JP2008282187A - Image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the following problem: though generating an artifact such as jaggies, ringing or the like in a slant line part or an edge part of an image in a nearest neighbor method, a linear interpolation method, or a high-order interpolation method because regarding a pixel as a point in conventional magnification-reduction technique, the jaggies are not conspicuous or does not occur if the pixel is the ideal point, but an area of unsmoothness is largely magnified because the pixel has an area. <P>SOLUTION: The pixel is handled as a pixel having the area. A decision part 104 detects a slant line present inside input image data by a color difference value and a luminance value of each the pixel inside an input pixel group and the input pixel group around an output pixel after magnification processing. A coefficient calculation part 105 configures a slant line zone by the pixels each having the area based on slant line direction information, decides whether the slant line zone crosses the output pixel having the area or not, and calculates an interpolation coefficient from an area ratio of a crossing portion between the slant line zone and the input pixel when they cross. An interpolation part 106 interpolates the color difference value and the luminance value of the output pixel based on the interpolation coefficient or the like. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image processing device for enlarging or reducing an image. For example, when an image or video is displayed on a display device such as a home television, the resolution of the input image is matched with the resolution of the display device. The present invention relates to an image processing apparatus.

  A conventional image processing apparatus performs enlargement or reduction of an image using any of the nearest neighbor method, the bilinear method, and the bicubic method (see, for example, Non-Patent Document 1). .

Among these methods, the nearest neighbor interpolation method is the luminance value, color difference value or RGB value of the pixel P _{i, j + 1} located closest to the interpolation pixel P _{u, v} newly increased by the image enlargement process. In this method, values indicating color information (hereinafter collectively referred to as pixel values) are used as pixel values of the interpolated pixels _{Pu, v} as they are. The nearest neighbor method has the advantage that it can be processed at high speed because it is the simplest method. This method also has the advantage of high contrast after image enlargement. However, in this method, jaggies are generated in the diagonal lines, so that the image quality becomes very poor.

The bilinear interpolation method uses pixel values of four pixels P _{i, j} , P _{i, j + 1} , P _{i + 1, j} , P _{i + 1, j + 1} around the interpolation pixel P _{u, v.} This is a technique for performing planar interpolation. In this method, jaggy does not occur in the oblique line unlike the nearest neighbor method, and an image smoothing effect can be obtained. However, there is a problem that a portion where the pixel value changes rapidly such as an edge portion is blurred.

The bicubic method is a method of performing curved surface interpolation using a cubic function using pixel values from 16 pixels P _{i, j} to P _{i + 3, j + 3} around the interpolation pixel P _{u, v.} is there. With this method, an edge enhancement effect is obtained, blurring is less likely, and jaggies with diagonal lines are less likely to occur. However, there is a problem that a large amount of calculation is required as compared with the above two interpolation methods.

  Thus, as an improvement measure of these classical methods, an image processing apparatus is proposed that includes means for adaptively correcting the edge portion in consideration of human visual characteristics (see, for example, Patent Document 1). . A multi-rate processing unit developed from the classical method, and a processing unit that adaptively determines whether enlargement or reduction is possible from the relationship between the target pixel to be enlarged and its neighboring pixels before enlargement, and performs enlargement or reduction. There has also been proposed an image processing apparatus provided with (see, for example, Patent Document 2).

Japanese Patent Application Laid-Open No. 2004-133592 (pages 8 to 12, FIGS. 2 and 3) JP 2001-251500 A (pages 7 to 14, FIG. 1) Mikio Takagi, “Image Analysis Handbook”, University of Tokyo Press, January 17, 1991, p. 441-444

  When the above classical method is used, when the enlargement ratio is large or the reduction ratio is small, blurring increases or artifacts such as ringing and jaggy occur around the edge of the image. There is. Moreover, even if adaptive processing is added to the classical method to reduce these artifacts, there is a problem that the maximum magnification that can be obtained is only twice, and the effect is limited. In addition, since the adaptive processing is introduced, there is a problem that the calculation cost is significantly increased as compared with the classical method.

  In addition, the pixel area is not considered at all by considering the pixel as a point, or even when the pixel area is considered, the overlap between the diagonal line and the pixel is not considered, so jaggies, ringing, etc. There is also a problem that artifacts are generated.

  In addition, since the conventional method (Patent Document 2) introduces multi-rate processing, which is an extension of the classical method, to the enlargement / reduction algorithm, this point alone is at least as high as or more than the classical method. There is a problem that the calculation cost is high. In addition, as in Patent Document 1 and Patent Document 2, when adaptive processing is applied to the image content, there is a problem that the calculation cost is significantly increased compared to the classical method. As described above, the conventional method has a problem that the calculation cost is increased.

  The present invention has been made in view of such a technical situation, and an object of the present invention is to use an image interpolation method that takes into account the area of pixels and the area of oblique line components that constitute an image, and reduces the calculation cost to a classical method. It is an object to provide an image processing apparatus for enlarging or reducing an image, capable of suppressing blurring and reducing artifacts such as ringing while suppressing the image to the same level as the above.

  An image processing apparatus according to a subject of the present invention includes an output pixel position determination unit that determines position coordinates of each output pixel of an output image after image enlargement processing in accordance with a desired image enlargement ratio, and the output pixel position determination A pixel group of an input image necessary for obtaining a pixel value composed of color information values such as a luminance value and a color difference value of the output pixel based on the position coordinates of the output pixel output from the unit, and the input pixel group An input image pixel coordinate determination unit that determines the coordinates of each input pixel belonging to the input image, and an input image storage unit that holds pixel values composed of color information values such as luminance values and color difference values of the input pixels of the input image And the pixel value of each input pixel in the input pixel group from the input image storage unit based on the coordinates of each input pixel in the input pixel group output from the input image pixel coordinate determination unit. Input image acquisition unit to acquire The coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit and the pixel value of each input pixel belonging to the input pixel group output from the input image acquisition unit, In the case where each input pixel belonging to the input pixel group is regarded as one point, the pixel among all the input pixels belonging to the input pixel group determined in the vicinity of a predetermined dimension of the output pixel to be enlarged An oblique line direction determination unit that detects an oblique line passing through two input pixels whose values are interrelated and outputs information of the oblique line, and each input pixel and the output pixel belonging to the input pixel group is a point Instead, the position coordinates of the output pixel that is the enlargement target transmitted from the output pixel position determination unit and the output from the oblique line direction determination unit The diagonal line information (1) The oblique line given by the oblique line information is a band-like oblique line band having a predetermined width determined based on the size of the shape of each pixel in a direction perpendicular to the direction of the oblique line. (2) Whether or not the diagonal line zone intersects with the output pixel as a two-dimensional body that exists at the position coordinates of the output pixel and has a region having the same size as the shape of each pixel (3) When the diagonal line zone intersects with the output pixel as the two-dimensional body, it is determined that the diagonal line zone is an effective diagonal line zone, and then intersects the diagonal line zone. A coefficient for calculating the area of the intersection of each input pixel of the two-dimensional body and the oblique line band, further calculating the area ratio of the area of each intersection, and calculating an interpolation coefficient based on at least the area ratio The calculation unit and the input image output from the input image acquisition unit Based on the pixel value of each input pixel in the prime group and the interpolation coefficient output from the coefficient calculation unit, output from the interpolation unit that interpolates the pixel value of the output pixel and the output pixel coordinate determination unit An output unit that outputs the position coordinates of the output pixel and the pixel value of the interpolated output pixel output from the interpolation unit.

