JP2016031548A - Image processing apparatus, imaging device, program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to reduce storage capacity and time and effort for a user.SOLUTION: An information processing part (300) of an imaging device (10) includes a specification part (307) and an extraction part (308). The specification part (307) specifies, in a photographed image (500), pixels determined not to be background color pixels and determined to be edge pixels. The extraction part (308) refers to the pixels specified by the specification part (307) to extract a photographic object area (510) from the photographed image (500).SELECTED DRAWING: Figure 1

Description

  The present invention relates to an image processing apparatus that processes image data generated by photographing.

  In Patent Document 1, when a pedestal such as a desk is used as a background and a sheet of paper placed on the pedestal is photographed as a photographing target, a photographing target region (a region indicating the photographing target) is extracted from the photographed image ( Technology) is disclosed.

JP 2007-201948 A JP2013-168119A JP 2013-4088 A

  In the technique of Patent Document 1, a pedestal in a state where there is no photographing target (paper or the like) is photographed in advance, and then a photographing target placed on the pedestal is photographed. Subsequently, a difference image between a photographed image obtained by photographing a pedestal having no photographing object and a photographed image obtained by photographing the photographing object placed on the pedestal is obtained, and contour information (photographing is performed from the difference image. The contour of the target area) is extracted.

  Therefore, the technique of Patent Document 1 has a problem in that it takes time for the user to photograph the pedestal before placing the photographing object before photographing the photographing object. Moreover, when re-photographing a photographing target by changing the photographing position and layout, it is necessary to take a picture of the pedestal before placing the photographing target again, which is troublesome.

  Further, in the technique of Patent Document 1, a storage capacity is required to hold two images, a captured image showing a pedestal with no shooting target and a captured image showing a shooting target placed on the pedestal. The problem also arises.

  SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing apparatus that can suppress the storage capacity and the user's effort compared to the prior art.

  An image processing apparatus according to an aspect of the present invention is formed by connecting, from a photographed image, a detection unit that detects candidate vertices that are candidates for vertices of a photographing target region, and connecting a plurality of candidate vertices of the photographed image with lines. A setting unit that sets a gradation value condition that is the same as or similar to the representative color of the background area, and a background color pixel that satisfies the condition for each pixel of the photographed image. A first determination unit configured to determine whether or not, a second determination unit configured to determine whether or not each pixel is an edge pixel indicating an edge, and the edge determined to be not the background color pixel in the captured image A specifying unit that specifies a pixel determined to be a pixel, and an extraction unit that extracts the shooting target region from the captured image with reference to the pixel specified by the specifying unit.

  According to one aspect of the present invention, since a portion to be photographed can be extracted with high accuracy without requiring an image in which a background (a pedestal such as a desk) without a subject to be photographed is used as in Patent Document 1, photographing is performed. There is no need to perform the process twice, so that the storage capacity and the labor required for the user can be reduced as compared with Patent Document 1.

1 is a block diagram illustrating a schematic configuration of a photographing apparatus according to an embodiment of the present invention. It is the figure which showed typically the picked-up image image | photographed with the imaging device of FIG. It is a flowchart which shows the processing content of a four corner detection part. It is the figure which showed an example of the moving average filter. It is a figure which shows eight sub area | regions obtained by dividing | segmenting the background area | region contained in a picked-up image. 6 is a table showing the statistic of the gradation value of the R channel in the reference range for each of the sub-regions in FIG. 5. FIG. 6 is a table showing R channel tone value statistics for each of the sub-regions in FIG. 5, after rearrangement so that the average value of the tone values is in ascending order. The background color image produced | generated for every channel of R, G, B is shown. It is the figure which showed an example of the Gaussian filter. It is the figure which showed an example of the Sobel filter. It is the flowchart which showed the content of the process (canny method) by an edge detection part. The edge image produced | generated for every channel of R, G, B is shown. It is a figure which shows the non-background color / edge image which shows the pixel determined not to be a background color pixel, and was determined to be an edge pixel. It is the schematic diagram which showed one of the pixel columns searched about each of a 1st search process-a 4th search process. It is the figure which showed the mask image produced | generated in the extraction part. It is the figure which showed the extracted image produced | generated in the extraction part. It is a figure which shows four sub area | regions used as the object of a statistical process among the background area | regions contained in a picked-up image. It is a figure which shows seven sub area | regions used as the object of a statistical process among the background area | regions contained in a picked-up image.

Embodiment 1
(Overall configuration of the imaging device)
Hereinafter, the configuration of an imaging apparatus according to an embodiment of the present invention will be described in detail. FIG. 1 is a block diagram illustrating a schematic configuration of a photographing apparatus according to the present embodiment.

  The imaging device 10 of the present embodiment is a portable terminal device having a camera function for capturing an image, and includes, for example, a tablet. As illustrated in FIG. 1, the imaging device 10 includes at least a touch panel unit 100, an imaging unit 200, an information processing unit 300, and a storage unit 400.

  Furthermore, although not shown, the photographing apparatus 10 is normally provided in a portable terminal device such as a communication unit, an Internet communication unit, a voice operation unit, a voice output unit, a television broadcast receiver, and a GPS (Global Positioning System). It has various hardware.

  The touch panel unit 100 is an apparatus in which a display unit that displays an image and an input unit that generates an input signal corresponding to a user's touch operation are integrated, and a known touch panel can be used. That is, the display unit is configured by an LCD (Liquid Crystal Display) or the like, and the input unit is configured by a capacitance sensor integrated with the display unit. In addition, although the imaging device 10 of the present embodiment includes the touch panel unit 100 as an image display unit, the image display unit is not limited to the touch panel unit 100. That is, in the photographing apparatus 10, the display unit and the input unit may be configured separately. In this case, the display unit is an LCD or the like, and the input unit is a keyboard or the like.

  The photographing unit 200 is an image sensor (camera) such as a charge coupled device (CCD) or a complementary metal-oxide semiconductor (CMOS). The photographing unit 200 performs photographing processing according to the photographing instruction signal input by the user, and generates photographed image data indicating the photographed image. The photographed image data is color image data indicating the gradation values (0 to 255 in the case of 8 bits) of each channel of R (red), G (green), and B (blue) for each pixel. .

  The information processing unit (image processing apparatus) 300 is a block that controls the operation of each hardware of the photographing apparatus 10, and is realized by, for example, a microcomputer or a microprocessor including a CPU (Central Processing Unit). The information processing unit 300 extracts various information stored in the storage unit 400 and a program for performing various controls, performs information processing, and controls each hardware of the photographing apparatus 10 based on the result of the information processing. And the information are sent to control the operation of each hardware of the photographing apparatus 10.

