JP2007214887A - Digital still camera and image composition method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To compose a pan focused image without unnaturalness from a plurality of images, even when blurring occurs. <P>SOLUTION: This digital still camera, with which an image with the entire screen coming into focus is composited, by consecutively photographing a plurality of images different in the focal position of the same object, and segmenting and pasting together part areas coming into focus from the respective photographed images, is provided with a motion correcting means 3 for matching the viewing angles of the plurality of images with each other; an image compositing means 4 for generating a single composite image by segmenting and pasting together part areas having focus from the plurality of images corrected by the motion correcting means 3; and a boundary correction means 5 for modifying a boundary part between different images in the composited image. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a digital still camera and an image composition method, and more particularly, to a digital still camera and an image composition method capable of capturing and generating a focused image from a distant view to a close view.

  In a digital still camera that images a subject using a solid-state imaging device such as a CCD type image sensor or a CMOS type image sensor, for example, as described in Patent Documents 1 and 2 below, the same subject image with different focal lengths is recorded. A pan-focus image is obtained by capturing a plurality of images and cutting out and synthesizing in-focus portions from the respective images.

  However, in the prior art described in Patent Document 1, the incident light from the subject under the microscope is divided into three parts using a prism, and each of them is received by different solid-state image sensors placed at different focal lengths. The incident light quantity of each solid-state image sensor becomes 1/3. Further, when a color solid-state image sensor equipped with, for example, a primary color filter is used as the solid-state image sensor, for example, a pixel equipped with a red filter does not use green and blue light, so the light use efficiency becomes 1/3. In order to obtain a bright image, it is necessary to devise a considerably increased amount of exposure.

  In the prior art described in Patent Document 2, a plurality of the same subject images with different focal lengths are captured by a digital still camera that takes a general landscape photograph, and the focused portions of the images are bonded together to achieve pan-focus. I have an image. However, since the image is taken with the camera in hand, the shift of the image composition boundary portion is noticeable when the camera shake occurs because the image is in focus.

JP 2005-57503 A JP 2002-51254 A

  The user holds the camera, picks up multiple images of the same subject at different focal lengths, and synthesizes a single pan-focus image by cutting out and pasting the focused portion from each image. In this case, there is a problem in that an image becomes unnatural unless the image shift at the boundary between the bonded portions is corrected well.

  An object of the present invention is to provide a digital still camera and an image composition method capable of synthesizing a pan-focus image having no unnaturalness even when camera shake occurs.

  The digital still camera of the present invention continuously captures a plurality of images of the same subject with different focus positions, and cuts out and pastes a partial area that is in focus from each captured image so that one image is focused on the entire screen. In the digital still camera for synthesizing a plurality of images, motion correction means for matching the angles of view between the plurality of images, and a focused partial area from the plurality of images corrected by the motion correction means It is characterized by comprising image combining means for generating a single composite image by combining, and boundary correction means for correcting a boundary portion between the different images in the composite image by correction calculation.

  The boundary correction means of the digital still camera of the present invention calculates the image data of the boundary part by weighted addition averaging of the image data of the boundary part cut out from the two original images used for the synthesis. And

  The image synthesizing unit of the digital still camera of the present invention divides each of the plurality of different images into the same mesh, and most of the divided images at the same position of the different images among the divided images. It is characterized in that a divided image having a high degree of focus is selected and used as a divided image at the position in the composite image.

  The image synthesizing unit of the digital still camera according to the present invention is characterized in that, of the plurality of different images, an image in which the positional deviation correction by the motion correcting unit is not less than a predetermined value is not used for the image synthesizing.

  The image composition method according to the present invention is an image composition method in which a single focused image is synthesized on the entire screen by cutting out and pasting partial regions in focus from a plurality of images with different focus positions of the same subject. , Correcting the angle of view between the plurality of images, generating a single composite image by cutting out and pasting the focused partial area from the plurality of corrected images, and generating the composite image A boundary portion between the different images in the image is corrected by a correction calculation.

  In the image composition method of the present invention, the correction calculation is performed by weighting and averaging the image data of the boundary portion obtained by cutting out the image data of the boundary portion from the two original images used for the composition. To do.

  In the image composition method of the present invention, the image composition method divides each of the plurality of different images into the same mesh, and among the divided images, the most focused image among the divided images at the same position of the different images. A divided image having a high degree is selected as a divided image at the position in the composite image.

  In the image composition method of the present invention, an image in which positional deviation correction by correction for matching the angle of view among the plurality of different images is not used for the image composition.

