KR20190002281A - Method for removing processing target object in image and apparatus for performing the method - Google Patents

Abstract

The present invention relates to a method for removing a processing target object from an image and an apparatus for performing the same. The method for removing a processing target object from an image comprises the steps of: determining a processing target object in an image by an image processing apparatus; and post-processing the image on the processing target object by the image processing apparatus, wherein the processing target object can be an object in which imaging is not intended.

Description

TECHNICAL FIELD The present invention relates to a method for removing an object to be processed from an image, and a device for performing the method.

The present invention relates to a method for removing an object to be processed from an image and an apparatus for performing such a method. And more particularly, to an image processing method and apparatus for eliminating an object unnecessarily (or improperly) included in moving image information captured by a video processing apparatus through post-image processing.

An omnidirectional imaging system refers to an imaging system capable of recording image information in all directions (360 degrees) based on a specific point in time. Since the field-of-view image can be obtained in comparison with the conventional image system, it has recently been applied to a field of view such as a computer vision, a mobile robot, a surveillance system, a virtual reality system, a pan-tilt-zoom ) Applications such as cameras and video conferencing are becoming more and more widespread.

Various methods can be used to obtain omni-directional images. For example, an omnidirectional image may be generated by joining an image obtained by rotating one camera with respect to an optical axis satisfying a single view point. Or a method of arranging a plurality of cameras in an annular structure and combining images obtained from the respective cameras may be used. The user can generate omni-directional images using various omni-directional image processing apparatuses (or omnidirectional image processing cameras, 360 degree cameras).

When an omnidirectional image is generated by an omnidirectional image processing apparatus, unnecessarily (or improperly) captured objects may exist, and such unnecessary (or improperly) captured objects may require post-processing.

It is an object of the present invention to solve all the problems described above.

In addition, the present invention is a method for determining an object that is unnecessarily (or inappropriately) picked up as an object to be processed in an omnidirectional image picked up by an omnidirectional image processing apparatus, It is intended to process through.

Another object of the present invention is to generate a natural omni-directional image even after removal of an object to be processed through image interpolation after image post-processing for an object to be processed in an omnidirectional image.

In order to accomplish the above object, a representative structure of the present invention is as follows.

According to an aspect of the present invention, there is provided a method of removing an object to be processed from an image, the method comprising: an image processing apparatus determining the object to be processed in the image; and an image post- The object to be processed may be an object for which imaging is not intended.

According to another aspect of the present invention, an image processing apparatus for removing an object to be processed from an image includes a communication unit for communication with an external device and a processor operatively connected to the communication unit, The object to be processed, which can be implemented to determine the object to be processed and perform image post-processing on the object to be processed, may be an object for which imaging is not intended.

According to the present invention, an object unnecessarily (or inappropriately) picked up in an omnidirectional image picked up by an omnidirectional image processing apparatus is determined as a processing object, and the object to be processed is subjected to post- Can be processed.

In addition, according to the present invention, it is possible to generate a natural omni-directional image even after removal of an object to be processed through image interpolation after image post-processing for an object to be processed in an omnidirectional image.

1 is a conceptual diagram illustrating an omni-directional image processing apparatus according to an embodiment of the present invention.
FIG. 2 is a conceptual diagram illustrating the characteristics of a plurality of image pickup units positioned in an omni-directional image processing apparatus according to an embodiment of the present invention.
3 is a conceptual diagram showing an image pickup line of a plurality of image pickup units according to an embodiment of the present invention.
4 is a conceptual diagram showing imaging lines of a plurality of image sensing units according to an embodiment of the present invention.
5 is a conceptual diagram illustrating an image processing method for an object to be processed according to an embodiment of the present invention.
6 is a conceptual diagram illustrating a method of determining a determination subject area according to an embodiment of the present invention.
7 is a conceptual diagram illustrating a method for determining a determination subject area according to an embodiment of the present invention.
8 is a conceptual diagram illustrating a method of determining an object to be processed according to an embodiment of the present invention.
9 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.
10 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.
11 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.
12 is a conceptual diagram illustrating an operation of an omni-directional image processing apparatus according to an embodiment of the present invention.

The following detailed description of the invention refers to the accompanying drawings, which illustrate, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, the specific shapes, structures, and characteristics described herein may be implemented by changing from one embodiment to another without departing from the spirit and scope of the invention. It should also be understood that the location or arrangement of individual components within each embodiment may be varied without departing from the spirit and scope of the present invention. Therefore, the following detailed description is not to be taken in a limiting sense, and the scope of the present invention should be construed as encompassing the scope of the appended claims and all equivalents thereof. In the drawings, like reference numbers designate the same or similar components throughout the several views.

Hereinafter, various 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.

Hereinafter, the image processing apparatus of the embodiment of the present invention may include an omni-directional image processing apparatus. The omni-directional image processing apparatus may include an omni-directional camera (360-degree camera) capable of imaging omni-directional (or 360-degree images).

Hereinafter, the image information and moving image information disclosed in the embodiments of the present invention may include an omni-directional image (or a 360-degree image).

1 is a conceptual diagram illustrating an omni-directional image processing apparatus according to an embodiment of the present invention.

1, a structure of an omnidirectional image processing apparatus is disclosed.

