JP2012053740A - Image processing method and image processing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and an image processing system with which the shape of a picture after distortion correction is not in a polygonal line but in a smoothly changing shape.SOLUTION: Based on the angle OT of a polar coordinate showing the notice pixel of a picture subject to distortion correction, distortion amount tables A, B which are set to be biaxial and sandwich the notice picture are selected and the distortion amounts A, B corresponding to the distance OR are calculated. From trigonometrical function of the angle OT, an interpolation coefficient is calculated, thereby performing interpolation of the distortion amounts A, B. By using an interpolation amount C, the distance OR is corrected. The polar coordinate expressed by the angle OT and the corrected distance IR is converted into the orthogonal coordinate, thereby obtaining reference coordinate showing the notice pixel of the image before distortion correction.

Description

  The present invention relates to an image processing method and an image processing apparatus that correct distortion generated in an image of captured data.

  When a digital image is captured by a digital camera or the like, an image in the digital image may be distorted due to the influence of distortion of the lens. This phenomenon is likely to occur in a camera using a lens with a simple structure. Techniques for correcting such image distortion in a digital image by digital image processing have been proposed so far.

  For example, the distortion correction apparatus described in Patent Document 1 calculates reference coordinates corresponding to a pixel of interest in an image to be output after correcting the distortion from the uncorrected image having distortion, and the calculated reference. By interpolating the pixels in the image before correction based on the coordinates, the coordinates of the pixels in the image after correction are calculated. When calculating reference coordinates from an image before correction, data representing distortion amount called distortion data is referred to. However, the distortion data can only represent a distortion having a shape in which the distortion amount depends only on the distance from the center of the image.

  As shown in FIG. 10, the lens 100 and the light receiving surface 101 of the image sensor are not necessarily directly facing each other, and are often slightly inclined as represented by an inclination φ in the drawing. When there is no inclination φ of the light receiving surface 101, the locus of points having the same distortion amount in the image is concentric as shown in FIG. 11A, but when there is an inclination φ, it is not concentric. A distorted circle shown in FIG. The distortion shape as shown in FIG. 11B cannot be correctly corrected by the correction method of Patent Document 1 because the distortion amount differs even when the distance from the center of the image is the same.

  As a method of correcting a distortion in which the shape of the distortion changes due to a factor other than the distance from the center as shown in FIG. 11B, there is a method described in Patent Document 2. FIG. 12 is a diagram for explaining a distortion correction method by the imaging apparatus according to Patent Document 2. In this method, a plurality of axes L1 to L8 extending radially outward from the center O in the image are set in advance. On each axis, a distortion amount curve with respect to the distance from the center O is designated to create a distortion amount curve. This distortion amount curve can be set independently for each axis. Then, the imaging apparatus performs distortion correction processing of the captured image using the approximate expression of the distortion amount curve created for each axis. Thereby, the distortion which changes with factors other than the distance from the center of an image can be corrected.

Japanese Patent No. 4104571 JP 2008-227582 A

  However, Patent Document 2 discloses a method for designating the amount of distortion independently for each axis, but does not disclose a specific method for calculating the amount of distortion for a pixel not on each axis. Absent.

  For example, as shown in FIG. 12, the distortion amount in the area ABCD surrounded by the points A and B on the axis L1 and the points C and D on the axis L8 is represented by points A, B, C, and D. If it is expressed by linear interpolation using the amount of distortion designated by each point, the locus of the same amount of distortion has the shape of a line segment represented by a dotted line in FIG. Since the shape of the locus of the same distortion amount is often an arc shape, when an image is corrected according to the distortion amount calculated by linear interpolation, the distortion-corrected pattern changes to a polygonal line shape at the boundary of the region ABCD.

  The present invention has been made to solve the above problems, and when correcting distortion that changes due to factors other than the distance from the distortion center position on the image, the shape of the pattern after distortion correction is It is an object of the present invention to provide an image processing method and an image processing apparatus that can be corrected to a shape that changes smoothly without being broken.