  Hereinafter, various embodiments of the subject of the present invention will be described in detail along with the effects and advantages thereof with reference to the accompanying drawings.

  According to the subject of the present invention, the pixel is treated as a two-dimensional body having an area rather than a point, and image interpolation is performed when the image is enlarged or reduced. As a result, it becomes possible to fundamentally reexamine the cause of jaggies that occur, and as a result, the jaggies can be suppressed without increasing the calculation cost. In addition, there is an effect that ringing generated at the edge portion during interpolation can be suppressed.

(Embodiment 1)
The feature point of the present embodiment is that “the input pixel group is based on the position coordinate and pixel value information of each input pixel (point) in the input pixel group located in the vicinity of the output pixel after image enlargement / reduction. , A straight line (corresponding to a straight line extending in the horizontal direction or the vertical direction), a diagonal line segment, or a diagonal line that exists in the input image data through two input pixels (two points) having a correlation between the pixel values. A line (hereinafter also simply referred to as a diagonal line) is detected, and then the pixel is regarded as a two-dimensional body having an area rather than a point, and the detected straight line is represented by the pixel having an area, Whether or not a diagonal line segment or a diagonal line band having a predetermined width (hereinafter referred to as a diagonal line band) is formed, and the diagonal line band intersects an area of the output pixel having an area. When crossing, each input that belongs to the input pixel group and crosses the diagonal line Obtain the area of the intersection of the pixel and the diagonal line band, calculate the interpolation coefficient based on the ratio of the areas of the intersection (hereinafter simply referred to as area ratio), and output the output based on the calculated interpolation coefficient The pixel value of the pixel is interpolated ". As a result, a function of suppressing jaggies that occur in the oblique line portion (jaggies occur due to rattling of pixel values of the target pixel and its surrounding pixels) and ringing that occurs in the edge portion is realized. Yes. Hereinafter, the characteristic points of the present embodiment will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration example of an image processing apparatus according to the present embodiment. This image processing apparatus has a function of enlarging or reducing a digital image by a rational number or an integer multiple. In the following, regarding reduction, it is considered to be enlargement of 1 times or less, for example, 1/2 enlargement, so “reduction” is also unified with the term “enlargement”. In the present embodiment, the pixel is not regarded as an ideal point (having no area) as in the conventional example, but as a two-dimensional body having an area as illustrated in FIG. think of. In FIG. 3, each pixel is a unit square having a length of 1, and the area of each pixel is also 1. Here, the shape of the pixel is not necessarily a square as shown in FIG. 3, but may be an arbitrary shape such as a polygon such as a rectangle, a circle, or an ellipse. However, it is necessary to have a shape that can define the area of the pixel. Further, the length forming the pixel may not be 1. Of course, the area may not be 1. Hereinafter, for convenience of description, it is assumed that the shape of the pixel is a unit square of length 1. In addition, the coordinates of the pixels (input pixels and output pixels) are set with reference to the coordinate axis of the output image.

  In FIG. 1, an output pixel position determination unit 101 determines each pixel position (output pixel position) of an image (output image) after enlargement processing in accordance with a desired image enlargement ratio. The input image pixel coordinate determination unit 102 belongs to the input image pixel group and the input pixel group necessary for obtaining the pixel value of the output pixel based on the output pixel coordinates output from the output pixel position determination unit 101. Determine the coordinates of each input pixel. The input image frame memory 109 is connected to a TV tuner or the like, and each input transmitted from the TV tuner or the like to each address corresponding to the coordinate position of each input image determined in advance by the coordinate axis of the input image. The pixel value of the pixel is held for one frame. The input image acquisition unit 103 is based on the coordinates of each input pixel in the input pixel group output from the input image pixel coordinate determination unit 102 (in this case, each input pixel output from the input image pixel coordinate determination unit 102). The coordinates of each input pixel can be associated with the address of the input image frame memory 109 by performing coordinate conversion for converting the coordinate of the output image to the coordinate axis of the input image. From the frame memory 109, pixel values (color information values such as luminance values and color difference values) of each input pixel in the input pixel group are acquired.

  Further, the oblique line direction determination unit 104 determines the coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit 102 and each input pixel belonging to the input pixel group output from the input image acquisition unit 103. Based on the pixel value, all inputs belonging to the input pixel group determined in the vicinity of the predetermined dimension of the output pixel (one point) to be enlarged when each input pixel is regarded as one point at this stage A straight line, a diagonal line segment, or a diagonal line (diagonal line) passing through two input pixels (two points) whose pixel values are in a correlation relationship among pixels is detected, and information on the diagonal lines (positions of the two points) Coordinate or diagonal equation).

  Further, at this stage, the coefficient calculation unit 105 regards each pixel as a two-dimensional body having an area or a region instead of a point, and outputs the output as an enlargement target transmitted from the output pixel position determination unit 101. Based on the position coordinates of the pixel and the oblique line information output from the oblique line direction determination unit 104, first, (1) the oblique line given by the oblique line information is set in a direction perpendicular to the direction of the oblique line. A band-shaped oblique band having a predetermined width determined based on the dimension of the shape of each pixel is configured. (2) The diagonal line band exists at the position coordinates of the output pixel and has the same dimension as the dimension of each pixel. It is determined whether or not it intersects with an output pixel as a two-dimensional body having an area determined from this area or the dimension. (3) When intersecting, it is determined that the constructed diagonal line band is an effective diagonal line band. And included in the diagonal line Calculating the area of the intersection of the input pixel (the two-dimensional form) with the oblique line zone, after further calculating the area ratio of the area of each intersection, to calculate the interpolation coefficient based on the area ratio.

  In addition, the interpolation unit 106 determines the pixel value of the output pixel based on the pixel value of each input pixel in the input pixel group output from the input image acquisition unit 103 and the interpolation coefficient output from the coefficient calculation unit 105. Is interpolated.

  In addition, the output unit 107 outputs the position coordinates of the output pixel output from the output pixel coordinate determination unit 101 and the interpolated pixel value of the output pixel output from the interpolation unit 106 in the correspondence in the frame memory 108. Output to the address you want. The frame memory 108 stores the position coordinates and pixel values of the output pixels that have been transmitted at the corresponding addresses.

  Next, functions or operations of the image processing apparatus of FIG. 1 will be described.