Referring to FIG. 1, the omnidirectional image processing apparatus 100 may have a wearable structure similar to a necklace that can be hung on a user's neck. The omnidirectional image processing apparatus 100 may have a shape of an open necklace on one side or a shape of a necklace on one side as shown in FIG. Hereinafter, in the embodiment of the present invention, it is assumed that the omnidirectional image processing apparatus 100 has a U-shape with one side open. The omni-directional omni-directional image processing apparatus 100 can be hung on a user's neck in the form of a wearable device to take omni-directional images.

In the embodiment of the present invention, it is assumed that the omnidirectional image processing apparatus 100 is hooked on the neck of a user as a necklace (or a necklace-shaped, U-shaped opening). However, the omnidirectional image processing apparatus 100 may not be merely a form hanging on the user's neck. For example, the omni-directional image processing apparatus 100 may be installed in other body parts or external objects (or objects) / devices / structures of the user in various forms that can be hooked / attached to obtain omni-directional images.

The user can acquire a plurality of images for creating an omnidirectional image in a state where both hands are free from the omnidirectional image processing apparatus 100 implemented as a wearable device.

The omnidirectional image processing apparatus 100 may include a plurality of image pickup units. Each of the plurality of image sensing units may be positioned at a predetermined interval (or a predetermined interval) in the omnidirectional image processing apparatus to separately capture images corresponding to the angle of view / image sensing line. Although the position of each of the plurality of image pickup units may be fixed in the omnidirectional image processing apparatus 100, each of the plurality of image pickup units may be movable, and the position of each of the plurality of image pickup units may be changed.

For example, the omnidirectional image processing apparatus 100 may include three image pickup units, and the three image pickup units may have a field of view (for example, 120 to 180 degrees) Can be picked up. The three image capturing units may be the image capturing unit 1 (110), the image capturing unit 2 (120), and the image capturing unit 3 (130).

Hereinafter, a structure in which three image pickup sections are included in the omnidirectional image processing apparatus 100 is described for convenience of explanation. However, a plurality of image pickup units (for example, 2, 4, 5, 6, etc.) other than three may be included in the omnidirectional image processing apparatus 100 to capture an omnidirectional image, And are included in the scope of the invention.

The image capturing unit 1 (110), the image capturing unit 2 (120), and the image capturing unit 3 (130) can capture an image according to the angle of view. The image 1 is generated by the image capturing unit 1 110 on the same time resource, the image 2 is generated by the image capturing unit 2 120, and the image 3 is generated by the image capturing unit 3 130 . The view angles of the image capturing unit 1 110, the image capturing unit 2 120 and the image capturing unit 3 130 may be 120 degrees or more and overlapping image regions may exist in the images 1, 2, . Thereafter, the omnidirectional image can be generated by stitching / correcting the image 1, the image 2, and the image 3 picked up on the same time resource by the omnidirectional image processing apparatus 100. The stitching and / or correction procedure for a plurality of images may be performed by the omni-directional image processing apparatus itself or may be performed based on a user apparatus (smart phone) capable of communicating with the omni-directional image processing apparatus 100. That is, the additional image processing procedures for the generated multiple images may be performed by the omnidirectional image processing apparatus 100 and / or another image processing apparatus (smart phone, personal computer, etc.).

Hereinafter, specific features of the omnidirectional image processing apparatus and omnidirectional image generating method are specifically disclosed.

FIG. 2 is a conceptual diagram illustrating the characteristics of a plurality of image pickup units positioned in an omni-directional image processing apparatus according to an embodiment of the present invention.

2, features of a plurality of image pickup units positioned in a U-shaped omni-directional image processing apparatus are disclosed. The position of the image pickup section disclosed in Fig. 2 is an exemplary one. Each of the plurality of image pickup units may be located at various positions on the omnidirectional image processing apparatus to pick up a plurality of images for generating omni-directional images.

At the top of Fig. 2, the rear portion of the omnidirectional image processing apparatus is started.

The image pickup unit 1 210 and the image pickup unit 2 220 included in the omnidirectional image processing apparatus may be located in a curved portion where the curvature exists in the omnidirectional image processing apparatus. Specifically, when the user wears the omnidirectional image processing apparatus as a wearable apparatus, the image pickup unit 1 210 and the image pickup unit 220 can be positioned in the bent region contacting the back of the neck. For example, when the image pickup unit 1 210 and the image pickup unit 220 are positioned at a certain distance on the basis of a maximum curvature point (for example, a U-shaped middle portion) of the U-shaped omni-directional image processing apparatus Can be located.

The image capturing unit 210 may capture an area including a left rear rectangular area on the basis of a line of sight direction of the user. The image capturing unit 2 (220) can capture an area including a rear right square area on the basis of the user's line of sight. Specifically, the image pickup section 210 (210) has a first angle of view and can perform image pickup with respect to an area corresponding to the first angle of view. The image pickup section 2 (220) has a second angle of view, and can perform imaging with respect to an area corresponding to the second angle of view. For example, the first angle of view and the second angle of view may be 120 to 180 degrees.

When the imaging by the image pickup unit 1 210 and the image pickup unit 220 is performed, a first overlapping area 215 overlapping with the first view angle and the second view angle may occur. Then, the omnidirectional image can be generated based on the stitching considering the overlap area.

2, the front side of the omnidirectional image processing apparatus is started.