  An image processing method according to the present invention includes a distortion center coordinate setting step for setting a distortion center coordinate at a predetermined position on a distortion-corrected image obtained by distortion-correcting imaging data, and a radial pattern from the distortion center on the image after distortion correction. A distortion amount table setting step for setting a distortion amount table in which a distortion amount corresponding to the distance from the distortion center is specified for each of a plurality of straight lines extending in a straight line, and converting the orthogonal coordinates of the pixel of interest on the image after distortion correction into polar coordinates Based on the polar coordinate conversion step expressed by the distance and angle centered on the distortion center, and the angle on the polar coordinate, two straight lines sandwiching the pixel of interest are selected from a plurality of straight lines, and the two straight lines are set. The distortion amount according to the polar coordinate distance from the distortion amount table selection step for selecting two distortion amount tables and the two distortion amount tables selected in the distortion amount table selection step. A distortion amount selection step for selecting each, a distortion amount interpolation step for calculating an interpolation distortion amount by interpolating the two distortion amounts selected in the distortion amount selection step by obtaining an interpolation coefficient from a trigonometric function representing an angle on polar coordinates, Orthogonal coordinate conversion step that corrects polar coordinate distance using the interpolation distortion amount calculated in the distortion amount interpolation step and converts it into orthogonal coordinates, and pixels in the vicinity of the orthogonal coordinates converted in the orthogonal coordinate conversion step on the image of the imaging data A pixel value setting step of interpolating the values and setting the pixel value of the pixel of interest on the image after distortion correction.

  An image processing apparatus according to the present invention sets a distortion center coordinate at a predetermined position on a distortion-corrected image obtained by distortion-correcting captured image data, and converts orthogonal coordinates of a target pixel on the distortion-corrected image into polar coordinates. Then, a polar coordinate conversion unit expressed by a distance and an angle centered on the strain center, and a strain amount corresponding to the distance from the strain center for each of a plurality of straight lines extending radially from the strain center on the strain-corrected image are specified. Based on the polar coordinates converted by the polar coordinate conversion unit, two distortion amounts are acquired from the two linear distortion amount tables sandwiching the pixel of interest, and an interpolation coefficient is obtained from a trigonometric function representing an angle on the polar coordinates. Interpolating the two distortion amounts to calculate the interpolation distortion amount, correcting the polar coordinate distance using the interpolation distortion amount and converting it to orthogonal coordinates, and on the image of the imaging data, Reference coordinate calculation There by interpolating the pixel values of the rectangular coordinate neighborhood converted, in which and a pixel interpolation processing unit for setting the pixel value of the target pixel on the distortion corrected image.

  According to this invention, based on the polar coordinates indicating the pixel of interest on the image after distortion correction, two distortion amounts are obtained from a two-line distortion amount table sandwiching the pixel of interest, and a trigonometric function representing an angle on the polar coordinate Since the interpolation is performed with the interpolation coefficient obtained from the above, when correcting distortion that changes due to factors other than the distance from the distortion center position on the image, the shape of the pattern after distortion correction is not a polygonal line, but smooth It is possible to provide an image processing method and an image processing apparatus that can correct to a changing shape.

It is a block diagram which shows the structure of the imaging device which concerns on Embodiment 1 of this invention. 3 is a flowchart illustrating an operation of an image processing unit in the imaging apparatus according to Embodiment 1. 6 is a diagram illustrating a distortion amount table set in a distortion-corrected image in the imaging apparatus according to Embodiment 1. FIG. 6 is a block diagram illustrating an operation of a reference coordinate calculation unit according to Embodiment 1. FIG. 6 is a flowchart illustrating an operation of a reference coordinate calculation unit according to Embodiment 1. 6 is a diagram illustrating an outline of a method of interpolating a distortion amount using an angle OT of polar coordinates by the reference coordinate calculation unit according to the first embodiment. FIG. FIG. 6 is a continuation diagram illustrating an outline of a method for interpolating a distortion amount using an angle OT of polar coordinates by the reference coordinate calculation unit according to the first embodiment. In the imaging device which concerns on Embodiment 2 of this invention, it is a figure explaining the distortion amount table set to the image after distortion correction. FIG. 10 is a block diagram illustrating an operation of a reference coordinate calculation unit according to Embodiment 2. It is a figure which shows the positional relationship of the lens of an imaging device, and an image sensor light-receiving surface. FIGS. 11A and 11B are diagrams illustrating distortion of a captured image. FIG. 11A illustrates a locus of the same distortion amount when the image sensor light-receiving surface faces the lens, and FIG. 11B illustrates the image sensor light-receiving surface with respect to the lens. The locus | trajectory of the same distortion amount when is inclined is shown. It is a figure explaining the distortion correction method of the imaging device concerning patent documents 2.

Embodiment 1 FIG.
An imaging apparatus 1 shown in FIG. 1 includes a lens 2, an image sensor 3 that photoelectrically converts light collected by the lens 2 and outputs a captured image (digital data), and an image processing unit 4 that corrects distortion generated in the captured image. And a memory 8 storing information necessary for distortion correction. The digital camera body that has the lens 2 and the image sensor 3 to generate a captured image, and the image processing apparatus that has the image processing unit 4 and the memory 8 to correct distortion of the captured image are configured separately. May be.

