license plate area as feature information through image processing and acquires the vehicle number in the license plate area as recognition information. If the vehicle is recognized in this manner, it will be possible to crack down on illegal parking regulations.

On the other hand, when the error value between the first control value and the second control value deviates from the setting range due to positional deviation between the cameras, the correction unit 240 determines that the license plate area is not detected in the image of the second camera 120 And a case where the vehicle number recognition rate is less than the set value, calculates the error of the first control value, and corrects the first control value based on the calculated error.

Thereafter, the control unit 230 controls the second camera 120 based on the corrected first control value to enable detection of the license plate area and vehicle number recognition. As described above, if the second camera 120 is re-imaged through the corrected control value, accurate recognition of the object becomes possible, and object detection or object recognition rate degradation due to positional error between cameras can be solved.

In addition, the control unit 230 can detect and recognize the object using only the second camera 120 when the error value exceeds the set range by a predetermined number of times or when the first camera 110 has an error .

Fig. 3 is a diagram for explaining a configuration for correcting inter-camera distortion in the embodiment of the present invention. Fig. 3 (a) is an image of the first camera 110, and (b) and (c) are images of the second camera 120 before and after correcting the first control value.

As shown in FIG. 3 (a), when the coordinate of the vehicle detected in the image of the first camera 110 is P, the controller 230 stores the first control value corresponding to the coordinate P And controls the rotation / zooming of the second camera 120 based on the first control value searched by the first camera 210 and the searched first control value. If there is no change between the cameras, the second camera 120 may acquire an image of focusing the vehicle at the center of the screen as shown in FIG. 3 (c).

However, if the second camera 120 is driven using the existing control value in the case where a change occurs between the cameras, it is difficult to expect the image as shown in (c) of FIG. For example, as shown in FIG. 3 (b), the vehicle can be photographed while being positioned on the outskirts of the screen. In this case, it is difficult to detect the license plate and it becomes difficult to recognize the license number. However, when photographing the first control value after correction, the second camera 120 can acquire an image as shown in FIG. 3 (c), and can correctly recognize the vehicle number.

4 is a view for explaining a configuration for calculating an error in the embodiment of the present invention. 4A is an image of the first camera 110 and FIG. 4B is a view of the image of the second camera 120 taken close to the reference area of the screen (ex, . 4 shows a road sign instead of a vehicle for convenience of explanation.

4A, the corrector 240 sets a predetermined minutiae point coordinate P (position coordinate of the object) for the object in the image of the first camera 110, The second camera 120 for controlling the PTZ of the second camera 120 through the control unit 230 and photographing a point corresponding to the minutia coordinate P from the reference area ex 120 based on the first control value.

The correction unit 240 calculates an error value between the first control value matched to the minutia point coordinate and the second control value, and corrects the first control value based on the calculated error value. At this time, not only the first control value matched to the corresponding feature point coordinate P but also the control values matched to the other coordinates may be corrected together. Since the rotation of the camera is related to the pan and tilt angles, the matched control values for each coordinate can be collectively corrected by a wrong pan or tilt error.

On the other hand, the corrector 240 calculates an error using a predetermined feature point coordinate in the above-described method at the time when the vehicle is not detected or the intermission is not performed in the image of the first camera 110, It is possible to correct the control value in advance.

At this time, the predetermined feature point may correspond to a reference feature point that can be set in the first camera image. The reference feature point is a predetermined feature point that can be arbitrarily selected in the image, and it can utilize the feature point coordinates of the road sign, the facility, the street lamp, the building sign etc. in the image, and any coordinates in the image may be used.

According to this, it is possible to correct the matching information that is periodically changed at the time when the interception is not performed or at the time when the object is not detected (lunch time, nighttime), and correction is performed from time to time. You can reduce errors.

In addition, the corrector 240 may correct the inter-camera distortion using not only the predetermined feature point coordinates in the captured image of the first camera 110 but also the predetermined feature point coordinates in the captured image of the second camera 120. Specifically, the correction unit 240 corrects the control value based on the error calculated between the coordinates of the predetermined feature point acquired in the current captured image of the second camera 120 and the coordinates of the feature point acquired in the past captured image .

For example, in the captured image of the second camera 120, the position of the guide plate 'X' has been recognized as the absolute coordinates (100,100), but the position of the guide plate 'X' When it is recognized at the point, the error between the past coordinates and the current coordinates is calculated and the control value is corrected based on the calculated error.

Here, the error may include displacement information (movement distance, direction, etc. of the coordinates) of the current coordinate relative to the past coordinates. The correction between the past and current coordinates of the minutiae of the same object corresponds to a change between the cameras. The corrector 240 corrects the matched control values for each coordinate by the pan or tilt angle corresponding to the calculated error .

In the embodiment of the present invention, the vehicle license plate area is not detected even after correcting the distortion as described above, or the recognition rate of the vehicle number from the license plate area is smaller than the set value, or the image of the first camera 110 is received The mode is switched to the single detection mode using only the second camera 120. For example, if the license plate number can not be found, or if the current # number of recognition times is 50%, use the single detection mode.

On the other hand, if the error value between the first control value and the second control value exceeds the range set by repetition (by the set number of times) or the first camera is abnormal (phenomenon such that the object is not detected or the image is not received) The controller 230 uses only the second camera 120 alone and detects the vehicle while directly changing the PTZ control value of the second camera 120 at a set interval and recognizes the vehicle number, I will crack down.