  Further, the information processing unit 300 is designed to shift to a document shooting mode in response to an input instruction from a user. The document photographing mode is a mode in which when photographing is performed by the photographing unit 200, document image processing is performed on photographed image data obtained by photographing. Document image processing refers to image processing based on the premise that a document (book, document, etc.) placed on a desk is a subject to be photographed.

  That is, in the case where the information processing unit 300 has shifted to the document photographing mode and the photographed image data is generated by photographing by the photographing unit 200, the photographed image data is subjected to document image processing on the photographed image data. Is stored in the storage unit 400. The document image processing performed in the information processing unit 300 will be described in detail later.

  The storage unit 400 stores various application programs, OS programs, control programs, image processing programs executed by the information processing unit 300, various data (setting values, tables, etc.) read when executing these programs, and the captured image data. This is a storage area for storing. The storage unit 400 includes various conventionally known storage means such as read only memory (ROM), random access memory (RAM), flash memory, EPROM (Erasable Programmable ROM), EEPROM (registered trademark) (Electrically EPROM). A hard disk drive (HDD) or the like can be used. In addition, various data handled by the information processing unit 300 and data being processed are temporarily stored in the working memory of the storage unit 400.

(Information Processing Unit 300)
Next, each block included in the information processing unit 300 will be described in detail. As shown in FIG. 1, the information processing unit 300 includes a four-corner detection unit 301, a smoothing processing unit 302, a statistical processing unit 303, a condition setting unit 304 (setting unit), a background color determination unit 305 (first determination unit), An edge detection unit 306 (second determination unit), an identification unit 307, an extraction unit 308, a distortion correction unit 309, and a format processing unit 310 are provided. Each of these units is a processing unit that performs the document image processing.

  That is, when the document shooting mode is set, when shooting is performed by the shooting unit 200, four corner detection unit 301, smoothing processing unit 302, statistical processing unit are performed on the shot image data obtained by this shooting. 303, the condition setting unit 304, the background color determination unit 305, the edge detection unit 306, the specifying unit 307, the extraction unit 308, the distortion correction unit 309, and the format processing unit 310 execute image processing in this order.

  Note that a shooting target area (area indicating a shooting target) included in the shot image is extracted (specified) by the processes from the four corner detection units 301 to the extraction unit 308, and the distortion correction unit 309 performs processing on the shooting target area. On the other hand, distortion correction is executed, and an image file indicating the shooting target area is formed by the processing of the format processing unit 310. Hereinafter, each block included in the information processing unit 300 will be described in order.

(Four corner detector 301)
The photographed image according to the present embodiment is an image obtained by photographing a document on a two-page spread of a book placed on a desk as a subject of photographing and including the desk as a background. In other words, as shown in FIG. 2, the photographed image 500 shows a photographing target area (book) 510 and a background area (desk) 550 that are image areas indicating the photographing target.

  The four-corner detection unit 301 is a block that detects candidate vertices that are candidates for vertices of the shooting target area 510 by regarding that the shooting target area 510 included in the shot image 500 is a substantially rectangular document. Since a generally rectangular shape usually has four vertices, four candidate vertices are detected.

  There are various known methods for detecting the four candidate vertices of the substantially rectangular imaging target area 510, and the four-corner detection unit 301 may use any known method, but in the present embodiment, The technique shown in paragraphs [0056] to [0106] of Patent Document 2 (Japanese Patent Laid-Open No. 2013-168119) is used. Below, the outline | summary of the processing content of this method is demonstrated.

  FIG. 3 is a flowchart showing processing of the four corner detection unit 301. First, the four-corner detection unit 301 extracts edge pixels constituting an edge from the captured image 500, and generates an edge image (binary image) in which the edge pixel is “1” and the non-edge pixel is “0”. Generate (S1). There are various known methods for extracting edge pixels, and the four-corner detection unit 301 may use any known method. However, in this embodiment, the captured image is converted into luminance data, and edge extraction is performed on the luminance data by the Canny method.

  In the present embodiment, the canny method is also used in the edge detection unit 306, and details of edge extraction using the canny method will be described later together with the edge detection unit 306.

  Note that the four-corner detection unit 301 may perform morphological transformation such as expansion and contraction on the edge image generated by the Canny method.

  Next, the four corner detection unit 301 performs a labeling process in which a group of edge pixels adjacent (connected) in four directions or eight directions is set as a connected edge region, and a different label is attached to each connected edge region (S2).

  Subsequently, the four-corner detection unit 301 performs a process of extracting a connected edge region including a boundary between the imaging target region 510 and the background region 550 as a feature region from each of the labeled connected edge regions (S3). The subject is usually photographed so that its center is near the center of the photographed image and occupies most of the photographed image. Therefore, the four corner detection unit 301 extracts a connected edge region that satisfies the following condition A as a feature region.

  Condition A: In the captured image, the upper left corner is the origin, the right direction (width direction) is the x axis, the lower direction (height direction) is the y axis, the x coordinate of the right end of the captured image is Xmax, and the lower end of the captured image is The y coordinate is Ymax. At this time, the length in the width direction of the connected edge region is ¼ or more of the width (Xmax) of the captured image, and the length in the height direction of the connected edge region is 1 of the height of the captured image (that is, Ymax). / 4 or more, and the x coordinate of the center point of the connected edge region is Xmax / 4 or more and 3 × Xmax / 4 or less, and the y coordinate of the center point is Ymax / 4 or more and 3 × Ymax / 4. It is as follows.

  An example of a method for obtaining the coordinates of the center point of the connected edge region is a method in which the average of the coordinate values of all the pixels in the connected edge region is used as the coordinate of the center point.

  Subsequently, the four-corner detection unit 301 performs processing (straight line extraction processing) for identifying straight lines that form the upper side, the left side, the right side, and the lower side of the substantially rectangular region formed from the four candidate vertices from the feature region (S4). ).

  Here, there is a high probability that the upper side is located in the upper half of the captured image (that is, the y coordinate is in the range of 0 to Ymax / 2) and is parallel to the width direction of the captured image. Further, the left side is located in the left half of the captured image (that is, the x coordinate is in the range of 0 to Xmax / 2) and has a high probability of being parallel to the height direction of the captured image. The right side is located in the right half of the captured image (that is, the x-coordinate is in the range of Xmax / 2 to Xmax) and has a high probability of being parallel to the height direction of the captured image. The lower side is located in the lower half of the captured image (that is, the y coordinate is in the range of Ymax / 2 to Ymax) and has a high probability of being parallel to the width direction of the captured image.