  The image processing unit 4 includes a polar coordinate conversion unit 5, a reference coordinate calculation unit 6, and a pixel interpolation processing unit 7. In order to correct distortion in the captured image, that is, the image before distortion correction, attention is paid to the image after distortion correction. The pixel value of the pixel is calculated. FIG. 2 is a flowchart showing the operation of the image processing unit 4. When the coordinates of the target pixel to be corrected when the captured image is subjected to distortion correction to generate an image after distortion correction are input to the image processing unit 4, the image processing unit 4 first stores the distortion center stored in the memory 8. With reference to the coordinates and the pixel aspect ratio, the pixel of interest is converted from orthogonal coordinates to polar coordinates (step ST1). Next, the reference coordinate calculation unit 6 refers to the distortion amount parameter stored in the memory 8, and calculates the reference coordinates in the pre-distortion correction image (that is, the captured image) corresponding to the polar coordinates (step ST2). Finally, the pixel interpolation processing unit 7 performs an interpolation operation using the calculated reference coordinates and the pixel value in the pre-distortion image output from the image sensor 3, and calculates the pixel value of the target pixel in the post-distortion correction image. Calculate (step ST3).

  In step ST1, the polar coordinate conversion unit 5 calculates polar coordinates corresponding to the orthogonal coordinates representing the target pixel of the image after distortion correction according to the following equations (1) and (2). At this time, the polar coordinate conversion unit 5 acquires the distortion center coordinates (CX, CY) and the pixel aspect ratio ASP from the memory 8.

OR = SQRT (POW ((OX−CX) × ASP, 2) + POW (OY−CY, 2))
... (1)
OT = atan ((OY−CY) ÷ (OX−CX) × ASP): When OX ≠ CX OT = 90: When OX = CX and OY> CY OT = 270: When OX = CX and OY <CY OT = 0: OX = CX and OY = CY (2)
(OR, OT): Polar coordinates of the pixel of interest after distortion correction (OX, OY): Coordinates of the pixel of interest after distortion correction (CX, CY): Center coordinates of distortion ASP: Pixel aspect ratio (horizontal / vertical)
SQRT (p): a function for calculating the square root of p POW (p, q): a function for calculating the value of p to the power of q atan (p): a function for calculating the arc tangent of p

  The coordinates (OX, OY) of the distortion-corrected image target pixel instructed from the outside are orthogonal coordinates represented by the coordinates of the X axis and the Y axis that are orthogonal to the horizontal direction and the vertical direction of the image data. Similarly, the distortion center (CX, CY) is also an orthogonal coordinate. On the other hand, the polar coordinates (OR, OT) converted by the polar coordinate conversion unit 5 are based on the distance OR from the distortion center coordinates (CX, CT) to the target pixel coordinates (OX, OY) and the starting line passing through the distortion center. It is represented by an angle OT that indicates the inclination of.

  A known method may be used to calculate the distortion center coordinates, and distortion center coordinates (CX, CT) are calculated based on the degree of distortion of the grid when the imaging apparatus 1 images the grid-shaped subject as a calibration process. May be calculated. In the first embodiment, the previously calculated distortion center coordinates (CX, CT) are stored in the memory 8.

  In the first embodiment, since the distortion amount is calculated according to the distance from the distortion center coordinates (CX, CT), when the pixel itself is not square (for example, the shape of one pixel of the image sensor 3 is When the image data is not square), the distance between the horizontal direction and the vertical direction of the image data is different, and the amount of distortion cannot be calculated correctly. Therefore, in the above formulas (1) and (2), a correct distance can be calculated by multiplying the horizontal or vertical direction by a coefficient ASP corresponding to the pixel aspect ratio.

  In subsequent step ST <b> 2, the reference coordinate calculation unit 6 refers to the target polar coordinates of the image after distortion correction and the distortion amount parameter held in the memory 8, and calculates the reference coordinates of the image before distortion correction.

FIG. 3 is a diagram for explaining a reference coordinate calculation method of the reference coordinate calculation unit 6. Holds the amount of distortion corresponding to the distance from the distortion center for the four axes 10-1 to 10-4 extending radially from the distortion center coordinates (CX, CY) in the image to the periphery and orthogonal to each other. The distortion amount tables 20-1 to 20-4 to be set are set in advance. The distortion amount tables 20-1 to 20-4 have a structure in which a plurality of combinations of distortion amounts of output values with respect to distances of input values are arranged in ascending order of input values. The distance as the input value and the distortion amount as the output value change discretely, and also change discretely among a plurality of sets.
The value set as the distortion amount in the distortion amount table uses the ratio of the distance from the distortion center before distortion correction to the distance from the distortion center after distortion correction. Instead of this ratio, the difference between the distance from the distortion center before distortion correction and the distance from the distortion center after distortion correction may be used.