Hereinafter, a method for acquiring object information of an object information acquiring apparatus according to an embodiment of the present invention will be described in detail with reference to FIG. 5 is a diagram illustrating a method of acquiring object information using the apparatus of FIG.

5, the detecting unit 220 of the object information obtaining apparatus 200 according to the embodiment of the present invention analyzes an image of the first camera 110 to detect an illegal parking space vehicle in the image (S502) .

The control unit 230 searches the storage unit 210 for the first control value corresponding to the position coordinates of the vehicle detected in the image of the first camera 120 so that close-up photographing of the vehicle is possible, And controls the second camera 120 (S504). Here, the position coordinates of the vehicle may correspond to predetermined coordinates in the vehicle area, that is, the coordinates of the center point and the coordinates of the minutiae point.

Then, the recognition unit 250 extracts the license plate area of the vehicle from the image of the second camera 120, and recognizes the vehicle number (S506).

Here, the control unit 230 compares the recognition rate of the vehicle number calculated based on the recognition result of the recognition unit 250 with the set value (S508). At this time, the vehicle number recognition rate may mean the recognition rate of characters constituting the license plate area, or may be the latest # car number recognition rate performed on the latest # cars.

If the recognition rate is equal to or greater than the set value, the controller 230 performs illegal parking intermittently using the recognized car number and image (S510).

If the recognition rate is less than the set value, it is determined that the camera-to-camera misalignment has occurred due to the influence of the external force or the like, and the matching information constructed in the past is corrected.

For this, the controller 240 controls the PTZ of the second camera 120 to estimate the second control value so that the vehicle is located in the reference area of the first camera 110 (S512).

That is, the controller 240 controls the PTZ of the second camera 120 so that a predetermined coordinate point of the vehicle in the first camera image is located at the center of the screen, and estimates a second control value matching the coordinates.

Then, an error value is calculated by comparing the estimated second control value with a first control value that is matched with the existing coordinate (S514), and the first control value is corrected based on the error value (S516).

Thereafter, step S504 is performed again to detect the number plate of the vehicle or recognize the vehicle number from the image of the second camera 120. [ Such a recalibration process may be performed in the vehicle interrupting process and may be performed at a time when the interrupting process is not performed.

In the case where it is difficult to detect the license plate or the number despite the above-mentioned correction, the vehicle is switched to the single detection mode using only the second camera 120 and performs the vehicle detection and the number recognition while controlling the second camera 120 . Of course, even when it is detected that an image is not received from the first camera 110 due to a network error, a camera malfunction, or the like, it can be automatically switched to the intermittent detection mode of the second camera 120.

According to the object information acquiring method of the object information acquiring apparatus according to the present invention, when the object is difficult to be detected or recognized due to a physical location change between the cameras, the matching error between the cameras can be automatically corrected, .

In addition, since the object recognition error due to the positional change between the cameras is automatically corrected within the system, it is not necessary for the operator to physically rearrange the camera facilities and positions, and the facility can be easily maintained and managed.

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 apparatus 110: first camera
120: second camera 200: object information acquisition device
210: storage unit 220: detection unit
230: control unit 240:
250:

Claims (14)

An object information acquiring apparatus using at least one camera installed at a predetermined point,
A storage unit for storing a plurality of first control values for controlling a second camera corresponding to a plurality of position coordinates in an image of the first camera;
A detector for detecting an object by analyzing an image of the first camera;
Controls the second camera based on a first control value corresponding to a position coordinate of the object in the image so as to recognize the object, and controls the second camera To estimate a second control value; And
And a correction unit that calculates an error value between the first control value and the second control value and corrects each of the plurality of first control values based on the error value,
Wherein each of the first and second control values comprises:
Wherein the object information includes at least one of a pan, a tilt, and a zoom. The method according to claim 1,
Further comprising a recognition unit for extracting the feature information by image processing the object from the image of the second camera, and recognizing the object based on the feature information. The method of claim 2,
Wherein,
And corrects the first control value using the error value if the recognition rate calculated based on a result of the recognition unit is less than a preset value. The method according to claim 1,
Wherein,
Wherein the object detection and recognition unit detects and recognizes the object using only the second camera when the error value exceeds the setting range by a preset number or an error occurs in the first camera. delete delete delete A method for acquiring object information using an object information acquiring apparatus using at least one camera installed at a predetermined point,
Storing a plurality of first control values for controlling a second camera corresponding to a plurality of position coordinates in an image of the first camera;
Analyzing an image of the first camera to detect a predetermined object;
Controls the second camera based on a first control value corresponding to a position coordinate of the object in the image so as to recognize the object, and controls the second camera To estimate a second control value; And
Calculating an error value between the first control value and the second control value and correcting each of the plurality of first control values based on the error value,
Wherein each of the first and second control values comprises:
Wherein the object information includes at least one of a pan, a tilt, and a zoom. The method of claim 8,
Extracting feature information by image processing the object from the image of the second camera, and recognizing the object based on the feature information. The method of claim 9,
Wherein the correcting comprises:
And correcting the first control value using the error value if the recognition rate calculated based on the recognition result is less than a predetermined value. The method of claim 8,
Further comprising detecting and recognizing the object using only the second camera when the error value exceeds the set range by a preset number or an error occurs in the first camera. delete delete delete

Images