  First, the output pixel position determination unit 101 calculates the output image size from the input image size in accordance with a desired magnification, and determines the output pixel position in the order of raster scanning. As shown in FIG. 2, the raster scan order is a method in which an order is assigned from the upper left to the right of an image and is used in a scanning line method such as a TV. The output size need not be calculated every time the output pixel position is determined, and the output size may be calculated only once at the first time. Further, the output pixel positions need not be arranged in the raster scan order, and may be arranged in an arbitrary order. However, in the following description, for convenience, it is assumed that the output pixels are arranged in the raster scan order.

Next, the input image pixel coordinate determining unit 102 determines an input pixel position group necessary for calculating or interpolating the pixel value of each output pixel position determined by the output pixel position determining unit 101. Here, the number of input pixels included in the input image position group necessary for calculating the output pixel value varies depending on the selected interpolation method. For example, as shown in FIG. 3, an input pixel is defined as X and an output pixel is defined as Y. In the case of enlarging the image at the same time in the vertical and horizontal directions, when considering up to the primary vicinity, the input image pixel coordinate determination unit 102 determines that the four input pixels X _{m, n} , X _{m + 1, n} , X _{m, n + 1} , X _{m + 1, n + 1} is determined, and each input pixel X _{m, n} , X _{m + 1, n} , X _{m, n + 1} , X belonging to the input image position group is determined. Determine the coordinates of _{m + 1, n + 1} and output. Similarly, when considering up to the second-order neighborhood, X _{m−1, n−1} , X _{m, n−1} , X _{m + 1, n−1} , X _{m + 2, n−1} , X _m− The coordinates of 16 pixels of _{1, n} ,... X _{m + 2, n + 2} are determined and output. At this time, the distance of the X _{i, j} coordinate based on the coordinate axis of the output pixel Y _{i, j} is also calculated according to the enlargement ratio. For example, in the case of FIG. 3 _, the distance between the input pixel X _{i, j} and the input pixel X _{i + 1, j} is 2. In addition, the range of the input image position group may be larger than the secondary vicinity, and when different algorithms are used in the vertical direction and the horizontal direction, for example, it is a secondary vicinity in the vertical direction and 1 in the horizontal direction. When the range of the input image position group is determined in the next vicinity, a rectangular region may be considered as the input image position group.

  Further, the enlargement may be performed independently in the vertical and horizontal directions. In that case, the system of FIG. 1 is prepared independently in the vertical and horizontal directions, or the same system is used twice in the vertical and horizontal directions by adding a function of rotating the image.

  Except for the case shown in FIG. 3, the positional relationship between the input pixel and the output pixel differs depending on the magnification. Also, the input pixel and the output pixel may overlap depending on the magnification. FIG. 4 shows an example where there is an overlap between the input pixel and the output pixel. As shown in FIG. 4, when the input pixel and the output pixel do not overlap at all, when the output pixel overlaps with one input pixel, when the output pixel overlaps with two input pixels, and when there are four output pixels May overlap with the input pixel.

  When the output pixel position is located at the edge of the screen, an approximation operation is performed. For example, there may be no 16 input pixels. The input image pixel coordinate determination unit 102 also detects the screen edge at the same time. In particular, in the case of the screen edge, the input image pixel coordinate determination unit 102 ignores the pixel (interpolates only with the effective pixel) or enters a zero value for a portion where the secondary neighboring region does not exist. Alternatively, an approximation operation in the vicinity of the second order, such as folding an existing pixel at the screen edge, is performed. In the following, an example in which the pixel value of the output pixel is interpolated only with the effective pixel at the screen edge in consideration of the secondary vicinity will be described.

  The input image acquisition unit 103 stores pixel values (color information values such as luminance values and color difference values) of the input pixels belonging to the input pixel position group determined by the input image pixel coordinate determination unit 102 as input image frame memories. Obtained by performing the above-described coordinate transformation from 109 corresponding addresses.

Next, the oblique line direction determination unit 104 acquires the position coordinates and the pixel value of each input pixel in the input pixel group from the input image pixel coordinate determination unit 102 and the input image acquisition unit 103, respectively. At this stage, since each pixel is regarded as one point, for example, it is assumed that the positional relationship between the input pixel group and the output pixel is as shown in FIG. 5 (each input pixel in FIG. 5). As shown later, the squares shown in FIG. 6 refer to the case where each input pixel is considered to be a two-dimensional body having an area (the same applies to FIG. 6). In this case, the oblique line direction determination unit 104 includes 16 input pixels X _{m−1, n−1} , X _{m, n−1} , X _{m + 1, n} that form the input pixel group of the output pixel Y _{i, j. −1} , X _{m + 2, n−1} , X _{m−1, n} ,..., X _{m + 2, n + 2} , _two of which have a certain correlation between pixel values An oblique line passing through the input pixel is detected. For example, as illustrated in FIG. 6, the oblique line direction determination unit 104 detects an oblique line DL that passes through two input pixels X _{m−1, n + 1} and X _{m + 1, n} . The “oblique line” mentioned here includes a straight line extending in the horizontal direction or the vertical direction. After detecting the oblique line DL, the oblique line direction determination unit 104 causes the coefficient calculation unit 105 to send oblique line information (for example, two input pixels X _{m−1, n + 1} , X _{m + 1, n} Corresponds to vector information such as position coordinates and oblique line DL equation). Note that the diagonal line may be a line segment that exists only in the vicinity of the secondary.

  When the diagonal line direction determination unit 104 determines that there is no diagonal line in the set input pixel group, the unit 104 detects a diagonal line extending in the vertical or horizontal direction. And At this time, which straight line in the vertical direction or horizontal direction is selected by the same unit 104 depends on the higher linearity of each input pixel (for example, the pixel value of each input pixel in the input pixel group is closer) There is also a method of selecting whether the input pixels are close to each other in the vertical direction or the horizontal direction. However, in this embodiment, the oblique line detection method in this case is not defined.

Next, the coefficient calculation unit 105 receives the diagonal line information and the position coordinates of the enlarged output pixel Y _{i, j} from the diagonal line direction determination unit 104 and the output pixel position determination unit 101, respectively. Image processing that is handled as a two-dimensional body having a predetermined area instead of a point is executed as follows. That is, since the output pixel Y _{i, j} and each input pixel are two-dimensional bodies having a predetermined shape and area, the coefficient calculation unit 105 determines the diagonal line determined based on the diagonal line information as the extension direction of the diagonal line. A belt-like oblique line band having a predetermined width dimension determined according to the dimension of each pixel is formed with respect to a direction orthogonal to. Then, it is determined whether or not the constructed diagonal line intersects the region of the output pixel Y _{i, j} having a predetermined two-dimensional shape and area. If it intersects, the “oblique line” detected by the oblique line direction determination unit 104 is appropriate. On the other hand, when it is detected that the intersection does not occur, the operation of the oblique line direction determination unit 104 and the determination operation of the coefficient calculation unit 105 are repeated until the intersection is detected.