  Therefore, for each of the upper side, the left side, the right side, and the lower side, the number of edge pixels connected in a specific direction is the maximum in the range where the existence probability is high, and a line-shaped edge pixel group having a predetermined length or more is selected as the feature region. Extract from inside. The expression of the approximate straight line of the extracted edge pixel group is specified by the method of least squares. The specified approximate straight line becomes the upper side, the left side, the right side, and the lower side.

  When the four-side straight line extraction process in S4 is completed, the four-corner detection unit 301 obtains the intersection coordinates based on the straight line formula obtained in S4 (S5). That is, the four corner detection unit 301 sets the intersection coordinates of the left-side straight line and the upper-side straight line as the upper-left vertex coordinates, the intersection coordinates of the upper-side straight line and the right-side straight line as the upper-right vertex coordinates, and the intersection coordinates of the right-side straight line and the lower-side straight line as the lower-right vertex coordinates. And the coordinates of the intersection of the straight line on the left side are obtained as the coordinates of the lower left vertex. Then, the four corner detection unit 301 uses these four vertex coordinates as the coordinates of the candidate vertices, and stores the extraction result information indicating the vertex coordinates in the storage unit 400.

  In the above, the outline of the process of each step of FIG. 3 has been described. The specific contents of the process of each step of FIG. 3 are as shown in Patent Document 2. Further, the four corner detection unit 301 may obtain the coordinates of the four candidate vertices using a known method other than the method disclosed in Patent Document 2.

(Smoothing processing unit 302)
Next, the smoothing processing unit 302 will be described. In this embodiment, a statistical processing unit 303 (to be described later) obtains the statistic of the gradation value of the background area 550 included in the captured image 500. This statistic is the mean as well as the standard deviation. Here, if noise depending on the characteristics of the image sensor is included in the captured image data (input image data), the value of the standard deviation may greatly fluctuate due to the noise. Examples of this noise include dark part noise, thermal current, and noise due to output variations of the image sensor. Due to the influence of such noise, a plurality of different portions in the background area should originally exhibit substantially the same gradation value, but show significantly different gradation values. There is a possibility that the error of the calculated statistical value becomes large.

  Therefore, the smoothing processing unit 302 located in the preceding stage of the statistical processing unit 303 performs the smoothing processing for each of the R, G, and B channels in order to reduce or eliminate the influence of the noise as described above. It has become.

  As means for performing a smoothing process, a spatial filter that performs a convolution operation such as a moving average filter or a weighted average filter is generally known. An example of the moving average filter is shown in FIG. By applying the same filter, the gradation value (pixel value) is averaged between the target pixel of the captured image and the surrounding pixels, so the fluctuation of the gradation value becomes gradual and smoothing processing is performed. Will be made.

  Note that a median filter may be used instead of the smoothing processing by the smoothing processing unit 302.

(Statistical processing unit 303)
The statistical processing unit 303 is a block that performs statistical processing of gradation values for the background region 550 included in the captured image 500. The processing contents of the statistical processing unit 303 will be specifically described as follows.

  The statistical processing unit 303 refers to the extraction result information stored in the storage unit 400 (information indicating the processing result of the four-corner detection unit 301), and the statistical processing unit 303 of the captured image smoothed by the smoothing processing unit 302 The background area 550 is specified from among them. Further, the statistical processing unit 303 divides the background region 550 into a plurality of sub-regions, and sets a reference range of 64 × 64 pixels centering on the central pixel of each region for each sub-region. The statistical processing unit 303 obtains the statistical value of the gradation value for each channel (each channel of R, G, B) in the reference range for each sub-region. The obtained statistical values are an average value of gradation values (gradation average value) of each pixel in the reference range and a standard deviation of gradation values of each pixel in the reference range.

  More specifically, the statistical processing unit 303 includes a rectangular area (broken line in FIG. 5) formed by connecting four candidate vertices detected by the four corner detection unit 301 with a straight line (a broken line in FIG. 5) and the captured image 500. An area between each side is treated as a background area 550. Further, the statistical processing unit 303 divides the background region 550 into eight sub-regions partitioned by each extension line of each side of the rectangular region (broken line in FIG. 5), and for each sub-region, the statistics of the reference range are divided. The value is calculated.

  That is, as shown in FIG. 5, the background area 550 is divided into sub-areas A to H. Then, for each sub-region, the gradation average value (average gradation value) and standard deviation for each of the R, G, and B channels are obtained for the reference range. FIG. 6 shows the grayscale average value and standard deviation of the R channel for each of the sub-regions A to H in FIG. Although not shown, for the G channel and the B channel, as in FIG. 6, the gradation average value and the standard deviation in each of the sub-regions A to H in FIG. 5 are obtained.

(Condition setting unit 304)
The condition setting unit 304 is a block for setting a numerical range (condition) of gradation values of the same or similar color as the representative color of the background region 550 based on the result of the statistical processing of the statistical processing unit 303. This numerical range is a range of (average value of gradation average value) ± (n × average value of standard deviation) in each of the RGB channels (n is a predetermined coefficient).

  For example, when the grayscale average values of the R channel in FIG. 6 are rearranged in ascending order, the result is as shown in FIG. 7, and the grayscales of the six sub-regions excluding the region where the grayscale average values are the maximum value and the minimum value are obtained. From the average value and the standard deviation, the average value of the R channel gradation average value (Rgra.avg.) And the average value of the standard deviation (Rσavg.) Are calculated.

That is,
Rgra.avg. = ((56.15 + 57.71 + 59.14 + 61.27 + 68.53 + 76.34)) / 6 = 63.19
It becomes.
The average value Rσavg. Of the R channel standard deviation is
Rσavg. = ((1.41 + 1.09 + 1.19 + 0.98 + 1.22 + 1.21)) / 6 = 1.18
The numerical range is set based on these values.

Specifically, for the R channel, the numerical value range Rcand of gradation values of the same or similar color as the representative color is:
(Rgra.avg.−n × Rσ) ≦ Rcand ≦ (Rgra.avg. + N × Rσ)
And
If the predetermined coefficient n is set to 3,
59.6 ≦ Rcand ≦ 66.7
It becomes.