  According to the present invention, since a plurality of images subjected to motion correction are used when synthesizing images, unnaturalness is eliminated in the synthesized image. In addition, since the boundary portion between different images is corrected, the connection between the different images becomes smooth, and a pan-focus image without a sense of incongruity can be synthesized.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a functional explanatory diagram of a digital still camera according to an embodiment of the present invention. In the digital still camera of this embodiment, when the user presses the release button, three color subject images (first image, second image, and third image) are continuously captured by the color solid-state imaging device 1.

  The first image is picked up when the focus position of the focus lens 2 is automatically adjusted to a distant view, the second image is picked up when the focus position of the focus lens 2 is automatically set to a middle position, and the third image is focused. An image is taken when the focus position of the lens 2 is automatically adjusted to the near view.

  The image data of the first image, the second image, and the third image output from the color solid-state imaging device 1 are stored in the frame memory. When an image is taken with a digital still camera in hand, a slight shift occurs between each of the first image, the second image, and the third image due to camera shake.

  Therefore, the minute motion correction means (camera shake correction means) 3 built in the camera generates an image in which the shift between the first image, the second image, and the third image is corrected. For example, when the first image, the second image, and the third image are captured in this order, the difference between the first image and the second image is detected, and the angle of view of the second image is set so that the difference disappears. The second image is corrected to match the angle of view. Similarly, the difference between the first image and the third image is detected, and the third image is corrected so that the angle of view of the third image matches the angle of view of the first image so that the difference is eliminated.

  This correction can be performed using a conventional electronic (soft) camera shake correction technique. By this correction, the etched portion in the second image matches the edge portion in the first image, and the edge portion in the third image matches the edge portion in the first image.

  At this time, for example, when the camera shake is so large that the positional deviation of the third image cannot be corrected, that is, when the positional deviation correction amount is larger than a predetermined value, the third image is deleted, A composite image described below is generated from the second image. When the positional deviation is small, an image composition process is performed using all images as described below.

  Note that when the first image → the second image → the third image is captured in this order, if the shift amount increases in this order, the positional shift amount of the third image with respect to the first image increases, and the third image is deleted. Become. However, the third image can be used by using the second image captured in the middle as a basic image and correcting the first image and the third image so as to match the angle of view of the second image. In some cases. It is also possible to obtain an image that becomes the center of blurring among the three images by arithmetic processing, and to use this as a basic image.

  Next, the image synthesizing means 4 built in the camera analyzes three images (the corrected first image, the second image (basic image), and the corrected third image), and cuts out and pastes the focused portion. Thus, one composite image is generated as will be described in detail later.

  Further, the boundary correction means 5 built in the camera will be described in detail later from the boundary portion images cut out from the corrected first image, the second image, and the corrected third image of the boundary portion in the composite image. And this is replaced with the image of the boundary portion in the composite image to obtain an output image.

  FIG. 2 is a detailed explanatory diagram of the composition processing performed by the image composition means 3 shown in FIG. A corrected first image (hereinafter referred to as A image), a second image (hereinafter referred to as B image), and a corrected third image (hereinafter referred to as C image) are respectively meshed. Divide into As the mesh is finer, a finely focused portion can be extracted. However, in the illustrated example, the mesh is divided into 4 × 4 = 16 meshes. The divided images in the i-th row and the j-th column are illustrated as Aij, Bij, and Cij.

  First, the image synthesizing means 3 detects a focus value (for example, a high frequency component value or a contrast value) of the divided image A11 in the first row and first column in the A image through a filter (not shown). The in-focus value of the divided image B11 in the first row and the first column and the in-focus value of the divided image C11 in the first row and the first column in the C image are detected.

  Then, a divided image having the highest in-focus value is selected from the divided images A11, B11, and C11, and set as the divided image in the first row and first column of the composite image. In the example shown in FIG. 2, the divided image B11 is selected as the divided image in the first row and first column of the composite image.

  Similarly, the divided image of the i-th row and j-th column constituting the composite image is selected from the divided images Aij, Bij, and Cij to generate the composite image.

  Even in this synthesized image, since the edge portion in each image is corrected by the minute motion correcting unit 3 in FIG. 1, the deviation between the subject images in the boundary portion is not conspicuous. However, since the A image (far view image), B image (middle view image), and C image (near view image) are originally captured images with different focus lens focus positions, in this embodiment, further boundary correction processing is performed.

  FIG. 3 is a detailed explanatory diagram of the boundary correction process performed by the boundary correction unit 5. The composite image described with reference to FIG. 2 is an image obtained by cutting and pasting an A image, a B image, and a C image in units of divided images, so that a boundary 7 exists between different images A, B, and C.