In the case of crossing, the coefficient calculation unit 105 sequentially calculates the area of each crossing portion between the input pixel region included in the configured diagonal line band and the diagonal line band, and these crossing parts. The area ratio of the areas is calculated. Then, the coefficient calculation unit 105 calculates an interpolation coefficient based on the calculated area ratio. For example, consider a case where the positional relationship between the area of the output pixel Y _{i, j} and each input pixel in the input pixel area is given as shown in FIG. In this case, the coefficient calculation unit 105 intersects the area of the input pixels X _{m−1, n + 1} and X _{m + 1, n and} the area of the output pixel Y _{i, j} from the information of the oblique line DL in FIG. The diagonal line DLB having a width dimension of 1 is formed. That is, the oblique line band DLB has a rectangular shape having the same line width as the unit square. In this case, since the oblique line band DLB intersects the area of the output pixel Y _{i, j} , the coefficient calculating unit 105 intersects the area of the input pixel X _{m−1, n + 1} and the oblique line band DLB. The area of the hatched area (hatched area with hatching in FIG. 7) and the intersection of the input pixel X _{m + 1, n} region and the slanting line band DLB (hatched line with hatched hatching in FIG. 7) Part). After that, the coefficient calculation unit 105 calculates the ratio of the two areas obtained, and calculates the interpolation coefficient according to the area ratio. In the case of FIG. 7, there are only two pixels on the obtained oblique line band DLB, but the formed oblique line area overlaps more input pixel areas depending on the inclination of the oblique line and the output pixel position. It is also possible. In addition, the diagonal line band formed by the coefficient calculation unit 105 only needs to intersect the region of the output pixel Y _{i, j} even a little.

Now, in order to show a specific example of the operation of the coefficient calculation unit 105 and for the sake of simple explanation, consider the case of double enlargement considering the vicinity of the first order. In the case where the vertical and horizontal directions are enlarged / reduced at the same time, only in the vertical direction, or even when the magnification is changed, the interpolation coefficient in the case of double enlargement can be calculated in the same way. FIG. 8 shows an example of double enlargement considering the vicinity of the primary. Assuming that the oblique line direction determination unit 104 detects an auxiliary line as shown in bold in FIG. 8 as an oblique line, the oblique line band formed by the coefficient calculation unit 105 intersects the region of the output pixel Y _{i, j.} At the same time, it intersects with the area of each input pixel X _{m, n} , X _{m, n + 1} , X _{m + 1, n} , X _{m + 1, n−1} , and the area of each intersection can be obtained. If the area of each intersection is defined as A, B, C, and D, and the center of the output pixel Y _{i, j} is the origin, the equation of the auxiliary line is expressed by Equation (1).

  On the other hand, since the width of the diagonal line band is 1, the integral of the upper and lower line equations (Equation 2 (2), Equation 3 (3)) of the diagonal line band and the boundary line of the input pixel is obtained by integration. The area of each intersection can be calculated.

  From the above, the areas are obtained as A = D = √2−1 / 2≈0.91 and B = C = 0.25. By normalization, the final interpolation coefficients are A = D = 0.39 and B = C = 0.11.

The interpolation unit 106 uses the pixel value of the input pixel that intersects the diagonal line band among the pixel values of the input pixel output from the input image acquisition unit 103 and the interpolation coefficient information output from the coefficient calculation unit 105. The pixel value of the interpolated output pixel Y _{i, j} is calculated. In the case of the example of FIG. 8, the pixel value of the output pixel Y _{i, j} is calculated by the following equation (4).

  The output unit 107 receives the output pixel coordinate information from the output pixel position determination unit 101 and the output pixel value from the interpolation unit 106, and stores the output pixel coordinate information and the output pixel value at a predetermined address in the frame memory 108. .

  As described above, according to the present embodiment, interpolation is performed in consideration of the area of each pixel, so that it is possible to suppress the occurrence of pixel shakiness that is the root cause of jaggy that occurs in the oblique line portion. . In addition, according to the present embodiment, adaptive processing for diagonal lines (diagonal line band configuration, detection of presence / absence of crossing of diagonal line band and area of output pixel having area, calculation of area of intersection) In order to perform interpolation coefficient calculation based on the area ratio and interpolation processing based on the interpolation coefficient), the influence of noise generated by erroneously determining a part other than a diagonal line as a diagonal line is suppressed. I can do it.

(Embodiment 2)
FIG. 9 is a block diagram illustrating a configuration example of the image processing apparatus according to the present embodiment. In FIG. 9, the difference from the image processing apparatus in FIG. 1 according to the first embodiment is a coefficient calculation unit 205, and other components are the corresponding components in FIG. 1 according to the first embodiment. Is the same.

In the coefficient calculation unit 205, the center of each input pixel X _{m, n} and output pixel Y _{i, j} included in the oblique line band is compared with the interpolation coefficient calculation processing described above in the coefficient calculation unit 105 in FIG. An algorithm that takes into account the distance is also added. Specifically, it is as follows.

In the case shown in FIG. 8, the input pixel X _{m, n + 1} and the input pixel X _{m + 1, n−1} are equidistant from the output pixel Y _{i, j} . However, In one example shown in FIG. 7 described above, with respect to two input pixels included in the diagonal band _{DLB X m-1, n +} 1, X m + 1, n, the input pixel X _{m-1 , n + 1} and the output pixel Y _{i, j} are different from the center distance between the input pixel X _{m + 1, n} and the output pixel Y _{i, j} . Therefore, in such a case, not only the interpolation coefficient is calculated based on the above-described area ratio alone, but in addition to the area ratio, each input pixel and output pixel Y _i, If the interpolation coefficient is calculated in consideration of the inverse ratio of the center-to-center distance to _j , it is possible to obtain a higher-precision interpolation coefficient that reflects the actual pixel arrangement on the diagonal band. Therefore, the coefficient calculation unit 205 includes the position coordinate data of the output pixel Y _{i, j} transmitted from the output pixel position determination unit 101 _{, and} the position coordinate data of the input pixel transmitted from the input image pixel coordinate determination unit 102. , The center distance d1 between the input pixel X _{m−1, n + 1} and the output pixel Y _{i, j and} the center distance d2 between the input pixel X _{m + 1, n} and the output pixel Y _{i, j} calculate. In addition, the coefficient calculation unit 205, like the coefficient calculation unit 105, includes the diagonal line band DLB configured based on the diagonal line information output from the diagonal line direction determination unit 104 and each input pixel included in the diagonal line DLB. After calculating the area at the intersection of X _{m-1, n + 1} and X _{m + 1, n} , the area ratio of both areas is calculated, and in addition to this area ratio, the above calculated center-to-center The interpolation coefficient is calculated in consideration of the inverse ratio of the distances d1 and d2. Now, the area of the intersection (hatched portion) of the input pixel X _{m−1, n + 1} having the area and the oblique line band DLB is S1, and the other input pixel X _{m + 1, n} having the area and the oblique line band When the area of the intersection (shaded portion) with DLB is displayed as S2, the coefficient calculation unit 205 displays the interpolation coefficient for the input pixel X _{m−1, n + 1} as A, and on the other hand, the input pixel X _{m + 1. , n} is expressed as B, when the interpolation coefficient is normalized, the interpolation coefficient for each input pixel can be calculated based on the following equations (5) and (6). .