The numerical range Rcand is
(Rgra.avg.−n × Rσ) ≦ Rcand ≦ (Rgra.avg. + N × Rσ)
Of course, (Rgra.avg.−n × Rσ) <Rcand <(Rgra.avg. + N × Rσ) may be satisfied.

  The above processing is similarly performed for the G and B channels, and Gcand and Bcand are obtained in addition to Rcand.

  Note that, in the statistical processing unit 303 and the condition setting unit 304, the average value may be obtained not only by arithmetic average but also by weighted average or arithmetic average. In the present embodiment, the coefficient n is set to 3. However, the coefficient n is not particularly limited, and may be a decimal number as well as an integer. In the statistical processing unit 303 and the condition setting unit 304, a median value may be used instead of the average value.

(Background color determination unit 305)
The background color determination unit 305 determines whether each pixel of the captured image 500 is a background color pixel that satisfies the numerical range set by the condition setting unit 304 for each of the R, G, and B channels. That is, in the R channel, the background color determination unit 305 determines that a pixel whose gradation value is Rcand is a background color pixel, and determines that other pixels are not background color pixels. In the G channel, the background color determination unit 305 determines a pixel whose gradation value is Gcand as a background color pixel, and determines that other pixels are not background color pixels. In the B channel, the background color determination unit 305 determines a pixel whose gradation value is Bcand as a background color pixel, and determines that other pixels are not background color pixels.

  Further, the background color determination unit 305 generates a background color image indicating the determination result of whether or not it is a background color pixel for each of the R, G, and B channels. The background color image is a binary image in which the background color pixel is “1” and the non-background color pixel is “0”. FIG. 8 shows a background color image generated for each of the R, G, and B channels.

  In the background region 550 included in the photographed image 500, white noise or the like generated due to the characteristics of the image sensor appears as an isolated point, and this isolated point may not be determined as a background color pixel.

  Therefore, the background color determination unit 305 may convert the isolated points into background color pixels by performing morphological conversion such as expansion and contraction on the background color image. As a result, a background color image from which isolated points are removed can be generated.

(Edge detection unit 306)
The edge detection unit 306 performs edge pixel detection processing (extraction processing) from the captured image 500 for each of the R, G, and B channels. Specifically, for each of R, G, and B channels, it is determined whether or not each pixel constituting the captured image 500 is an edge pixel (a pixel that constitutes an edge).

  As edge pixel detection processing by the edge detection unit 306, various well-known edge detection methods (edge extraction methods) can be used. In this embodiment, a canny method (canny filter) is used. The Canny method refers to a method of detecting a thinned edge using a Gaussian filter (FIG. 9) and a Sobel filter (FIG. 10). Hereinafter, processing (canny method) by the edge detection unit 306 will be described.

  FIG. 11 is a flowchart showing the processing contents of the edge detection unit 306. The edge detection unit 306 executes the processing of FIG. 11 for each of R, G, and B channels.

  First, the edge detection unit 306 smoothes the captured image using a Gaussian filter (S101). Thereby, the noise of a picked-up image is suppressed.

Next, the edge detection unit 306 obtains edge strength and edge direction for each pixel using a Sobel filter (S201). Specifically, the edge detection unit 306 obtains the edge intensity ▽ L and the edge direction θ of the target pixel (X, Y) by the following equations (1) and (2).
▽ L = √ (Lx ² + Ly ² ) Formula (1)
θ = tan ⁻¹ (Ly / Lx) Equation (2)
Lx is a horizontal differential value of the Sobel filter, and Ly is a vertical differential value of the Sobel filter.

  Subsequently, the edge detection unit 306 quantizes θ (edge direction) obtained in S201 (S301). Specifically, quantization is performed to any angle of 0 °, 45 °, 90 °, and 135 °.

Next, the edge detection unit 306 performs a thinning process (S401) for thinning the edge with reference to the edge intensity ▽ L obtained in S201 and the θ quantized in S301. The thinning process is executed as shown below.
The edge detection unit 306 selects ▽ L (X, Y)> ▽ L (X-1, Y) and ▽ L (X, Y)> ▽ L (X + 1) for the target pixel (X, Y) at θ = 0 °. , Y), the value of ▽ L (X, Y) is maintained (does not change), and if this condition is not satisfied, the value of ▽ L (X, Y) is changed to 0.
The edge detection unit 306 selects ▽ L (X, Y)> ▽ L (X-1, Y + 1) and ▽ L (X, Y)> ▽ L (X + 1) for the pixel of interest (X, Y) at θ = 45 °. , Y-1), the value of ▽ L (X, Y) is maintained (does not change), and if this condition is not satisfied, the value of ▽ L (X, Y) is changed to 0.
The edge detection unit 306 selects ▽ L (X, Y)> ▽ L (X, Y-1) and ▽ L (X, Y)> ▽ L (X) for the target pixel (X, Y) at θ = 90 °. , Y + 1), the value of 変 更 L (X, Y) is maintained (does not change), and if this condition is not satisfied, the value of ▽ L (X, Y) is changed to 0.
The edge detection unit 306 selects ▽ L (X, Y)> ▽ L (X-1, Y-1) and ▽ L (X, Y)> ▽ L for the pixel of interest (X, Y) at θ = 135 °. When the condition of (X + 1, Y + 1) is satisfied, the value of ▽ L (X, Y) is maintained (not changed), and when the condition is not satisfied, the value of ▽ L (X, Y) is changed to 0.

  As described above, the thinning process of S3 can be executed by using ▽ L and θ. Next, the edge detection unit 306 determines whether or not it is an edge by performing hysteresis threshold processing on the edge intensity for each pixel (S501).

  Specifically, the edge detection unit 306 determines that the pixel of interest (X, Y) is an edge pixel if ▽ L (X, Y)> high level threshold, and ▽ L (X, Y) <low level. If it is a threshold value, the pixel of interest (X, Y) is determined as a non-edge pixel. When the high level threshold value> the low level threshold value and the image data is 8 bits, for example, the high level threshold value is 144 and the low level threshold value is set to 64. However, it goes without saying that the value is not limited to 144 or 64 as long as the high level threshold value> the low level threshold value holds.

  Further, the edge detection unit 306 determines that the target pixel (X, Y) is connected (adjacent) to the edge pixel when the low level threshold ≦≦ L (X, Y) ≦ high level threshold. Y) is determined as an edge pixel, and if not connected to the edge pixel, the target pixel (X, Y) is determined as a non-edge pixel.