The image sensing unit 3 230 may be positioned on the front of the omnidirectional image processing apparatus. Specifically, the image capturing unit 3 230 may be located at the distal end (the end portion of the U-shape) of the omnidirectional image processing apparatus. When the user places the omnidirectional image processing apparatus on the neck as a wearable apparatus, the distal end portion of the U-shaped omnidirectional image processing apparatus may be positioned in the front direction of the user (the direction in which the user's gaze is viewed). The omnidirectional image processing apparatus may include a first end portion and a second end portion, and the image pickup portion 3 230 may be located at one end portion of the first end portion and the second end portion.

The image capturing unit 3 230 may perform image capturing in the same direction as the direction of the user's gaze and perform image capturing of an area corresponding to the user's gaze.

More specifically, the image pickup section 3 230 has a third angle of view, and can perform imaging with respect to an area corresponding to the third angle of view. For example, the third angle of view may be between 120 and 180 degrees. When the image pickup by the image pickup section 3 230 is performed, the first overlap angle 225 of the image pickup section 1 210 and the third overlap angle 225 of the image pickup section 3 230 may occur have. When the imaging by the imaging unit 3 230 is performed, the second angle of view of the imaging unit 2 (220) and the third angle of overlap 235 of the imaging angle of the imaging unit 3 (230) have.

When the omnidirectional image processing apparatus is hooked on the neck of the wearable apparatus, the image pickup unit 1 210 and the image pickup unit 2 220 are positioned at a relatively higher position than the image pickup unit 3 230 Can be located. Further, the image pickup section 3 (230) is located at only one end portion.

In the conventional omni-directional image processing apparatus, a plurality of image pickup units positioned at the same height are implemented with a certain angle, whereas the omnidirectional image processing apparatus according to the embodiment of the present invention has different angles between the plurality of image pickup units, can be different. Therefore, the sizes / shapes of the first overlapping area 215, the second overlapping area 225, and the third overlapping area 235 may be different from each other for a plurality of images generated by each of the plurality of image pickup units.

The image pickup unit 1 210, the image pickup unit 2 220 and the image pickup unit 3 230 taking the first overlapping area 215, the second overlapping area 225 and the third overlapping area 235 into consideration, An omnidirectional image can be generated based on image processing procedures (stitching / correction, etc.) for the image 1, the image 2, and the image 3 generated by each.

The sizes of the first angle of view, the second angle of view, and the angle of the third angle of view may be set to be the same, but they may be set differently, and these embodiments are also included in the scope of the present invention.

3 is a conceptual diagram showing an image pickup line of a plurality of image pickup units according to an embodiment of the present invention.

In Fig. 3, an image pickup line of each of a plurality of image pickup units provided in the omnidirectional image processing apparatus is started. When the ground is assumed to be parallel to the X-Z plane formed by the X-axis and the Z-axis, the image capturing lines are arranged on the space represented by the X-axis / Y-axis / Z- As shown in Fig.

Conventional omni-directional image processing apparatuses can be implemented with a plurality of image pickup units at the same height at a certain angle (for example, 120 degrees). In this case, a plurality of imaging lines of the plurality of image pickup units included in the conventional omni-directional image processing apparatus are arranged in parallel to the ground (or XZ plane) and have a plurality of (for example, 120 degrees) .

As described above, the omnidirectional image processing apparatus according to the embodiment of the present invention is configured so that the heights of the plurality of image pickup units (or the positions at which the plurality of image pickup units are implemented) and the angles between the plurality of image pickup units May differ from each other at the time of imaging. Therefore, the characteristics of the imaging lines of the omni-directional image processing apparatus according to the embodiment of the present invention are different from those of the imaging lines of the conventional omni-directional image processing apparatus.

The image pickup lines of each of the plurality of image pickup units shown in Fig. 3 are illustrative examples for showing the difference (e.g., height, angle) between the pickup lines of each of the plurality of image pickup units due to the characteristics of the wearable device have. 3 may be an imaging line in the case where there is no movement by a user wearing the omnidirectional image processing apparatus or when the omnidirectional image processing apparatus is fixed in a specific state.

The top of Fig. 3 starts imaging lines of the image sensing unit 1 (310) and the image sensing unit 2 (320).

The image pickup unit 1 310 and the image pickup unit 2 320 can be implemented at a relatively higher position than the image pickup unit 3 330. In the case where the user wearing the omnidirectional image processing apparatus is assumed to be in the Y-axis direction, the image pickup unit 1 310 and the image pickup unit 2 320 in the omnidirectional image pickup apparatus structured to wear on the neck The curved portion (curve / central portion in the U-shape) relatively rises and the leg portion (the end portion in the U-shape) where the image pickup portion 3 330 is located can be relatively lowered.

For example, the image pickup line 1 315 of the image pickup unit 1 310 is parallel to the XZ plane, and has a first angle 1 with the X axis at the coordinate a of the Y axis, a second angle with the Y axis, It can have an angle.

The image pickup line 2 325 of the image pickup unit 2 320 may have a fourth angle with the X axis at the point a in the Y axis, a fifth angle with the Y axis, and a sixth angle with the Z axis in parallel with the XZ plane .