  In the equations (5) and (6), if d1 = d2 = 1, the normalized interpolation coefficients A and B are both of the intersection portions as in the first embodiment. It is calculated based only on the area ratio of the areas S1 and S2.

Further, depending on the magnification, as illustrated in FIG. 10, the output pixel Y _{i, j} and the input pixel X _{m, n} may partially overlap. The example of FIG. 10 is an example of 5/2 times enlargement only in the horizontal direction. At this time, in the above-described method described in the first half of the present embodiment based on the example of FIG. 7, the overlap between the input pixel and the output pixel is only considered in the distance between the centers.

Therefore, as a next method, the coefficient calculation unit 205 calculates an interpolation coefficient in consideration of the area of the overlapping portion. In the case of FIG. 10, the output pixel Y _{i, j} and the input pixel X _{m, n} have an overlap of half the area (0.5). Therefore, the coefficient calculation unit 205 considers the overlapping area as a contribution of the input pixel X _{m, n} and interpolates the remaining area from the diagonal line zone and the center-to-center distance. For example, it is assumed that an oblique line band as shown in FIG. 11 is detected by the coefficient calculation unit 205. At this time, as shown in FIG. 11, the results of obtaining the area of overlap between the diagonal line band and each input pixel of the input pixel group are displayed as A, B, C, and D, respectively. Further, the distance between the centers of the output pixel Y _{i, j} and each input pixel X _{m, n} , X _{m, n + 1} , X _{m + 1, n−1} , X _{m + 1, n} is defined as d1, It represents as d2, d3, d4. When all of the area ratio, the center-to-center distance, and the overlapping area are taken into account, the coefficient values A ′, B ′, C ′, and D ′ calculated and output by the coefficient calculation unit 205 are the following (7). It is given by the equations (10) to (10).

  When considering the area ratio, the center-to-center distance, and the overlapping area, the interpolation coefficient may be calculated in consideration of all elements as described above, or only the area ratio and the center-to-center distance may be calculated. The interpolation coefficient may be calculated in consideration, or the interpolation coefficient may be calculated in consideration of only the area and the overlapping area.

  As described above, according to the present embodiment, the pixel area, the center-to-center distance between the pixels, and the overlapping area are taken into consideration, so the occurrence of pixel shakiness that is the root cause of jaggies occurring in the oblique line portion is reduced. In addition to being able to suppress, the pixel value of the output pixel can be interpolated while guaranteeing the general property that there is a strong correlation between a certain pixel and its neighboring pixels. In addition, since adaptive processing is performed on diagonal lines, it is possible to suppress noise caused by erroneous determination of parts other than diagonal lines.

(Embodiment 3)
FIG. 12 is a block diagram illustrating a configuration example of the image processing apparatus according to the present embodiment. 12 is different from the image processing apparatus shown in FIG. 1 or FIG. 9 in the first and second embodiments in the configuration and function of the oblique line direction determination unit 304 and the coefficient calculation unit 305. The constituent elements are the same as the corresponding constituent elements of the image processing apparatus shown in FIG. That is, in the first and second embodiments, the effectiveness of the oblique line band formed by the coefficient calculation unit is determined in accordance with the position of the output pixel Y _{i, j} as a two-dimensional body. However, in the present embodiment, in order to reduce the amount of calculation _, the position of the output pixel Y _{i, j} as a two-dimensional body is not included in the oblique line band configured by the coefficient calculation unit. As shown in FIG. 13, the input pixels X _{m, n} , X _{m + 1, n} , X _{m, n + 1} , X _{m + 1, n + 1} located in the primary vicinity of the output pixel Y _{i, j} Using the center of gravity or the center position of the image (the position of the center of gravity Yc in FIG. 13) as a representative position, the effectiveness of the oblique line band is detected.

First, the oblique line direction determination unit 304 inputs the input pixels X _{m, n} , X _{m + 1, n} , X _{m, n + 1} , X _{m + 1} when each pixel is regarded as a point at this stage. _{, n + 1} center of gravity Yc (in this stage, the center of gravity Yc is also a point), and the coordinates of the center of gravity Yc are output to the coefficient calculation unit 305. The diagonal line direction determination unit 304 performs the same processing as the diagonal line direction determination unit 104 of FIG. 1 described in the first embodiment to detect diagonal lines, and the detected diagonal line information is used as a coefficient calculation unit. To 305. Here, if the position coordinate of the center of gravity Yc of the input pixel group is the same for the new output pixel Y _{i, j + 1} to be subjected to the enlargement process, the oblique line direction determination unit 304 determines that new output pixel. When newly calculating the interpolated pixel value of Y _{i, j + 1} , it is only necessary to send the same oblique line information to the coefficient calculation unit 305, so that it is not necessary to detect the oblique line every time.

The coefficient calculation unit 305 performs the following processing based on the position coordinates of the centroid Yc of the input pixel group and the diagonal line information sent from the diagonal line direction determination unit 304. First, the coefficient calculation unit 305 configures the diagonal line band as described above based on the diagonal line information, and then the configured diagonal line band has the same shape and the same size as each input pixel. It is detected whether or not it intersects the area of the center of gravity Yc as a two-dimensional body. If not intersecting, each of the units 304 and 305 performs the above processing until an intersecting diagonal line band is detected. When the crossing is detected, the coefficient calculation unit 305 obtains the area of each crossing portion between each input pixel that intersects the diagonal line band and the diagonal line band as described in the first embodiment, and further calculates those areas. The area ratio is calculated. Here, as in the case of the oblique line direction determination unit 304, the oblique line information is not changed unless the coordinates of the center of gravity Yc are changed. Therefore, the coefficient calculation unit 305 determines the area ratio regarding the new output pixel Y _{i, j + 1.} It is only necessary to calculate once for each center of gravity Yc.

In addition to the area ratio, the coefficient calculation unit 305 overlaps not only the area ratio but also the distance between the centers of the input pixels and the output pixels Y _{i, j} intersecting the diagonal line bands, as in the coefficient calculation unit 205 of the second embodiment. The interpolation count is calculated in consideration of the area. The center-to-center distance and the overlapping area need to be calculated every time regardless of the position of the center of gravity Yc.

Here, an example is shown in which the coefficient calculation unit 305 calculates the interpolation count in consideration of the area ratio, the input pixel value, the center-to-center distance between the output pixels Y _{i, j} and the overlapping area. The interpolation count may be calculated considering only the area ratio and the center-to-center distance, or the interpolation count may be calculated only considering the area ratio and the overlapping area.