  As the distortion amount parameter, in addition to the distortion amount table set independently for each of the axes 10-1 to 10-4, the horizontal axis passing through the distortion center (the broken line shown in FIG. The inclination θ of the axis 10-1 relative to the line) is also set in advance.

  When the target pixel represented by the target polar coordinates (OR, OT) is not directly above the axes 10-1 to 10-4, the distortion amount tables set on the two axes located between the target pixels. It is necessary to obtain a distortion amount by interpolating the value of. Hereinafter, a distortion amount interpolation method will be described.

FIG. 4 is a block diagram for explaining the operation of the reference coordinate calculation unit 6. FIG. 5 is a flowchart for explaining the operation of the reference coordinate calculation unit 6.
The reference coordinate calculation unit 6 refers to the angle OT of the pixel of interest obtained in the previous polar coordinate conversion process (step ST1) and the inclination θ of the axis 10-1 preset in the memory 8 to determine the pixel of interest. The distortion amount table set for one of the two axes sandwiched between them is selected as the distortion amount table A (step ST2-1A). Here, when viewed from the distortion center, the distortion amount table set on the axis on the left side of the straight line connecting the distortion center and the pixel of interest is the distortion amount table A, and the distortion amount table set on the axis on the right side of the straight line Is a distortion amount table B. First, the conditions for selecting the distortion amount table A are listed below.
When 0 degree <= OT−θ <90 degrees: Select the distortion amount table 20-1 When 90 degrees <= OT−θ <180 degrees: Select the distortion amount table 20-2 180 degrees <= OT−θ < When 270 degrees: distortion amount table 20-3 is selected. When 270 degrees <= OT-θ <360 degrees: distortion amount table 20-4 is selected.

Similarly, the reference coordinate calculation unit 6 refers to the angle OT of the target polar coordinate and the inclination θ of the axis 10-1 preset in the memory 8, and the other axis (the other axis between the two axes sandwiching the target pixel) ( That is, the distortion amount table set to the distortion amount table B is selected by selecting the distortion amount table set on the right side of the straight line connecting the distortion center and the pixel of interest when viewed from the distortion center (step ST2-1B). The conditions for selecting the distortion amount table B are listed below.
When 0 degree <= OT−θ <90 degrees: Select the distortion amount table 20-2 When 90 degrees <= OT−θ <180 degrees: Select the distortion amount table 20-3 180 degrees <= OT−θ < When 270 degrees: distortion amount table 20-4 is selected. When 270 degrees <= OT-θ <360 degrees: distortion amount table 20-1 is selected.

Subsequently, the reference coordinate calculation unit 6 calculates the distortion amount A from the distortion amount table A with reference to the distance OR of the target polar coordinates (step ST2-2A). When the distance OR matches the distance that is the input value of the distortion amount table A, the reference coordinate calculation unit 6 calculates the distortion amount that is the corresponding output value as it is as the distortion amount A.
On the other hand, when the distance OR does not coincide with the distance that is the input value of the distortion amount table A, the distortion amount A is calculated by interpolating neighboring input values and output values. Examples of the interpolation method include linear interpolation with respect to distance, spline interpolation, and the like, and any method may be used. As an example, the following formula (3) shows a formula for calculating the distortion amount A by linear interpolation.

DA = (TAO [N + 1] −TAO [N]) ÷ (TAI [N + 1] −TAI [N])
× (OR-TAI [N]) + TAO [N]
... (3)
However, when OR is equal to or greater than the maximum element of TAI [N], DA is set as the maximum element of TAO [N].
DA: Distortion amount A
OR: Polar coordinate distance OR
TAI [N]: Nth element of distortion amount table A input value (distance) TAO [N]: Nth element of distortion value table A output value (distortion amount) N: TAI [N] <= OR <TAI [N + 1] ] Element number N satisfying

  Similarly, the reference coordinate calculation unit 6 refers to the distance OR of the target polar coordinates and calculates the distortion amount B from the distortion amount table B (step ST2-2B). Similarly to the processing of the distortion amount A, the reference coordinate calculation unit 6 calculates the distortion amount that is the corresponding output value as the distortion amount B when the distance OR matches the distance that is the input value of the distortion amount table B. To do. If the distance OR does not match the distance that is the input value of the distortion amount table B, the distortion amount B is calculated by interpolating nearby input values and output values. Examples of the interpolation method include linear interpolation with respect to distance, spline interpolation, and the like, and any method may be used. As an example, the following formula (4) shows a formula for calculating the distortion amount B by linear interpolation.