3, the image pickup line 3 (335) of the image pickup section 3 (330) is parallel to the XZ plane, and has a seventh angle with the X axis at the coordinate b of the Y axis, an eighth angle with the Y axis, And a ninth angle. b may be a value less than a. The imaging line 3 335 of the image sensing unit 3 330 may be configured to be parallel to the XZ plane and to look at the entire surface (for example, a direction perpendicular to the XY plane) as the user's line of sight.

That is, the imaging line 1 (315) and the imaging line 2 (325) have the same height with respect to the Y axis and the imaging line 3 (335) is located at a position relatively lower than the imaging line 1 and the imaging line 2 can do. The imaging line 1 315, the imaging line 2 325, and the imaging line 3 335 disclosed in FIG. 3 are examples of imaging lines having different characteristics, and various other imaging lines are defined and an omni-directional image Can be picked up.

4 is a conceptual diagram showing imaging lines of a plurality of image sensing units according to an embodiment of the present invention.

In Fig. 4, imaging lines of a plurality of image pickup units different from those in Fig. 3 are started. Similarly, in FIG. 4, it is assumed that the paper surface is parallel to the XZ plane formed by the X axis and the Z axis.

4, the image pickup lines of the image pickup section 1 (410) and the image pickup section 2 (420) are started.

The image pickup unit 1 410 and the image pickup unit 2 420 may be positioned at a relatively higher position than the image pickup unit 3 430. Likewise, when it is assumed that the direction in which the user is standing is the Y-axis direction, the constructional omnidirectional image pickup device of the wearable device hanging on the neck has a curved portion ( The curved portion in the U-shape) relatively rises and the leg portion (the end portion in U-shape) where the image pickup unit 3 430 is located can be relatively downwardly captured.

For example, the image pickup line 1 415 of the image pickup unit 1 (410) is parallel to the XZ plane, and has a first angle 1 with the X axis at the coordinate a of the Y axis, a second angle with the Y axis, It can have an angle.

The imaging line 2 415 of the image sensing unit 2 420 may have a fourth angle with the X axis, a fifth angle with the Y axis, and a sixth angle with the Z axis in the coordinate a of the Y axis, parallel to the XZ plane .

The lower end of Fig. 4 starts the image pickup line of the image pickup section 3 (430).

The image pickup line 3 435 of the image pickup section 3 430 may not be parallel to the XZ plane and the x-axis and the seventh angle, the y-axis and the eighth angle, and the Z- It may have a ninth angle.

Since the image capturing unit 3 430 is located at the distal end of the omnidirectional image processing apparatus, the image capturing line is not parallel to the XZ plane and can have a certain angle (for example, 0 to 30 degrees) with the XZ plane.

That is, the imaging line 1 415 and the imaging line 2 425 have the same height with respect to the Y axis, and the imaging line 3 435 has the same height as the imaging line 1 415 and the imaging line 2 425 It can be located at a relatively low position. The imaging line 1 415 and the imaging line 2 425 are parallel to the XZ plane, but the imaging line 3 435 may not be parallel to the XZ plane.

In another embodiment of the present invention, for example, the image pickup line 1 of the image pickup section 1 forms a first angle with the XZ plane, and the coordinate a of the Y axis as a starting point, A second angle, and a third angle with the Z axis. The image pickup line 2 of the image pickup section 2 forms a first angle with the XZ plane and has a fourth angle with the X axis, a fifth angle with the Y axis, and a sixth angle with the Z axis with the coordinate a of the Y axis as the starting point Lt; / RTI > The image pickup line 3 of the image pickup section 3 forms a second angle with the XZ plane and has the seventh angle with the X axis and the eighth angle with the X axis and the ninth angle with the Z axis with the coordinate b of the Y axis as the starting point have.

In another embodiment of the present invention, for example, the image pickup line 1 of the image pickup section 1 forms a first angle with the XZ plane, and the coordinate a of the Y axis as a starting point, And a third angle with the Z axis. The image pickup line 2 of the image pickup unit 2 forms a second angle with the XZ plane and has a fourth angle with the X axis, a fifth angle with the X axis, and a sixth angle with the Z axis with the coordinate a of the Y axis as the starting point Lt; / RTI > The image pickup line 3 of the image pickup section 3 forms a third angle with the XZ plane and has the seventh angle with the X axis and the eighth angle with the X axis and the ninth angle with the Z axis with the coordinate b of the Y axis as the starting point have.

In other words, the image processing apparatus according to the embodiment of the present invention is different from the image processing apparatus in which the image pickup lines of the plurality of image pickup sections have the same angle as the ground at the same Y-axis point, The imaging lines are located at different Y-axis points and can have different angles from the ground (or XZ plane).

The wearable omni-directional image processing apparatus according to the embodiment of the present invention has a difficulty in freely specifying a subject desired to be photographed, and there is a possibility of unnecessary photographing of a body part. There is a possibility of photographing a body part like the object 450 in Fig.

For example, when the omni-directional image processing apparatus is photographed, a body part close to the omni-directional image processing apparatus can be photographed. If the omnidirectional imaging device is located near the head, if the image of the hair or the like is included in the captured omnidirectional image and if the omnidirectional imaging device is in the neck, the omnidirectional image .

If a part of the body captured in the omnidirectional image is included, it may interfere with appreciation of the omnidirectional image. Therefore, the function of judging and automatically excluding objects (for example, body parts) unnecessarily (or improperly) included in the omnidirectional image, or by designating a valid shooting region and photographing the object, It is necessary that a function that can be implemented in the omnidirectional image processing apparatus is implemented.