  Subsequently, the edge detection unit 306 generates an edge image indicating a determination result of whether or not the pixel is an edge pixel (S601). The edge image is a binary image in which the edge pixel is “1” and the non-edge pixel is “0”. Note that the edge detection unit 306 may perform morphological conversion such as expansion and contraction on the edge image.

  The edge detection unit 306 executes the above-described processing of FIG. 11 for each of R, G, and B channels. As a result, an edge image is generated for each of the R, G, and B channels. FIG. 12 shows an edge image generated for each of R, G, and B channels.

  Note that since the Canny method is applied to edge extraction (S1 in FIG. 3) by the four corner detection unit 301, the same processing as the processing in FIG. 11 is performed. However, while the edge detection unit 306 performs the Canny method shown in FIG. 11 for each of the R, G, and B channels, the four-corner detection unit 301 executes the Canny method on the luminance data.

(Specific part 307)
The specifying unit 307 refers to the background color image generated by the background color determination unit 305 and the edge image generated by the edge detection unit 306, and determines the background color pixels in all the R, G, and B channels. A process of specifying a pixel that is determined to be an edge pixel and determined not to be performed is performed.

  That is, the specifying unit 307 performs processing for specifying pixels that are not background color pixels in the background color image of FIG. 8 but are edge pixels in the edge image of FIG. 12 in all channels R, G, and B. . In the background color image of FIG. 8, a pixel with a 0 flag (a flag indicating black) is equivalent to a pixel that is not a background color pixel. In the edge image of FIG. ) Stands for the edge pixel.

  Subsequently, the specifying unit 307 generates a non-background color / edge image that specifies pixels that are determined not to be background color pixels and determined to be edge pixels in all of the R, G, and B channels. FIG. 13 shows a non-background color / edge image 900. The non-background color / edge image 900 in FIG. 13 is set to 1 (white) as the flag of the pixels that are determined not to be background color pixels in all the R, G, and B channels and determined to be edge pixels. This is a binary image with the flag set to 0 (black).

(Extractor 308)
The extraction unit 308 performs processing for extracting the imaging target area 510 from the captured image 500 (FIG. 2) input from the imaging unit 200. Hereinafter, the processing of the extraction unit 308 will be described.

  The captured image 500 (FIG. 2) of the present embodiment includes a background area 550 indicating a desk or the like, and an imaging target area 510 indicating a document or the like. Therefore, in the processing by the edge detection unit 306, the outline of the imaging target area 510 is detected as an edge. However, when a scratch or the like is formed on the desk, the scratch or the like is also detected as an edge in the background area 550. It may be detected.

  However, although the edge of the background area 550 (such as a scratch on the desk) may be detected as an edge as described above, it does not change to indicate the background color, so it can be determined as a background color pixel. High nature. Therefore, the edge (scratches on the desk, etc.) of the background region 550 is determined from the specific pixel (pixel determined not to be the background color pixel and determined to be the edge pixel) specified in the non-background color / edge image of FIG. Will be removed.

  From the above points, in the non-background color / edge image 900 of FIG. 13, when searching for a specific pixel from the upper side in a direction orthogonal to the upper side, the specific pixel located closest to the upper side, and the lower side When searching for a specific pixel from the lower side toward the direction orthogonal to the specific pixel located closest to the lower side, and when searching for a specific pixel from the left side toward the direction orthogonal to the left side, closest to the left side When searching for a specific pixel from the right side in a direction orthogonal to the right side, a pixel group including the specific pixel closest to the right side is a contour line of the shooting target area 510 of the shot image 500. It seems to be equivalent.

  Therefore, the extraction unit 308 refers to the non-background color / edge image 900 in FIG. 13 and extracts the contour pixels that form the contour line of the imaging target region 510. Specifically, the extraction unit 308 (a) for each pixel that configures the upper side of the captured image 500, the extraction unit 308 is located at the position closest to the pixel from among the pixel columns that are aligned in the direction orthogonal to the upper side from the pixel. A first search process for searching for a specific pixel; and (b) for each pixel constituting the lower side of the captured image 500, the pixel from the pixel row arranged in a direction orthogonal to the lower side from the pixel is the most. A second search process for searching for a specific pixel at a close position; and (c) for each pixel constituting the left side of the photographed image 500, the pixel row is aligned from the pixel row arranged in a direction orthogonal to the left side. A third search process for searching for a specific pixel at a position closest to the pixel; and (d) for each pixel constituting the right side of the photographed image 500, a pixel row arranged in a direction orthogonal to the right side from the pixel. The pixel is the most Performing a fourth search processing for searching for a specific pixel at the position are. FIG. 14 is a schematic diagram illustrating one of the pixel columns to be searched for each of the first search process to the fourth search process. The pixels extracted in the first search process to the fourth search process are set as contour pixels.

  Further, the extraction unit 308 uses, as a mask image, a binary image in which the flags of the contour pixels and the pixels surrounded by the contour pixels are set to 0 (black) and the flags of the other pixels are set to 1 (white). Generate. FIG. 15 is a diagram showing a mask image 700 generated by the extraction unit 308.

  Subsequently, the extraction unit 308 extracts a photographing target area 510 from the photographed image 500 by performing a masking process on the photographed image 500 using the generated mask image 700. Specifically, the extraction unit 308 performs an imaging unit for the contour pixels and pixels surrounded by the contour pixels (pixels indicated by a 0 (black) flag) in the mask image 700 of FIG. A gradation value of the photographed image 500 input from 200 is given, and a pixel (a pixel indicated by a 1 (white) flag) located outside the contour pixel in the mask image of FIG. An extracted image 800 having a tone value (for example, “0” when the image data is represented by 8 bits) is generated. FIG. 16 is a diagram showing an extracted image 800. The extracted image 800 in FIG. 16 shows a shooting target area 510 extracted from the shooting image 500 in FIG. 2, and the periphery of the shooting target area 510 is a solid image area in which black is uniformly filled. That is, the extracted image 800 in FIG. 16 corresponds to an image in which a portion other than the shooting target area 510 in the captured image 500 in FIG. 2 is masked with a black solid area.

(Distortion correction unit 309)
The distortion correction unit 309 is inclined with respect to the normal direction of the front of the shooting target with respect to a rectangular shooting target area such as a document or a slightly curved shooting target area such as a spread page of a book. It is a block that corrects the distortion of the object to be photographed by photographing the object to be photographed from the direction and also corrects the inclination of the area to be photographed. As a correction method of the distortion correction unit 309, various known methods can be used. As an example, a method disclosed in Patent Document 3 (Japanese Patent Laid-Open No. 2013-4088) can be given.