DB = (TBO [M + 1] −TBO [M]) ÷ (TBI [M + 1] −TBI [M])
× (OR−TBI [M]) + TBO [M]
... (4)
However, when OR is equal to or greater than the maximum element of TBI [M], DB is set as the maximum element of TBO [M].
DB: Distortion amount B
OR: Polar coordinate distance OR
TBI [M]: Mth element of the distortion amount table B input value (distance) TBO [M]: Mth element of the distortion value table B output value (distortion amount) M: TBI [M] <= OR <TBI [M + 1 ] Element number M

  Subsequently, the reference coordinate calculation unit 6 interpolates the distortion amounts A and B, and calculates an interpolation distortion amount C corresponding to the position of the target polar coordinate (step ST2-3). That is, the distortion amount A calculated from the distortion amount table A set on one of the two axes sandwiching the pixel of interest, and the distortion amount B calculated from the distortion amount table B set on the other axis Is interpolated using the angle OT to calculate the interpolation distortion amount C of the pixel of interest.

  6 and 7 show an outline of the interpolation method using the angle OT. The illustrated example shows a state where the light receiving surface β of the image sensor 3 is inclined by an angle φ from the surface α perpendicular to the optical axis Z when the Z axis is the optical axis of the lens 2. At this time, the surface α coincides with the XY plane, and the central axis of the inclination of the light receiving surface β coincides with the Y axis.

The angle between the X′-Y ′ plane and the light receiving surface β is μ degrees on the X′-Y ′ plane (Y ′ axis is not shown) obtained by rotating the XY plane about the Z axis by θ degrees. And The angle θ is the inclination of the axis 10-1 described above. As shown in FIG. 7, when the distance from the strain center on the plane α without the inclination of φ degree is 1, the distance from the distortion center on the light receiving surface β inclined by φ degree can be expressed by the following equation (5). it can.
tan (μ) = cos (θ) ÷ cos (φ) (5)

  The reference coordinate calculation unit 6 can make the shape of the locus of the same distortion amount after the interpolation into an arc shape by using a trigonometric function of the angle OT as an interpolation coefficient between the distortion amounts according to the above equation (5). The shape changes smoothly.

  The following equation (6) is a specific example of an equation for calculating the interpolation distortion amount C from the distortion amount A and the distortion amount B by interpolation. In the following formula (6), a function that returns a cosine is used. However, as long as the content is the same, a function that returns a sine or a function that returns a tangent may be used.

DC = DA × | cos (OT−θ) | + DB × (1− | cos (OT−θ) |)
... (6)
DA: Distortion amount A
DB: Distortion amount B
DC: interpolation distortion amount C
OT: Polar coordinate angle OT
θ: inclination of axis 10-1 cos (p): function that returns cosine of p | p |: absolute value of p

Thus, when the target pixel represented by the target polar coordinates (OR, OT) is not directly above the axes 10-1 to 10-4, the two axes that are located between the target pixels are arranged on the two axes. The interpolation distortion amount C is calculated by interpolating the set distortion amount table values. In the interpolation, an angle from two axes is used, and the interpolation coefficient changes between 0: 1 and 1: 0.
On the other hand, when the pixel of interest is located immediately above any of the axes 10-1 to 10-4, the interpolation coefficient is 0: 1 or 1: 0, so the distortion amount set for the axis The distortion amount corresponding to the distance OR can be directly acquired as the interpolation distortion amount C with reference to the table.

  Subsequently, the reference coordinate calculation unit 6 corrects the distance OR of the polar coordinates of interest using the calculated interpolation distortion amount C, and calculates the distance IR from the distortion center in the image before distortion correction (step ST2-4). Distance IR calculation when the value set as the distortion amount in the distortion amount tables 20-1 to 20-4 is the ratio of the distance from the distortion center before distortion correction to the distance from the distortion center after distortion correction. The formula is shown in the following formula (7).

IR = OR × DC (7)
IR: distance from the distortion center in the image before distortion correction OR: distance from the distortion center in the image after distortion correction (focused polar coordinates)
DC: Interpolation distortion amount C

  In step ST2-4, the values set as distortion amounts in the distortion amount tables 20-1 to 20-4 are the distance from the distortion center before distortion correction, and the distance from the distortion center after distortion correction. In the case of the difference between them, the reference coordinate calculation unit 6 uses the following equation (8) instead of the above equation (7).