Hereinafter, in an embodiment of the present invention, an object unnecessarily (or inappropriately) included in an omnidirectional image may be expressed by the term " object to be processed ". The effective shooting region may be a region other than the shooting region including the object to be processed.

The determination and processing of the object to be processed may be performed by a separate external device that receives and processes the omnidirectional image information from the omnidirectional image processing apparatus or the omnidirectional image processing apparatus. Hereinafter, an apparatus for determining an object to be processed and performing an image post-processing may be expressed by the term image processing apparatus. The image processing apparatus may include an omnidirectional image processing apparatus and / or an external apparatus.

An image processing apparatus for processing an object to be processed in an image may include a communication unit for communication with an external device and a processor operatively connected to the communication unit. The processor can perform image processing on an object to be processed, which will be described later in the present invention.

Hereinafter, a method for removing an object to be processed from an omnidirectional image is disclosed in the embodiments of the present invention. However, the present invention can also be applied to a method for removing an object to be processed from a general image rather than an omnidirectional image , These embodiments are also included in the scope of the present invention.

Hereinafter, a determination method for an object to be processed and an image post-processing method for the object to be processed are specifically disclosed.

5 is a conceptual diagram illustrating an image processing method for an object to be processed according to an embodiment of the present invention.

Referring to FIG. 5, a determination target region may be determined in an omnidirectional image to determine whether an object to be processed exists (Step S500).

The determination target region may be an area for determining the presence or absence of the processing target object on the omnidirectional image. In other words, the determination target area may be a region including an object to be processed which is not intended to be imaged by the omnidirectional image processing apparatus in the omnidirectional image.

The region to be judged can be determined based on various methods. The determination subject area may be set in advance, or may be determined based on the imaging distance, or may be determined by learning. The determination method of the determination target area will be described later. It is possible to quickly determine whether a judgment object area is set and whether a processing object exists in the judgment object area.

The presence or absence of the object to be processed can be determined in the determination target area (step S510).

The object to be processed may be an object that is captured unnecessarily (or improperly) in a forward image. As described above, when the omnidirectional image processing apparatus is a wearable apparatus, a specific body part (jaw, shoulder, etc.) of the user may be unnecessarily (or improperly) picked up and included in the omnidirectional image. In other words, the object to be processed may be an object for which imaging is not intended.

The object to be processed may be a predetermined body part or an object that is imaged close to or below a critical distance based on the position of the image pickup unit. For example, a part such as a chin or a shoulder can be picked up at a close distance equal to or less than a critical distance based on the position of the image pickup unit, and an object picked up at a distance not exceeding the critical distance can be determined as a processing object.

Alternatively, the object to be processed may be an object which is imaged close-up at a critical distance or less based on the position of the image pickup unit and is out of focus.

An image post-process for the object to be processed can be performed (step S520).

If there is an object to be processed in the determination target area, the image post-processing may be performed on the object to be processed.

Specifically, the forward direction image includes a plurality of frames, and a post-process may be performed on a frame including the object to be processed among a plurality of frames. Specifically, the object to be processed may be removed or the object to be processed may be interpolated based on the surrounding image information of the frame including the image information of the other frame and the object to be processed. The image post-processing method will be described in detail later.

If the object to be processed does not exist in the judgment object area, an omnidirectional image can be provided without additional post-processing of the object to be processed.

Also, according to the embodiment of the present invention, the post-processing of the object to be processed may be performed after determining whether or not the object to be processed exists directly in the omnidirectional image without determining the determination object area. That is, the presence or absence of the object to be processed can be determined without setting a separate region in the forward direction image, and image post-processing can be performed on the object to be processed.

6 is a conceptual diagram illustrating a method of determining a determination subject area according to an embodiment of the present invention.

FIG. 6 illustrates a method for determining a determination target area for searching for a processing object in an omni-directional image.

Referring to FIG. 6, when the user wears the omnidirectional image processing apparatus on his neck, the omnidirectional image processing apparatus can pick up an object 600 (e.g., a shoulder region image of the user). That is, the omnidirectional image may include the object 600 to be processed.

The determination target area 650 may be an area including the processing object 600 that is not intended to be imaged by the omni-directional image processing apparatus in the omnidirectional image. In other words, the determination target area 650 may be defined as an area for determining whether the processing target object 600 exists in the omnidirectional image. That is, the determination subject region 650 may be a predefined region for determining whether the processing object 600 exists in the forward direction image.

Alternatively, the determination subject region 650 may be a region (or an out of focus region) extracted from a region of interest to be imaged based on the current omni-directional image processing apparatus. The region of interest may be a focused region of the omnidirectional image processing apparatus or a range of regions based on the center of the omni-directional image to be imaged. Alternatively, the determination subject area 650 may be a background area having a low degree of association with an object positioned at an imaging center of the omnidirectional image.

Alternatively, the determination subject area 650 may be an area where the imaging distance is closer to a critical distance (for example, 30 cm). The imaging distance may be the distance between the imaging unit and the object of the omnidirectional image processing apparatus. For example, when an omnidirectional image processing apparatus such as the omnidirectional image processing apparatus disclosed in the above-described embodiments of the present invention is located on the neck, the shoulder of a user wearing the omnidirectional image processing apparatus may have an imaging distance And an area corresponding to the imaging distance within the critical distance may be determined as the determination subject area.