  The distortion correction unit 309 performs distortion correction on the imaging target region indicated in the extracted image 800 (FIG. 16) generated by the extraction unit 308.

(Format processing unit 310)
The format processing unit 310 converts the extracted image 800 after distortion correction by the distortion correction unit 309 into an image file of a predetermined format. Examples of the predetermined format include JPEG and TIFF. The format processing unit 310 stores the generated image file in the storage unit 400.

  In the information processing unit 300 of the present embodiment, as described above, the extracted image 800 showing the shooting target area 510 extracted from the shot image 500 is generated by the processing of the four corner detection units 301 to 308, and the distortion is generated. The correction unit 309 performs distortion correction on the extracted image 800, and the extracted image 800 subjected to distortion correction is converted into an image file of a predetermined format and stored.

  In the present embodiment, the imaging target region 510 is extracted using both the edge detection result of the edge detection unit 306 and the background color detection result of the background color determination unit 305. Thereby, the extraction accuracy of the imaging target area 510 can be synergistically increased. This point will be specifically described below.

  When a book or the like is photographed as a subject of photographing, as shown in FIG. 2, in the photographed image 500, a shadow region 520 may be shown in the background region 550 (in FIG. 2, the upper side of the book in the background region 550). A shadow area 520 is formed in the vicinity). The shadow area 520 is an area showing a shadow due to the influence of illumination light and a user's shadow (unintended shadow) reflected in the background.

  The shaded area 520 is basically a gray color. Therefore, when the background color determination unit 305 determines whether or not each pixel is a background color pixel for the captured image 500 shown in FIG. 2, the shadow area 520 located in the background area 550 is not determined as a background color pixel. The case occurs.

  However, in the present embodiment, since the result of edge detection by the edge detection unit 306 is used together, the shadow area 520 is suppressed from being extracted as the imaging target area 510. This is because the shadow area 520 is not detected as an edge, and the boundary between the portion determined to be the background color pixel and the shadow area 520 has a lower edge strength than the boundary between the shooting target area 510 and the shadow area 520, This is because the boundary between the portion determined to be a color pixel and the shadow area 520 is rarely determined as an edge. That is, the drawback of the method using the determination result of whether or not it is a background color pixel is compensated by the method using the edge detection result.

  On the other hand, as described above, in the processing by the edge detection unit 306, the outline of the imaging target area 510 is detected as an edge. However, when a scratch or the like is formed on the desk, the background area 550 is also detected. In some cases, the scratch or the like is detected as an edge.

  However, although a scratch on the desk in the background region 550 may be detected as an edge as described above, it shows a color similar to the background color, and thus is highly likely to be determined as a background color pixel. Therefore, finally, the extraction unit 308 can suppress the extraction of the edge of the background area 550 (such as a scratch on the desk) as a part of the imaging target area 510. That is, the drawback of the method using the edge detection result is compensated by the method using the determination result of whether or not it is a background color pixel.

  As described above, in this embodiment, since the edge detection result of the edge detection unit 306 and the background color detection result of the background color determination unit 305 are used in combination, the disadvantages of each other are compensated. The extraction accuracy of 510 can be increased synergistically.

  FIG. 16 shows an extracted image 800 that is finally generated. In the extracted image 800, the shaded area 520 in FIG. 2 is filled with the solid image area that forms the background, and is extracted together with the imaging target area 510. You can see that it has not been done.

  In the above embodiment, as shown in FIG. 5, the background region 550 is divided into a plurality of sub-regions A to H, and a reference region having a predetermined size including the center position of the sub-region is set for each sub-region. Statistical processing of the gradation value of the reference area is performed for each area. As a result, the statistical processing can be performed on a region that is not easily affected by the edge or shadow of the subject to be photographed (such as a book), so that the judgment accuracy of the background color judgment unit 305 can be improved, thereby photographing. The extraction accuracy of the target area 510 can also be improved.

  In the present embodiment, as shown in FIG. 7, the condition setting unit 304 removes the sub-regions (A, E) where the gradation average value is the maximum value and the minimum value, The numerical value range of the same or similar color is obtained. The reason is as follows. In a part of the background area 550, the gradation value becomes extremely high as compared with other areas due to light reflection (reflection of illumination, etc.), or a shadow (shadow of the photographer) is accidentally formed. As a result, there may be a unique situation in which the gradation value becomes extremely low or the standard deviation becomes large compared to other regions. Therefore, by excluding the area where the average gradation value is the maximum value and the minimum value, the calculation accuracy of the numerical value range of the gradation value of the same or similar color as the representative color of the background area 550 is improved.

  The information processing unit 300 only needs to be able to generate an extracted image 800 (FIG. 16) obtained by extracting the shooting target region 510 from the shot image 500, and the distortion correction unit 309 that is not directly related to the generation of the extracted image 800 and The format processing unit 310 may not be particularly provided. Further, the smoothing processing unit 302 may not be provided.

  Further, although the edge detection unit 306 detects the edge by the canny method, a method other than the canny method may be used. For example, it may be determined whether the pixel of interest is an edge pixel by obtaining a difference from a pixel adjacent in the horizontal direction or the vertical direction and performing threshold processing on the difference. The threshold value is, for example, 48 when the image data is 8 bits.

  Alternatively, the edge strength may be obtained by a Sobel filter shown in FIG. 2 and a threshold process may be performed on the edge strength to determine whether or not the pixel is an edge pixel. The threshold is set to 96, for example, when the image data is 8 bits.

[Embodiment 2]
In the first embodiment, as shown in FIG. 5, the background region 550 is divided into eight sub-regions A to H, and statistical processing is performed on the sub-regions A to H. There is no need to perform statistical processing. For example, as shown in FIG. 17, the statistical processing unit 303 calculates four sub-regions i to 2 adjacent to each other in the vertical and horizontal directions of a rectangular region formed by connecting four candidate vertices with straight lines among the eight sub-regions. It is also possible to detect and set a reference range of 64 × 64 pixels centering on the central pixel of each region in each of the sub-regions a to b. Then, the statistical processing unit 303 obtains the statistical value of the gradation value for each channel (each channel of R, G, B) in the reference range in each of the sub-regions a to b. The example of FIG. 17 has an advantage that the calculation amount can be reduced as compared with the example of FIG. 5. According to the example of FIG. 5, the sample amount is larger than that of the example of FIG. It has the advantage of being.