IR = OR + DC (8)
IR: distance from the distortion center in the image before distortion correction OR: distance from the distortion center in the image after distortion correction (focused polar coordinates)
DC: Interpolation distortion amount C

  Subsequently, the reference coordinate calculation unit 6 converts the calculated distance IR of the target polar coordinates of the pre-correction image and the angle OT of the target polar coordinates of the corrected image into orthogonal coordinates, and calculates reference coordinates (IX, IY). (Step ST2-5). Expressions (9) and (10) for converting the distance IR and the angle OT of the target polar coordinates into the reference coordinates (IX, IY) are shown below. The reference coordinates are calculated up to a unit of less than one pixel and used in the interpolation calculation (step ST3) by the pixel interpolation processing unit 7 in the subsequent stage.

IX = IR × cos (OT) + CX (9)
However, when IR = 0, IX = CX.
IY = IR × sin (OT) + CY (10)
However, when IR = 0, IY = CY.
OT: Angle of polar coordinates focused on image after distortion IR: Distance of polar coordinates focused on image before distortion correction (IX, IY): Reference coordinates (orthogonal coordinates) of pixel focused on image before distortion correction
(CX, CY): strain center coordinates cos (p): function for calculating cosine of p sin (p): function for calculating sine of p

Finally, in step ST3, the pixel interpolation processing unit 7 obtains a pixel value in the vicinity of the calculated reference coordinates (IX, IY) from the pre-distortion image output from the image sensor 3, performs an interpolation operation, and performs distortion. The pixel value of the target pixel of the corrected image is calculated. At the time of interpolation calculation, a value less than one pixel of the reference coordinates is used as an interpolation coefficient. The pixel value interpolation calculation in the pixel interpolation process may be linear interpolation, bicubic interpolation, or other methods.
On the other hand, when the orthogonal coordinates in the target pixel of the image before distortion correction are exactly integer values (that is, when the reference coordinates are one pixel), the pixel at the orthogonal coordinate position is set by setting the interpolation coefficient to 1: 0. The value can be used as it is as the pixel value of the pixel of interest.

  In addition, each part of the polar coordinate conversion unit 5, the reference coordinate calculation unit 6, and the pixel interpolation processing unit 7 may be configured by a dedicated circuit, or a program describing the processing contents of each unit is stored in a memory, and the CPU A microprocessor such as (Central Processing Unit) or DSP (Digital Signal Processor) may be configured to execute the program in the memory.

  As described above, according to the image processing method according to the first embodiment, the polar coordinate conversion unit 5 sets the distortion center coordinate on the distortion-corrected image obtained by distortion-correcting the imaging data, The reference coordinate calculation unit 6 sets four axes 10-1 to 10-4 that extend radially from the distortion center on the distortion-corrected image in an angular direction orthogonal to each other, and for each straight line, from the distortion center. The distortion amount table setting step for setting the distortion amount tables 20-1 to 20-4 in which the distortion amount corresponding to the distance is set, and the reference coordinate calculation unit 6 uses the orthogonal coordinates of the pixel of interest on the distortion-corrected image as polar coordinates. The polar coordinate conversion step that converts and expresses the distance and angle centered on the distortion center and the reference coordinate calculation unit 6 pays attention to the four axes 10-1 to 10-4 based on the angle on the polar coordinate. 2 axes with pixels in between The distance of polar coordinates from the distortion amount table selection step for selecting and selecting the two distortion amount tables set for the two axes, and the two distortion amount tables selected by the reference coordinate calculation unit 6 in the distortion amount table selection step A distortion amount selection step for selecting a distortion amount according to each, and an interpolation coefficient is obtained from a trigonometric function representing an angle in polar coordinates, and the two distortion amounts selected in the distortion amount selection step are interpolated to calculate an interpolation distortion amount. A distortion amount interpolation step, an orthogonal coordinate conversion step in which the reference coordinate calculation unit 6 corrects the polar coordinate distance using the interpolation distortion amount calculated in the distortion amount interpolation step, and converts it into orthogonal coordinates, and a pixel interpolation processing unit 7 Pixel value setting to set the pixel value of the pixel of interest on the image after distortion correction by interpolating the pixel value in the vicinity of the Cartesian coordinate transformed in the Cartesian coordinate transformation step It was configured with a step. For this reason, when correcting the distortion that changes due to factors other than the distance from the distortion center position on the image, the shape of the pattern after distortion correction is not a polygonal line but can be corrected to a smoothly changing shape.

  Further, according to the first embodiment, the polar coordinate conversion unit 5 is configured to perform correction using the pixel aspect ratio when converting from the orthogonal coordinates to the polar coordinates in the polar coordinate conversion step, so that one pixel of the image sensor 3 is corrected. Even when the shape is not square, correct polar coordinates of the target pixel can be obtained, and as a result, distortion correction accuracy can be improved.