7 is a conceptual diagram illustrating a method for determining a determination subject area according to an embodiment of the present invention.

In Fig. 7, a method for determining a determination target area including an image to be processed in an omnidirectional image is disclosed. In Fig. 7, a method for determining a determination target area based on learning is disclosed.

Referring to FIG. 7, in consideration of the structure of the omni-directional image processing apparatus, there is a high possibility that a specific body part (for example, a shoulder) of the user is continuously imaged on a specific area of the omni-directional image. In this case, an area where the image of the user's specific body part is highly likely to be located on the omnidirectional image may be set as the determination object area 700. [

An area in which a user's shoulder image is likely to be located may be set in advance in the determination target area 700 or a determination target area 700 is set through learning about an area in which a user's shoulder image is likely to be positioned .

For example, omnidirectional images captured by the omnidirectional image processing apparatus are collected, and learning about a shoulder image existing as a processing object in the collected omnidirectional image can be performed. In this case, the determination target area 700 in which the processing object exists may be adaptively changed through learning. It is possible to determine the region where the possibility of existence of the object to be processed is higher than the threshold percentage is determined as the judgment object region 700. [

8 is a conceptual diagram illustrating a method of determining an object to be processed according to an embodiment of the present invention.

In FIG. 8, a method for determining an object to be processed in a determination object region of an omnidirectional image is disclosed.

Referring to FIG. 8, the object 840 to be processed may be preset. As described above, when the omnidirectional image processing apparatus is a wearable apparatus, a specific body part (jaw, shoulder, etc.) can be picked up as the object 840 to be processed.

The presence or absence of the processing object 840 in the determination object region or the entire omnidirectional image region in which the processing object 840 may be located in the omnidirectional image may be determined. When the object 840 exists in the omnidirectional image, the object 840 is removed through the post-processing of the image, or the object 840 is extracted from the surrounding image information of the frame including the object 840 And / or based on image information of another frame. The image post-processing method will be described in detail later.

The template image for the object 840 to be processed may be stored in advance as the reference object image 800 in order to determine whether or not the object 840 exists in the determination object area. If the object image to be processed is an image of a jaw, a shoulder, etc., of a user unnecessarily imaged due to the structure of the omnidirectional processing device, the reference object image 800 may be displayed on the jaw, shoulder, And may be an image serving as a criterion for discriminating an image of the user.

It can be determined whether the determination target object is the processing object 840 based on a comparison between the reference processing object image 800 and the determination object image 820 included in the captured omnidirectional image.

For example, when object 1, object 2, and object 3 exist in the determination target area, object 1, object 2, and object 3 are objects to be determined, and images of object 1, object 2, Image 820. < / RTI >

The degree of similarity (or degree of matching) between the judgment object object image 820 and the reference processing object image 800 is analyzed. If the degree of similarity is high, the judgment object can be determined as the processing object.

According to the embodiment of the present invention, learning on the object to be processed can also be performed. As described above, the reference image to be processed 800 can be updated through continuous learning of the object to be processed included in the omnidirectional image, and the accuracy of determination of the object to be processed can be increased.

If the object to be processed 840 is not set in advance, the following image may be set as the determination target object.

For example, the object to be judged may be 1) an object which exists for a predetermined time in the judgment subject area, 2) an object which repeatedly appears in the judgment subject area, or 3) An object that exists during the critical time or appears repeatedly in the area to be judged, 4) an object that exists during the threshold time in the omnidirectional image, 5) an object that repeatedly appears in the omnidirectional image, or 6) But may be an object that exists or appears repeatedly in the omnidirectional image for a critical period of time.

If there is an object to be judged as described above, the user who has performed the omnidirectional image capturing can request the judgment of whether the object is post-processed.

For example, when the hair of the user and the jaw image of the user are determined as the determination object, the user may be asked to confirm whether to perform the image post-processing on the user's hair and the jaw image of the user. If the post-processing of the object is requested from the user, the post-processing of the object can be performed.

9 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.

9, an image post-processing method for an object to be processed is disclosed. In particular, a method for performing interpolation on an image region including a processing object based on peripheral image information of a target frame including an object to be processed is disclosed.

Referring to FIG. 9, an omnidirectional image may be generated through separate image interpolation for an area including an object to be processed. The frame including the object to be processed is represented by the term of the object frame 900 and the region for performing the image interpolation including the object to be processed may be expressed by the term of the object region 920. [

Specifically, a post-processing all-directional image may be generated by performing image interpolation based on surrounding pixel information of a target frame.

When the interpolation object area 920 of the interpolation object frame 900 includes the object to be processed, the interpolation object area 920 is selected based on the pixel information located in the outer area 940 of the interpolation object area 920 Interpolation can be performed. For example, the outer area 940 may be an area where n pixels (where n is a natural number) are located in the opposite direction of the center of the object to be processed with respect to the outline of the interpolation object area 920, A region including pixels located at a certain distance in a direction opposite to the center direction of the object to be processed.

The interpolation of the interpolation object area 920 based on the pixel information located in the outer area 940 can be performed based on various methods.