  Alternatively, as shown in FIG. 18, among the blocks adjacent to the upper, lower, left, and right sides of a rectangular area formed by connecting four candidate vertices with straight lines, the block adjacent to the upper side is divided into seven sub-areas, etc. It may be divided at intervals (“A” to “MA” in FIG. 18), and for each sub-region, the statistical values of the gradation values of all the pixels in the sub-region may be obtained.

[Embodiment 3]
In the first embodiment, a tablet is used as an example of the imaging device 10, but the imaging device 10 may be various portable terminals (smartphones, tablets, notebook computers, mobile phones, and the like) having the imaging unit 200. Further, the photographing apparatus 10 may be a digital still camera, a general-purpose personal computer, or the like.

  In addition, the information processing unit 300 of the present embodiment is not necessarily provided in the imaging device 10 including the imaging unit 200, and may be provided in an apparatus that does not include the imaging unit 200. The point is that the information processing unit 300 can be provided as long as it is a device (a general-purpose personal computer or server device without a camera) that can receive a captured image from a mobile terminal with a camera. In addition, a multifunction peripheral or a large display device (such as an information display or an electronic blackboard) can receive a captured image, and the information processing section 300 may be provided in the multifunction peripheral or the large display device. Note that the communication means for the captured image between the portable terminal and the personal computer, server device, multifunction device, large display device, or the like may be wired or wireless. Also, communication via a network such as the Internet may be used.

  In addition, when transmitting a captured image from a mobile terminal to an apparatus including the information processing unit 300, the mobile terminal side performs a geometric correction process by performing a process for extracting a target area, and a target to be captured The user may be allowed to select the second mode in which the region extraction process and the geometric correction process are not performed. Mode information indicating which mode is selected is transmitted from the portable terminal together with the captured image to the apparatus including the information processing unit 300, and the information processing unit 300 is the first when the mode information indicates the first mode. The mode is executed, and when the mode information indicates the second mode, the second mode is executed.

[Embodiment 4]
As described above, the information processing unit 300 may be realized by software using a CPU, or may be realized by a logic circuit formed in an integrated circuit or the like. In the case of software, the information processing unit 300 includes a recording medium such as a ROM or a storage device in which a program that is the software is recorded so as to be readable by a computer (or CPU). The recording medium may be a “non-temporary tangible medium” such as a card, a disk, a semiconductor memory, or a programmable logic circuit. The program may be transmitted to the computer via any transmission medium (communication network, broadcast wave, etc.). Note that one embodiment of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission.

[Summary]
The information processing unit 300 (image processing apparatus) according to the first aspect of the present invention includes a four-corner detection unit 301 that detects candidate vertices that are candidates for vertices in the shooting target area 510 from the shot image 500, and a plurality of shot images 500. A region other than the region formed by connecting the candidate vertices with a line as a background region 550, and a condition setting unit 304 for setting a numerical value range (condition) of gradation values of the same or similar color as the representative color of the background region 550; A background color determination unit 305 (first determination unit) that determines whether or not each pixel of the photographed image 500 is a background color pixel that satisfies the numerical range, and whether or not each pixel is an edge pixel that indicates an edge. An edge detection unit 306 (second determination unit) for determination, a specifying unit 307 for specifying a pixel that is determined not to be the background color pixel and determined to be the edge pixel in the captured image 500, and specified to the specifying unit 307 The It was referring to pixels, an extraction unit 308 which extracts the imaging region 510 from the captured image 500, and further comprising a.

  According to the aspect 1 of the present invention, the imaging target region 510 is extracted from the captured image by using the edge detection result of the edge detection unit 306 and the determination result of the background color determination unit 305 in combination, and thereby the imaging target There is an effect that the extraction accuracy of the region 510 can be synergistically increased.

  That is, (a) in the configuration in which the imaging target region 510 is extracted using the determination result of the background color determination unit 305 without using the edge detection result of the edge detection unit 306, the shadow region 520 included in the background region 550 is captured. (B) In the configuration in which the imaging target area 510 is extracted using the edge detection result of the edge detection unit 306 without using the determination result of the background color determination unit 305, There is a drawback that the scratch image is extracted as the imaging target region 510 when the background region 550 has a scratch image (for example, a scratch on a desk). However, in the present invention, by using the edge detection result of the edge detection unit 306, the disadvantage (only the shadow region 520 is extracted as the imaging target region 510) when only the determination result of the background color determination unit 305 is used is suppressed. In addition, by using the determination result of the background color determination unit 305, it is possible to suppress a defect (only the scratch image is extracted as the imaging target region 510) when only the edge detection result of the edge detection unit 306 is used. That is, in the present invention, by using edge detection and background color determination together, it is possible to compensate for each other's defects and synergistically obtain the effect of improving the extraction accuracy of the photographing target.

  In addition, according to the information processing unit 300 according to the first aspect of the present invention, it is possible to accurately capture a shooting target portion without requiring an image in which a background (a pedestal such as a desk) in a state where there is no shooting target as in Patent Document 1 is required. Therefore, it is not necessary to take the image twice, and it is possible to reduce the storage capacity and the trouble for the user as compared with Patent Document 1.

  In addition to the configuration of the aspect 1, the information processing unit 300 of the aspect 2 of the present invention includes a statistical processing unit 303 that performs statistical processing on the gradation value of the statistical object, with the range included in the background region 550 as the statistical object, The condition setting unit 304 sets the numerical range based on the result of the statistical processing.

  Thereby, in the photographed image 500, a numerical value range of gradation values indicating the same or similar color as the representative color of the background region 550 can be set with high accuracy, and it can be accurately determined whether or not it is a background color pixel. There is an effect.

  In addition to the configuration of aspect 2, the information processing unit 300 according to aspect 3 of the present invention divides the statistical target range into a plurality of sub-regions, and generates a reference region of a predetermined size including the center position of each sub-region for each sub-region. And the statistical processing of the gradation value of the reference area is performed for each predetermined area.

  According to the third aspect of the present invention, the statistical processing can be performed on a region that is not easily affected by the edge or shadow of a shooting target (such as a book), and thus the determination accuracy of the background color determination unit 305 is improved. As a result, the extraction accuracy of the imaging target region 510 can be improved.

  In addition to the configuration of aspect 3, the information processing unit 300 according to aspect 4 of the present invention includes the condition setting unit 304 that includes a sub-region having a maximum average value or median value among the statistics obtained by the statistical processing, and the The condition is set with reference to the statistic of each sub-region excluding the sub-region having the minimum average value or median value.