Embodiment 2. FIG.
Since the imaging apparatus 1 according to the second embodiment has the same configuration as that of the imaging apparatus 1 shown in FIG. 1, the following description will be given with reference to FIG. In the second embodiment, the distortion amount parameter stored in the memory 8 is different from that in the first embodiment, and therefore the operation of the reference coordinate calculation unit 6 is also different. The following description will focus on the different parts.

  FIG. 8 is a diagram for explaining a reference coordinate calculation method of the reference coordinate calculation unit 6. A distortion amount table 20-holding distortion amounts corresponding to the distance from the distortion center with respect to K axes 10-1 to 10 -K extending in an arbitrary direction from the distortion center coordinates (CX, CY) in the image. 1 to 20-K is set in advance. Similar to the first embodiment, the distortion amount tables 20-1 to 20-K have a structure in which a plurality of combinations of distortion amounts of output values with respect to distances of input values are arranged in ascending order of input values. . The distance that is an input value and the amount of distortion that is an output value change discretely, and also change discretely between multiple sets.

  In the first embodiment, the inclination θ is set only for the axis 10-1. However, in the second embodiment, the inclinations θ1 to θK are set for the axes 10-1 to 10-K. .

FIG. 9 is a block diagram illustrating the operation of the reference coordinate calculation unit 6 according to the second embodiment, and corresponds to FIG. 4 in the first embodiment.
The reference coordinate calculation unit 6 uses the angle OT of the pixel of interest of the image after distortion correction obtained in the previous polar coordinate conversion process (step ST1 shown in FIG. 2) and the axes 10-1 to 10-10 set in advance in the memory 8. Referring to the inclinations θ1 to θK of −K, one of the two axes sandwiching the pixel of interest in between (ie, the axis on the left side of the straight line connecting the distortion center and the pixel of interest when viewed from the distortion center) Is selected as a distortion amount table A (step ST2-1A ′). The conditions for selecting the distortion amount table A are listed below.
If θ1 <= OT <θ2: Select the distortion amount table 10-1 If θ2 <= OT <θ3: Select the distortion amount table 10-2:::
When θK <= OT or OT <θ1: Select the distortion amount table 10-K

Similarly, the reference coordinate calculation unit 6 refers to the angle OT of the target polar coordinate and the inclinations θ1 to θK of the axes 10-1 to 10-K set in the memory 8 in advance, so that the biaxial axis sandwiching the target pixel therebetween. The distortion amount table set on the other axis (that is, the axis on the left side of the straight line connecting the distortion center and the pixel of interest when viewed from the distortion center) is selected as the distortion amount table B (step ST2-1B). '). The conditions for selecting the distortion amount table B are listed below.
When θ1 <= OT <θ2: Select the distortion amount table 10-2. When θ2 <= OT <θ3: Select the distortion amount table 10-3.
When θK <= OT or OT <θ1: Select the distortion amount table 10-1.

  The subsequent steps ST2-2A and ST2-2B are the same as steps ST2-2A and ST2-2B shown in FIG.

  Subsequently, the reference coordinate calculation unit 6 interpolates the distortion amounts A and B using the following equation (11), and calculates an interpolation distortion amount C corresponding to the position of the target polar coordinate (step ST2-3 ').

DC = DA × (| cos (OT−θ1) | − | cos (θB−θ1) |)
÷ (| cos (θA−θ1) | − | cos (θB−θ1) |)
+ DB × (| cos (OA−θ1) | − | cos (θT−θ1) |)
÷ (| cos (θA−θ1) | − | cos (θB−θ1) |)
(11)
DA: Distortion amount A
DB: Distortion amount B
DC: interpolation distortion amount C
OT: Polar coordinate angle OT
θA: inclination of the axis A where the distortion amount table A is set θB: inclination of the axis B where the distortion amount table B is set cos (p): function that returns the cosine of p | p |: absolute value of p

  Subsequent steps ST2-4 and ST2-5 are the same as steps ST2-4 and ST2-5 shown in FIG.

  As described above, according to the image processing method according to the second embodiment, the reference coordinate calculation unit 6 includes a plurality of axes 10-1 to 10-10 extending radially from the distortion center on the distortion-corrected image in the distortion amount table setting step. For each −K, a strain amount table 20-1 to 20-K in which a strain amount corresponding to the distance from the strain center is specified is set, and a plurality of axes are selected based on the polar coordinate angle in the strain amount table selection step. Two axes that sandwich the pixel of interest between 10-1 to 10-K are selected. For this reason, as in the first embodiment, when correcting distortion that changes due to factors other than the distance from the distortion center position on the image, the shape of the pattern after distortion correction is not a polygonal line, but smooth. The shape can be corrected to change. Further, by increasing the number of axes from four in the first embodiment to K, it is possible to cope with a finer distortion shape.