  As described above, according to the present embodiment, the pixel area, the pixel distance, and the overlapping area are taken into consideration, and thus the occurrence of pixel shakiness that is the root cause of jaggies occurring in the oblique line portion is suppressed. In addition, interpolation can be performed while guaranteeing the general property that the correlation in the vicinity of the pixel is strong. In addition, since the adaptive processing is performed on the diagonal lines, it is possible to suppress the generation of noise due to erroneous determination of portions other than the diagonal lines. In particular, according to the present embodiment, there is an advantage that the calculation amount of the oblique line direction determination unit and the coefficient calculation unit can be reduced.

(Embodiment 4)
FIG. 14 is a block diagram illustrating a configuration example of the image processing apparatus according to the present embodiment. In the present embodiment, the functions of the input image acquisition unit 403, the coefficient calculation unit 405, and the interpolation unit 406 are different in comparison with the first embodiment (FIG. 1) and the second embodiment (FIG. 9). The other components are the same as the corresponding components in FIGS.

  In the present embodiment, when a vertical or horizontal straight band is formed in the coefficient calculation unit 405 as a diagonal line band based on the diagonal line direction information of the diagonal line direction determination unit 104, for example, FIG. This relates to a case where a straight band extending in the vertical direction as shown is detected as an oblique line band. At this time, the area of each intersection formed by the detected diagonal line band and each input pixel that intersects the diagonal line band is 1. At this time, even if the area ratio and the center-to-center distance ratio are taken into account, the interpolation of the output pixel value is linear interpolation. Therefore, in the present embodiment, an adaptation process (algorithm) is added for the case of a diagonal line extending in the vertical direction or the horizontal direction, which is a special diagonal line.

The input image acquisition unit 403 acquires, from the input image frame memory 109, pixel values of input pixels that belong to a range that is equal to or higher than the secondary vicinity of the output pixel Y _{i, j} . Here, when the bicubic method is used in the coefficient calculation unit 405 or the interpolation unit 406, the input image acquisition unit 403 determines the pixel values of the input pixels belonging to the range up to the third order vicinity of the output pixel Y _{i, j.} You may get When a higher-order interpolation method is used, the pixel values of input pixels belonging to a range of the vicinity of the third order or higher may be acquired. Here, in order to reduce the memory capacity of a buffer unit (not shown) in the oblique line direction determination unit 104, the input image acquisition unit 403 belongs to a range that is equal to or higher than the secondary vicinity of the output pixel Y _{i, j.} When handling pixels, the pixel values of pixels belonging to the higher-order neighborhood range may be acquired only in the horizontal direction and the vertical direction. One such example is shown in FIG. In FIG. 16, the input image acquisition unit 403 does not need to acquire the pixel values of the pixels at the four corners. Therefore, the memory capacity in the buffer unit of the oblique line direction determination unit 104 can be reduced. is there.

First, an example in which the coefficient calculation unit 405 performs coefficient calculation for high-order interpolation will be described. When the oblique line direction determination unit 104 determines that the oblique line is a special oblique line extending in the vertical direction or the horizontal direction, and the oblique line information is sent to the coefficient calculation unit 405, the coefficient calculation is performed. It is determined whether or not the diagonal line band formed by the unit 405 based on the diagonal line information intersects with the two-dimensional region of the output pixel Y _{i, j} . Instead of calculating the interpolation coefficient taking into account the area ratio, distance ratio and area size, instead, the interpolation coefficient is calculated based on a higher-order interpolation method such as the bicubic method, The interpolation coefficient is transmitted to the interpolation unit 406. The interpolation unit 406 interpolates the pixel value of the output pixel Y _{i, j} based on the interpolation coefficient obtained by the coefficient calculation unit 405 based on the higher-order interpolation method.

Next, a case where the interpolation unit 406 performs interpolation of diagonal lines extending in the vertical direction or the horizontal direction will be described. When the oblique line information regarding the oblique line extending in the vertical direction or the horizontal direction is sent from the oblique line direction determining unit 104, the coefficient calculating unit 405 is configured based on the oblique line information. It is determined whether or not the oblique line band intersects the two-dimensional region of the output pixel Y _{i, j} , and when it is confirmed that the oblique line band intersects, the oblique line extends vertically or horizontally with respect to the interpolation unit 406. It simply sends diagonal line information that it is an extended diagonal line. When the interpolation unit 406 receives information on the diagonal line extending in the vertical direction or the horizontal direction, the interpolation unit 406 obtains an interpolation coefficient using a high-order interpolation method such as a bicubic method, and then outputs the output pixel Y _i, Interpolate the pixel value of _j . On the other hand, when an interpolation coefficient value such as an area ratio as described above is sent from the coefficient calculation unit 405, the interpolation unit 406 outputs the output pixel Y as described in the first embodiment according to the received coefficient value. Interpolates the pixel values of _{i and j} .

  Here, although the bicubic method is taken up as a high-order interpolation method, a high-order interpolation method such as a spline method may be used, and other methods may be used.

Alternatively, the pixel values of the output pixels Y _{i, j} may be interpolated by the high-order interpolation method described above only in the vertical direction, or output using the high-order interpolation method described above only in the horizontal direction. You may perform the interpolation of the pixel value of pixel _{Yi, j} .

  As described above, according to the present embodiment, the pixel value of the output pixel is interpolated while suppressing the occurrence of pixel shakiness in the oblique line portion and ensuring the general property that the neighborhood of the pixel is strong. I can do it. In particular, according to the present embodiment, it is possible to perform high-order interpolation with respect to an oblique line band extending in the vertical direction or the horizontal direction, and interpolate the pixel value of the output pixel while preserving detailed image information. There is a feature that it can be done, and noise generated when a line other than a diagonal line is erroneously determined as a diagonal line can also be suppressed.

(Embodiment 5)
FIG. 17 is a block diagram showing an image processing circuit according to the present embodiment, which corresponds to an example in which the feature points of the present embodiment are applied to the image processing apparatus of FIG. 1 shown in the first embodiment. The input image line memory 509, which is a core component in FIG. 17, is connected to an input image supply unit (not shown) such as a TV tuner.

  In the first to fourth embodiments described above, the processing described above is performed after the pixel values of the pixels of the input image are stored in the frame memory. The object of the present embodiment is to reduce the processing delay time by using a line memory 509 instead of the frame memory.

The input image line memory 509 is equipped with a line memory capable of storing pixel values of each pixel for four lines of the input pixel when considering up to the secondary vicinity of the output pixel Y _{i, j} (FIG. 5). It ’s fine. As shown in FIG. 18, when the interpolation processing of the pixel value of the output pixel Y _{i, j} is completed and the next output pixel Y _{i, j ′} is the processing target, the pixel value of each pixel belonging to the uppermost line. The data is discarded, and the pixel value data of each pixel belonging to the next new line is taken into the input image line memory 509. At this time, it is easy to perform the processing order of the output pixels Y _{i, j in} the raster scan order, but it may not be in the raster scan order. However, when processing is performed by skipping a line, the capacity of the line memory increases, so that the configuration is possible, but this is not preferable.