The interpolation subject area 920 is divided into a plurality of sub-interpolation subject areas, and each of the plurality of sub-interpolation subject areas can be interpolated based on pixel information (luminance information, chrominance information, etc.) of the nearest outer area.

The division of a plurality of lower interpolation target areas may be performed in consideration of pixel characteristics of the outer area 940 in order to increase interpolation accuracy. For example, the division of a plurality of sub-interpolation subject areas may be performed in consideration of the degree of change of the pixel characteristics of the outer area 940. [ Specifically, it is determined whether or not the degree of change is equal to or greater than a predetermined threshold change degree in consideration of the brightness information of the pixels included in the outer region 940 and the degree of change of the color difference information, and the division on the outer region 940 may be performed. This division can improve interpolation accuracy.

10 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.

In Fig. 10, an image post-processing method for an object to be processed is disclosed. FIG. 10 illustrates a method of interpolating an image region including an object to be processed in consideration of image information of a previous frame based on a target frame including an object to be processed.

Referring to FIG. 10, an n-th frame is an interpolation object frame 1000, and an interpolation object area 1020 may include an object to be processed. The n-a frame (where a is a natural number), which is a reference frame 1050, may be a frame to be referred to for image post-processing of the interpolation object frame 1000. The reference frame 1050 may be a frame generated within the threshold time based on the generation time of the interpolation object frame 1000. [ Alternatively, the reference frame 1050 may be a frame to be referred to for decoding / encoding the interpolation object frame 1000, or may be a frame having relatively high similarity of the image characteristic information. Interpolation can be performed on the interpolation object area 1020 based on the image information of the reference area included in the reference frame 1050. [

For convenience of explanation, one reference frame 1050 and a reference area 1070 are disclosed. However, the reference frame and the reference area may be plural, and the video information of a plurality of reference frames may be used for interpolation for the interpolation object frame 1000 It can also be utilized.

The reference frame 1050 may be determined in consideration of the image information of the interpolation object frame 1000. If the degree of similarity between the video information of the reference frame and the video information of the frame to be interpolated is not high, the sense of heterogeneity of the video at the time of interpolation may become larger. In this case, the image interpolation procedure based on the reference area may not be performed.

A frame in which the degree of similarity is determined to determine the reference frame 1050 may be expressed by the term reference candidate frame. The reference candidate frame may include a reference candidate region corresponding to the interpolation subject region 1020. [ The reference candidate frame may be determined as the reference frame 1050 and the reference candidate region may be determined as the reference region 1070 based on the similarity between the interpolation target frames.

The similarity between the image information of the reference candidate frame and the image information of the interpolation object frame 1000 is based on the pixel characteristic information of the outer region of the reference candidate region of the reference candidate frame and the pixel characteristic information of the outer region of the interpolation object region 1020 Can be determined. Specifically, the similarity between the pixel characteristic information (e.g., luminance information, chrominance information) of the outer region of the reference candidate region and the pixel characteristic information (e.g., luminance information, chrominance information) of the outer region of the interpolation subject region 1020 Can be determined.

As described above, the outer area is an area in which n pixels (where n is a natural number) are positioned in the opposite direction of the center of the object to be processed with respect to the outline of the object to be interpolated 1020, And a pixel located at a certain distance in the opposite direction of the center of the pixel.

The degree of similarity can be determined by comparing the pixel characteristic information of the interpolation subject area 1020 with the pixel characteristic information of the reference candidate area. Alternatively, the outer region of the interpolation subject region 1020 may be divided into a plurality of first lower outer regions, the candidate outer region of the reference candidate region may be divided into a plurality of second lower outer regions, The degree of similarity may be determined through comparison of the pixel characteristic information corresponding to each of the plurality of second lower outer regions. When the similarity degree is equal to or greater than the threshold similarity degree, the reference candidate frame is determined to be the reference frame 1050, and the reference candidate region can be determined to be the reference region 1070. [

The interpolation of the interpolation object area 1020 based on the image information located in the reference area 1070 can be performed based on various methods.

The image information located in the reference area can be utilized as the image information of the interpolation object area 1020. [ Alternatively, the interpolation subject area 1020 may be divided into a plurality of lower interpolation subject areas, and each of the plurality of lower interpolation subject areas may be interpolated based on pixel information (luminance information, chrominance information, etc.) of the nearest outer area. In order to increase the interpolation accuracy, the division of the plurality of sub-interpolation target areas may be performed in consideration of the pixel characteristics of the outer area. For example, the division of a plurality of sub-interpolation subject areas may be performed in consideration of the degree of change of the pixel characteristics of the outer area.

A method of interpolating the interpolation object area based on the image information of the interpolation object frame shown in FIG. 9 and a method of performing the image interpolation based on the image information of the reference area in FIG. 10 may be used in combination.

For example, interpolation can be performed on the interpolation object area in consideration of both the pixel information located in the outer area of the interpolation object area 1020 and the pixel information of the reference area 1070. [ More specifically, the central part of the interpolation object area 1020 is interpolated considering the pixel information of the reference area 1070, and the outer part of the interpolation object area 1020 excluding the center part is interpolated to the outer area of the interpolation object area 1020 And can be interpolated considering the pixel information located.