  Also in the second embodiment, as in the first embodiment, the polar coordinate conversion unit 5 may be configured to perform correction using the pixel aspect ratio when converting from the orthogonal coordinates to the polar coordinates in the polar coordinate conversion step. Good. With this configuration, the correct polar coordinates of the pixel of interest can be obtained even when the shape of one pixel of the image sensor 3 is not square, and as a result, distortion correction accuracy can be improved.

In the first and second embodiments, the cause of the distortion change has been described by taking the inclination of the light receiving surface of the image sensor with respect to the lens optical axis as shown in FIG. 6 as an example. When the image sensor light-receiving surface itself is slightly distorted from the plane, the shape of the distortion changes due to factors other than the distance from the center. Even in such a case, distortion of the image data can be corrected by using the image processing method according to the first and second embodiments. However, it is necessary to change the set position of the distortion center coordinates (CX, CT) according to the cause of the distortion change.
It should be noted that the embodiments can be appropriately combined, changed, omitted, etc. within the scope of the present invention.

  DESCRIPTION OF SYMBOLS 1 Imaging device, 2 Lens, 3 Image sensor, 4 Image processing part, 5 Polar coordinate conversion part, 6 Reference coordinate calculation part, 7 Pixel interpolation process part, 8 Memory, 10-1-10-K, L1-L8 axis | shaft, 20 −1 to 20-K Distortion amount table, 100 lens, 101 Image sensor light receiving surface.

Claims (4)

In an image processing method for correcting distortion generated in an image of imaging data,
A distortion center coordinate setting step for setting a distortion center coordinate at a predetermined position on the image after distortion correction obtained by distortion correcting the imaging data;
A distortion amount table setting step for setting a distortion amount table specifying a distortion amount according to a distance from the distortion center for each of a plurality of straight lines extending radially from the distortion center on the image after distortion correction;
A polar coordinate conversion step that converts the orthogonal coordinates of the pixel of interest on the image after distortion correction into polar coordinates, and represents the distance and angle centered on the distortion center;
A distortion amount table selecting step of selecting two straight lines sandwiching the pixel of interest between the plurality of straight lines based on the polar coordinate angle, and selecting two distortion amount tables set to the two straight lines;
A strain amount selection step of selecting a strain amount according to the distance of the polar coordinates from the two strain amount tables selected in the strain amount table selection step;
A distortion amount interpolation step of calculating an interpolation distortion amount by interpolating two distortion amounts selected in the distortion amount selection step by obtaining an interpolation coefficient from a trigonometric function representing an angle on the polar coordinate;
An orthogonal coordinate conversion step for correcting the distance of the polar coordinates using the interpolation distortion amount calculated in the distortion amount interpolation step and converting it into orthogonal coordinates;
A pixel value setting step of interpolating a pixel value in the vicinity of the orthogonal coordinates converted in the orthogonal coordinate conversion step on the image of the imaging data, and setting the pixel value of the pixel of interest on the image after distortion correction An image processing method characterized by the above.   The distortion amount table setting step sets four straight lines extending radially from the distortion center on the image after distortion correction and extending in an angular direction perpendicular to each other, and each line corresponds to a distance from the distortion center. The image processing method according to claim 1, wherein a distortion amount table specifying a distortion amount is set.   3. The image processing method according to claim 1, wherein the polar coordinate conversion step performs correction using a pixel aspect ratio when converting from orthogonal coordinates to polar coordinates. In an image processing apparatus that corrects distortion generated in an image of imaging data,
A distortion center coordinate is set at a predetermined position on a distortion-corrected image obtained by distortion-correcting the imaging data, and the orthogonal coordinates of the pixel of interest on the distortion-corrected image are converted into polar coordinates, so that the distortion center is centered. A polar coordinate conversion unit represented by the distance and angle,
For each of a plurality of straight lines extending radially from the distortion center on the image after distortion correction, a distortion amount table specifying a distortion amount according to a distance from the distortion center is set, and the polar coordinates converted by the polar coordinate conversion unit To obtain two distortion amounts from a two-line distortion amount table with the pixel of interest in between, obtain an interpolation coefficient from a trigonometric function representing an angle on the polar coordinates, and interpolate the two distortion amounts. A reference coordinate calculation unit that calculates an interpolation distortion amount, corrects the distance of the polar coordinates using the interpolation distortion amount, and converts it into orthogonal coordinates;
A pixel interpolation processing unit that interpolates the pixel values in the vicinity of the orthogonal coordinates converted by the reference coordinate calculation unit on the image of the imaging data and sets the pixel values of the pixel of interest on the image after distortion correction. An image processing apparatus.

Images