Even when the third order vicinity of the output pixel Y _{i, j} is considered, the number of line memories is increased, but the memory capacity can be made smaller than that of the frame memory. In this way, the present embodiment can also be applied to the case where interpolation is performed in the vicinity of the third order of the output pixel Y _{i, j} .

  The input image line memory 509 in this embodiment is applicable not only to the first embodiment but also to any of the interpolation methods of the second, third, and fourth embodiments.

  As described above, according to the present embodiment, it is possible to reduce the memory capacity of the pixel value of the input pixel by mounting the line memory in the memory for the input pixel, and occurrence of pixel shakiness in the oblique line portion. In addition, the interpolation can be performed while guaranteeing the general property that the correlation between the pixels is strong.

(Appendix)
While the embodiments of the present invention have been disclosed and described in detail above, the above description exemplifies aspects to which the present invention can be applied, and the present invention is not limited thereto. In other words, various modifications and variations to the described aspects can be considered without departing from the scope of the present invention.

  The image processing apparatus according to the present invention is suitable for application to, for example, an image processing microprocessor used in a digital TV or a personal computer.

1 is a block diagram showing an image processing apparatus according to Embodiment 1 of the present invention. It is a figure which shows a raster scan order. It is a figure which shows the positional relationship of an input pixel and an output pixel when a pixel has an area. It is a figure which shows the example in case an input pixel and an output pixel overlap in position. FIG. 3 is a diagram illustrating a positional relationship between each input pixel and an output pixel in the first embodiment. It is a figure which shows the positional relationship of an oblique line, an input pixel, and an output pixel. It is a figure which shows the area of the cross | intersection part of a diagonal line zone and each input pixel. It is a figure explaining calculating an interpolation coefficient from the area of the intersection of a diagonal line zone and each input pixel. It is a block diagram which shows the image processing apparatus which concerns on Embodiment 2 of this invention. It is a figure which shows the case where an input pixel and an output pixel have overlapped. It is a figure which shows calculation of the interpolation coefficient when there exists overlap with an input pixel and an output pixel. It is a block diagram which shows the image processing apparatus which concerns on Embodiment 3 of this invention. It is a figure which shows the case where diagonal direction determination is performed in the primary vicinity of an output pixel. It is a block diagram which shows the image processing apparatus which concerns on Embodiment 4 of this invention. It is a figure which shows the case where the diagonal line | wire zone of a perpendicular direction is detected. It is a figure explaining acquiring the pixel value of an input pixel to the 3rd order neighborhood only in the perpendicular direction and the horizontal direction. It is a block diagram which shows the image processing apparatus which concerns on Embodiment 5 of this invention. It is a figure which shows operation | movement of the line memory for input images.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 101 Output pixel position determination part, 102 Input image pixel coordinate determination part, 103 Input image acquisition part, 104 Diagonal line direction determination part, 105 Coefficient calculation part, 106 Interpolation part, 107 Output part, 108 Frame memory, 109 Input image frame Memory, 205 Coefficient calculation unit in the second embodiment, 304 Diagonal line direction determination unit in the third embodiment, 305 Coefficient calculation unit in the third embodiment, 403 Input image acquisition unit in the fourth embodiment, 405 Coefficient calculation unit in the fourth embodiment, 406 Interpolation unit in the fourth embodiment, 509 Input image line memory in the fifth embodiment.

Claims (7)