Alternatively, deletion may be performed for an area including the object to be processed without interpolation. The area to be deleted for the area including the object to be processed without interpolation may be the removal object area. For example, when a shoulder image of a user is continuously captured in a certain area of an omnidirectional image, the shoulder image is determined as the object to be processed, and the area including the shoulder image (or the area including the shoulder image Processing area may be removed to generate a post-processing omnidirectional image. The removal target area may be an area corresponding to a range of view angles, and the removal target area may be removed on an omni-directional image.

11 is a conceptual diagram illustrating an image post-processing method for an object to be processed according to an embodiment of the present invention.

In Fig. 11, a determination method for an object to be processed is disclosed.

Referring to FIG. 11, the determination of the presence or absence of an object to be processed for each frame may be a burden on image post-processing. Therefore, based on the determination result of the object to be processed for the specific frame, further determination can be performed on the object to be processed using the determination result of the specific frame for the subsequent frame of the specific frame.

Specifically, it can be determined that the object to be processed exists in the first frame 1100 based on the above-described method. In the case where the object to be processed exists in the first frame 1100, there is a high possibility that the object to be processed continuously or repeatedly appears in the image frame 1150 after the first frame. Accordingly, the determination of the presence or absence of the object to be processed for the image frame 1150 after the first frame may be performed based on the object to be processed existing in the first frame 1100.

It is possible to determine whether or not an object having characteristic information similar to the characteristic information of the object to be processed exists in the imaging frame 1150 after the first frame by extracting characteristic information of the object to be processed existing in the first frame 1100 have. That is, the object image to be processed of the first frame 1100 is set as a reference, and thereafter, based on the similarity of the first frame 1100 to the object image of the processing object 1100 with respect to the image frame 1150 after the first frame It can be determined whether or not the object to be processed exists.

That is, for the first frame and after the first frame 1150, it is not determined whether or not the object to be processed exists for the individual frame. Instead, the first frame 1100 is traced for existence of the object to be processed, Lt; / RTI > is included in a subsequent frame may be determined. The tracking of the presence or absence of the object to be processed may be performed based on the motion vector of the object, or may be performed in the set determination region.

12 is a conceptual diagram illustrating an operation of an omni-directional image processing apparatus according to an embodiment of the present invention.

12, the operation of the omnidirectional image processing apparatus is started.

Referring to FIG. 12, the omnidirectional image processing apparatus 1200 may include a plurality of image pickup units 1250 as described above.

The lens unit of each of the plurality of image sensing units 1250 can be physically driven based on hardware, so that the angle of view (or the angle of the sensing line) in a specific direction can be adjusted or the angle of view ) May be adjustable.

According to the embodiment of the present invention, the omnidirectional image processing apparatus 1200 may receive feedback on determination result information about an object to be processed. The omnidirectional image processing apparatus 1200 determines information (hereinafter, drive control information) about an angle of view (or an angle of an imaging line) in a specific direction to be changed based on the determination result information about the object to be processed (1250).

(Hereinafter, drive control information) for the angle of view (or the angle of the imaging line) in a specific direction to be changed based on the determination result information on the object to be processed may be determined, and the omni- It is possible to control the driving of each of the plurality of image pickup sections 1250 based on the drive control information.

The position of the object to be processed is determined in the omnidirectional image, and the drive control information can be determined based on the position of the object to be processed. The drive control information may include information on an angle of view for imaging except for the object to be processed in the omnidirectional image.

A first angle of view for imaging the object to be processed is determined in the omnidirectional image and information about the first angle of view can be transmitted to the omnidirectional image processing apparatus 1200 as the primary drive control information.

The image pickup unit 1250 of the omnidirectional image processing apparatus 1200 can change the angle of view based on the primary drive control information. It can be additionally determined whether the omnidirectional image captured according to the changed angle of view based on the primary drive control information includes the object to be processed. When the omnidirectional image includes the object to be processed, the secondary drive control information for additional drive control is generated and transmitted to the omnidirectional image processing apparatus 1200.

That is, the image captured by the omnidirectional image processing apparatus 1200 may not include the object to be processed or the area where the object to be processed can be seen may be reduced as much as possible by adjusting the angle of view through the feedback method.

It is possible to restrict the range of the angle of view to be adjusted in performing the angle of view adjustment for the omnidirectional image processing apparatus 1200 so as to prevent the image from being reduced due to excessive angle of view adjustment. For example, when the range of the angle of view limited or adjusted is large, imaging may not be performed for a partial angle of view of the omnidirectional image. Therefore, the range of the adjustable angle of view can be set, and the angle of view can be set so as not to adjust the range of the adjustable angle of view.

The embodiments of the present invention described above can be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable recording medium may be those specifically designed and configured for the present invention or may be those known and used by those skilled in the computer software arts. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, medium, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code, such as those generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be modified into one or more software modules for performing the processing according to the present invention, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, Those skilled in the art will appreciate that various modifications and changes may be made thereto without departing from the scope of the present invention.

Accordingly, the spirit of the present invention should not be construed as being limited to the above-described embodiments, and all ranges that are equivalent to or equivalent to the claims of the present invention as well as the claims .

Claims (1)

A method of removing an object to be processed from an image includes:
The image processing apparatus determining the object to be processed in the image; And
And performing image post-processing on the object to be processed by the image processing apparatus,
Wherein the object to be processed is an object for which imaging is not intended.

Images