  According to the aspect 4 of the present invention, even when the background region 550 includes a local portion in which the variation in the value increases due to the reflection of illumination or the shadow of the photographer, the influence of the local portion There is an effect that can be suppressed.

  The information processing unit 300 according to the fifth aspect of the present invention includes a smoothing processing unit 302 that performs a smoothing process on the captured image 500 in addition to any one of the second to fourth aspects, and the statistical processing unit 303 includes the smoothing process. The statistical processing is performed using the captured image 500 that has been subjected to the conversion processing.

  According to the aspect 5 of the present invention, it is possible to suppress the influence of noise in the captured image 500 from affecting the statistical processing, and it is possible to improve the accuracy of the statistical processing.

  The information processing unit 300 according to the sixth aspect of the present invention, in addition to any one of the first to fifth aspects, if the pixel specified by the specifying unit 307 is a specific pixel, the extraction unit 308 (a) For each pixel constituting the upper side, a first search process for searching for a specific pixel closest to the pixel from a pixel row arranged in a direction orthogonal to the upper side from the pixel; and (b) the captured image 500. A second search process for searching for a specific pixel at a position closest to the pixel from a pixel row arranged in a direction orthogonal to the lower side from the pixel for each pixel constituting the lower side of the pixel; and (c) a captured image. For each pixel constituting the left side of 500, a third search process for searching for a specific pixel closest to the pixel from a pixel row arranged in a direction orthogonal to the left side from the pixel; Configure the right side of image 500 For each pixel, a fourth search process for searching for a specific pixel closest to the pixel from a pixel row arranged in a direction orthogonal to the right side from the pixel; and (e) the first to fourth searches. An extraction process for extracting the specific pixel extracted by the process as an outline and extracting the outline and an area surrounded by the outline as an imaging target area 510 is performed.

  According to the sixth aspect of the present invention, it is possible to specify the outline of the shooting target area 510 from the specific pixel that is determined not to be the background color pixel and to be the edge pixel, and from this outline, the shooting target area can be specified. 510 can be specified.

  The photographing apparatus 10 according to the seventh aspect of the present invention includes the photographing unit 200 and the information processing unit 300 according to any one of the first to sixth aspects.

  The information processing unit 300 according to aspects 1 to 6 of the present invention may be realized by a computer. In this case, the information processing unit 300 is made to operate on each computer by causing the computer to operate as each unit included in the information processing unit 300. The program to be realized and the computer-readable recording medium on which the program is recorded also fall within the scope of the present invention.

  One aspect of the present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and the technical means disclosed in different embodiments can be appropriately combined. Such embodiments are also included in the technical scope of one aspect of the present invention.

  The present invention is applied to a device capable of installing image processing software for processing a photographed image or a device capable of mounting an image processing circuit for processing a photographed image. Examples of such devices include general-purpose computers, server devices, multifunction devices, digital still cameras, and mobile terminals (smartphones, tablets, notebook computers, mobile phones) and the like.

DESCRIPTION OF SYMBOLS 10 Imaging device 200 Imaging part 300 Information processing part (image processing apparatus)
301 Four corner detection unit 302 Smoothing processing unit 303 Statistical processing unit 304 Condition setting unit (setting unit)
305 Background color determination unit (first determination unit)
306 Edge detection unit (second determination unit)
307 Identification unit 308 Extraction unit 309 Distortion correction unit 310 Format processing unit 400 Storage unit 500 Captured image 510 Shooting target region 520 Shadow region 550 Background region 700 Mask image 800 Extracted image

Claims (9)

A detection unit that detects candidate vertices that are candidates for vertices in the shooting target region from the shot image;
A setting unit that sets a condition for a gradation value of the same or similar color as the representative color of the background area, with a background area other than an area formed by connecting a plurality of candidate vertices with lines in the captured image ,
A first determination unit that determines whether or not each pixel of the captured image is a background color pixel that satisfies the condition;
A second determination unit that determines whether or not each pixel is an edge pixel indicating an edge;
Among the captured images, a specifying unit that specifies a pixel that is determined not to be the background color pixel and determined to be the edge pixel;
An image processing apparatus comprising: an extraction unit that extracts the shooting target region from the captured image with reference to the pixel specified by the specifying unit. A statistical processing unit that performs statistical processing of gradation values using the range included in the background region as a statistical target;
The image processing apparatus according to claim 1, wherein the setting unit sets the condition based on a result of the statistical processing.   The statistical processing unit divides the range of the statistical object into a plurality of sub-regions, sets a reference range of a predetermined size including the center position of the sub-region for each sub-region, and sets a reference region gradation value for each sub-region. The image processing apparatus according to claim 2, wherein the statistical processing is performed.   The setting unit includes the statistics of each sub-region excluding the sub-region having the maximum average value or median value and the sub-region having the minimum average value or median value among the statistics obtained by the statistical processing. The image processing apparatus according to claim 3, wherein the condition is set with reference to an amount. A smoothing processing unit that performs a smoothing process on the captured image;
The image processing apparatus according to claim 2, wherein the statistical processing unit performs the statistical processing using the captured image that has been subjected to the smoothing processing. When the pixel specified by the specifying unit is a specific pixel,
The extraction unit searches (a) for each pixel constituting the upper side of the captured image, a search for a specific pixel located closest to the pixel from a pixel row arranged in a direction orthogonal to the upper side from the pixel. 1 search process, and (b) for each pixel constituting the lower side of the photographed image, a search is made for a specific pixel located closest to the pixel from a pixel row arranged in a direction orthogonal to the lower side from the pixel. 2 search processing, and (c) for each pixel constituting the left side of the photographed image, a search is made for a specific pixel located closest to the pixel from a pixel row arranged in a direction orthogonal to the left side from the pixel. 3 search processing, and (d) for each pixel constituting the right side of the photographed image, a search is performed for a specific pixel at a position closest to the pixel from a pixel row arranged in a direction orthogonal to the right side from the pixel. 4 search processing; and (e) the first The extraction process of extracting the specific pixel extracted in the fourth search process as an outline and extracting the outline and the area surrounded by the outline as the imaging target area is performed. The image processing apparatus according to claim 5.   An imaging apparatus comprising: an imaging unit; and the image processing apparatus according to claim 1.   A program that causes a computer to function as each unit of the image processing apparatus according to any one of claims 1 to 6.   A computer-readable recording medium on which the program according to claim 8 is recorded.

Images