  Therefore, the boundary correction unit 5 determines a narrow partial region sandwiching the boundary 7. FIG. 4 illustrates a partial region 8 defined along the boundary 7 between the A image and the C image. The boundary correction means 5 of this embodiment divides the partial region 8 into a narrower region 8y sandwiching the boundary 7, a region 8x on the A image side adjacent to the region 8y, and a region 8z on the C image side adjacent to the region 8y. And divided into three.

  Then, the boundary correction unit 5 cuts out the image of the region 8y from the A image and also cuts out the image of the region 8y from the C image, and calculates an image (for example, 0.5A + 0.5C) obtained by weighted addition averaging of the cut out images. Thus, an image of the composite image area 8y is obtained.

  Similarly, the image of the region 8x is cut out from the A image and the image of the region 8x is cut out from the C image, and an image (for example, 0.75A + 0.25C) obtained by weighted addition averaging of both the cut out images is calculated. Let it be an image of the region 8x.

  Similarly, the image of the region 8z is cut out from the A image, the image of the region 8z is cut out from the C image, and an image (for example, 0.25A + 0.75C) obtained by weighted addition averaging of both the cut out images is calculated. Let it be an image of the region 8z.

  Then, the image data of the calculated regions 8x, 8y, and 8z is overwritten on the image data of the boundary portion of the composite image obtained in FIG. 2 to obtain an output image.

  Such boundary correction processing is performed at all the boundaries 7 between the images A, B, and C shown in the left diagram of FIG. 3, and the image data of the partial region 8 along the boundary 7 is replaced with image data obtained by weighted addition averaging. As a result, the connection between the images A, B, and C is further smoothed, and it is possible to obtain a natural composite image with pan focus and no sense of incongruity.

  In the above-described embodiment, three images having different focus positions are continuously captured and combined. However, a large number of images having different focus positions, such as four, five,... The present embodiment can also be applied to image composition.

  The image composition method of the present invention is useful as a digital still camera that captures pan-focus images because the images are smoothly connected and synthesized without unnaturalness.

FIG. 3 is a functional explanatory diagram of a digital still camera according to an embodiment of the present invention. It is function explanatory drawing of the image synthetic | combination means shown in FIG. It is a function explanatory view of the boundary correction means shown in FIG. It is explanatory drawing of the weighted addition average process which the boundary correction means shown in FIG. 1 performs.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Color solid-state image sensor 2 Focus lens 3 Minute motion correction means (camera shake correction means)
4 Image composition means 5 Boundary correction means 7 Boundary between different images 8 Partial region sandwiching the boundary

Claims (8)

  A digital still camera that synthesizes a single focused image on the entire screen by continuously capturing multiple images of the same subject with different focus positions and cutting out and pasting the in-focus partial areas from each captured image In this case, a motion correction unit that matches the angle of view between the plurality of images, and a combined image obtained by cutting out and pasting the focused partial area from the plurality of images corrected by the motion correction unit A digital still camera comprising: an image synthesizing unit that generates a boundary, and a boundary correcting unit that corrects a boundary portion between the different images in the synthesized image by a correction operation.   The said boundary correction means calculates the image data of the said boundary part by carrying out the weighted addition average of the image data of the said boundary part cut out from the two original images used for the said composition. Digital still camera.   The image synthesizing unit divides each of the plurality of different images into the same mesh, and among the divided images, the divided image having the highest in-focus degree among the divided images at the same position of the different images. 3. The digital still camera according to claim 1, wherein the digital still camera is selected to be a divided image at the position in the composite image.   4. The image synthesizing unit according to claim 1, wherein the image synthesizing unit does not use, in the image synthesizing, an image in which positional deviation correction by the motion correcting unit is a predetermined value or more among the plurality of different images. The digital still camera described in Crab.   In an image composition method for synthesizing a single focused image on the entire screen by cutting out and pasting partial areas in focus from a plurality of images with different focus positions of the same subject, A correction that matches the angle of view is performed, and a single partial image is generated by cutting out and pasting a partial area that is in focus from the plurality of corrected images, and between the different images in the composite image An image composition method, wherein a boundary portion is corrected by a correction operation.   6. The correction operation according to claim 5, wherein the correction calculation is performed by performing weighted addition averaging of the image data of the boundary portion obtained by cutting out the image data of the boundary portion extracted from the two original images used for the synthesis. Image composition method.   In the image synthesis, each of the plurality of different images is divided into the same mesh, and among the divided divided images, a divided image having the highest degree of focus is selected from the divided images at the same position of the different images. The image composition method according to claim 5, wherein the image is a divided image at the position in the composite image.   8. The image according to claim 5, wherein an image in which positional deviation correction by correction for matching the angle of view is not more than a predetermined value among the plurality of different images is not used for the image synthesis. Image composition method.

Images