An output pixel position determining unit that determines the position coordinates of each output pixel of the output image after the image enlargement process in accordance with a desired image enlargement ratio;
Based on the position coordinates of the output pixel output from the output pixel position determination unit, the pixel group of the input image necessary for obtaining a pixel value composed of values indicating the color information of the output pixel, and the input pixel group An input image pixel coordinate determination unit that determines the coordinates of each input pixel belonging to
An input image storage unit having a pixel value including a value indicating color information of each input pixel of the input image;
Based on the coordinates of each input pixel in the input pixel group output from the input image pixel coordinate determination unit, the pixel value of each input pixel in the input pixel group is acquired from the input image storage unit. An input image acquisition unit;
From the coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit and the pixel value of each input pixel belonging to the input pixel group output from the input image acquisition unit, the When each input pixel belonging to the input pixel group is regarded as one point, the pixel value among all the input pixels belonging to the input pixel group determined in the vicinity of the predetermined dimension of the output pixel to be enlarged Detecting a diagonal line passing through two input pixels having a phase relationship and outputting information of the diagonal line;
Each output pixel belonging to the input pixel group and the output pixel is not a point, but is regarded as a two-dimensional body having an area, and the output to be enlarged transmitted from the output pixel position determination unit Based on the position coordinates of the pixel and the diagonal line information output from the diagonal line direction determination unit, (1) the diagonal line given by the diagonal line information is a direction perpendicular to the direction of the diagonal line And (2) the diagonal line band is present at the position coordinates of the output pixel and the shape of each pixel is configured to form a band-shaped diagonal line band having a predetermined width determined based on the dimension of the shape of each pixel. It is determined whether or not it intersects with an output pixel as a two-dimensional body having an area having the same size as the dimension, and (3) when the diagonal line zone intersects with the output pixel as the two-dimensional body, the diagonal line zone Is determined to be a valid diagonal band, After calculating the area of the intersection of each input pixel of the two-dimensional body that intersects the diagonal line and the diagonal line, and further calculating the area ratio of the area of each intersection, at least based on the area ratio A coefficient calculation unit for calculating an interpolation coefficient;
An interpolation unit that interpolates the pixel value of the output pixel based on the pixel value of each input pixel in the input pixel group output from the input image acquisition unit and the interpolation coefficient output from the coefficient calculation unit When,
An output unit that outputs the position coordinates of the output pixel output from the output pixel coordinate determination unit and the pixel value of the interpolated output pixel output from the interpolation unit. ,
Image processing device. The image processing apparatus according to claim 1,
The coefficient calculation unit calculates, for each input pixel, a center-to-center distance between each input pixel of the two-dimensional body and the output pixel of the two-dimensional body that intersects the diagonal line band, and calculates each center-to-center distance and the area. The interpolation coefficient is calculated based on the ratio,
Image processing device. The image processing apparatus according to claim 2,
When the output pixel of the two-dimensional body overlaps with any input pixel of the two-dimensional body belonging to the input pixel group,
The coefficient calculation unit, for each of all input pixels that overlap the diagonal line band in each input pixel of the two-dimensional body belonging to the input pixel group, the area of the overlapping part of the input pixel and the diagonal line band To calculate
The coefficient calculation unit calculates the interpolation coefficient based on the area of each overlapping portion, the distance between the centers, and the area ratio,
Image processing device. An output pixel position determining unit that determines the position coordinates of each output pixel of the output image after the image enlargement process in accordance with a desired image enlargement ratio;
Based on the position coordinates of the output pixel output from the output pixel position determination unit, the pixel group of the input image necessary for obtaining a pixel value composed of values indicating the color information of the output pixel, and the input pixel group An input image pixel coordinate determination unit that determines the coordinates of each input pixel belonging to
An input image storage unit having a pixel value including a value indicating color information of each input pixel of the input image;
Based on the coordinates of each input pixel in the input pixel group output from the input image pixel coordinate determination unit, the pixel value of each input pixel in the input pixel group is acquired from the input image storage unit. An input image acquisition unit;
From the coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit and the pixel value of each input pixel belonging to the input pixel group output from the input image acquisition unit, the When each input pixel belonging to the input pixel group is regarded as one point, (1) the position coordinate of the center of gravity of each input pixel belonging to the input pixel group is obtained and output, and (2) the enlargement target The diagonal line passing through the two input pixels whose pixel values are interrelated among all the input pixels belonging to the input pixel group determined in the vicinity of the predetermined dimension of the output pixel, An oblique line direction determination unit that outputs information;
Each input pixel belonging to the input pixel group, the centroid and the output pixel are not a point but a two-dimensional body having an area, and the centroid output from the oblique line direction determination unit Based on the position coordinates and the oblique line information, (1) a predetermined width determined based on the size of each pixel in the oblique line given by the oblique line information in a direction perpendicular to the oblique line direction. (2) The barycentric region as a two-dimensional body having a region having the same size as the shape of each pixel, wherein the diagonal line zone exists at the position coordinates of the barycenter (3) When the diagonal line zone intersects the center of gravity region as the two-dimensional body, the diagonal line zone is determined to be an effective diagonal line zone, Each input pixel of the two-dimensional body intersecting the diagonal line and the diagonal Calculating the area of the intersection of the strip, the area ratio of the area of each intersection further on calculated based on at least the area ratio, and the coefficient calculation unit for calculating the interpolation coefficients,
An interpolation unit that interpolates the pixel value of the output pixel based on the pixel value of each input pixel in the input pixel group output from the input image acquisition unit and the interpolation coefficient output from the coefficient calculation unit When,
An output unit that outputs the position coordinates of the output pixel output from the output pixel coordinate determination unit and the pixel value of the interpolated output pixel output from the interpolation unit. ,
Image processing device. An output pixel position determining unit that determines the position coordinates of each output pixel of the output image after the image enlargement process in accordance with a desired image enlargement ratio;
Based on the position coordinates of the output pixel output from the output pixel position determination unit, the pixel group of the input image necessary for obtaining a pixel value composed of values indicating the color information of the output pixel, and the input pixel group An input image pixel coordinate determination unit that determines the coordinates of each input pixel belonging to
An input image storage unit having a pixel value including a value indicating color information of each input pixel of the input image;
Based on the coordinates of each input pixel in the input pixel group that is output from the input image pixel coordinate determination unit, the input image storage unit belongs to a range that is equal to or higher than the secondary vicinity in the input pixel group. An input image acquisition unit for acquiring the pixel value of each input pixel;
The coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit and the pixel value of each input pixel belonging to a range equal to or higher than the second order output from the input image acquisition unit All the input pixels belonging to a range determined by a predetermined dimension that is equal to or higher than the second order of the output pixel to be enlarged when each input pixel belonging to the input pixel group is regarded as one point. An oblique line direction determination unit that detects oblique lines passing through two input pixels in which the pixel values are interrelated, and outputs information of the oblique lines;
The diagonal lines are perpendicular to each other when the input pixels belonging to a range determined by a predetermined dimension equal to or greater than the second order of the output pixels and the output pixels are not points but are two-dimensional bodies having an area. Alternatively, when the oblique line information indicates that the straight line extends in the horizontal direction, the straight line has a predetermined width determined based on the size of each pixel in a direction perpendicular to the direction of the straight line. Whether or not the straight line intersects with the output pixel as a two-dimensional body that exists at the position coordinates of the output pixel and has a region having the same size as the shape of each pixel. When determining and intersecting, based on a high-order interpolation method, a coefficient calculation unit that calculates an interpolation coefficient;
Based on the pixel value of each input pixel belonging to a range determined by a predetermined dimension equal to or greater than the second order of the output pixel output from the input image acquisition unit, and the interpolation coefficient output from the coefficient calculation unit, An interpolation unit for interpolating a pixel value of the output pixel;
An output unit that outputs the position coordinates of the output pixel output from the output pixel coordinate determination unit and the pixel value of the interpolated output pixel output from the interpolation unit. ,
Image processing device. An output pixel position determining unit that determines the position coordinates of each output pixel of the output image after the image enlargement process in accordance with a desired image enlargement ratio;
Based on the position coordinates of the output pixel output from the output pixel position determination unit, the pixel group of the input image necessary for obtaining a pixel value composed of values indicating the color information of the output pixel, and the input pixel group An input image pixel coordinate determination unit that determines the coordinates of each input pixel belonging to
An input image storage unit having a pixel value including a value indicating color information of each input pixel of the input image;
Based on the coordinates of each input pixel in the input pixel group that is output from the input image pixel coordinate determination unit, the input image storage unit belongs to a range that is equal to or higher than the secondary vicinity in the input pixel group. An input image acquisition unit for acquiring the pixel value of each input pixel;
The coordinates of each input pixel belonging to the input pixel group output from the input image pixel coordinate determination unit and the pixel value of each input pixel belonging to a range equal to or higher than the second order output from the input image acquisition unit All the input pixels belonging to a range determined by a predetermined dimension that is equal to or higher than the second order of the output pixel to be enlarged when each input pixel belonging to the input pixel group is regarded as one point. An oblique line direction determination unit that detects oblique lines passing through two input pixels in which the pixel values are interrelated, and outputs information of the oblique lines;
The diagonal lines are perpendicular to each other when the input pixels belonging to a range determined by a predetermined dimension equal to or greater than the second order of the output pixels and the output pixels are not points but are two-dimensional bodies having an area. Alternatively, when the oblique line information indicates that the straight line extends in the horizontal direction, the straight line has a predetermined width determined based on the size of each pixel in a direction perpendicular to the direction of the straight line. Whether or not the straight line intersects with the output pixel as a two-dimensional body that exists at the position coordinates of the output pixel and has a region having the same size as the shape of each pixel. In the case of determining and intersecting, a coefficient calculation unit that outputs the oblique line information as it is;
An interpolation coefficient is calculated by a high-order interpolation method based on the oblique line information output from the coefficient calculation unit, and is within a range determined by a predetermined dimension equal to or higher than the second order of the output pixel output from the input image acquisition unit. An interpolation unit that interpolates the pixel value of the output pixel based on the pixel value of each input pixel to which it belongs and the calculated interpolation coefficient;
An output unit that outputs the position coordinates of the output pixel output from the output pixel coordinate determination unit and the pixel value of the interpolated output pixel output from the interpolation unit. ,
Image processing device. The image processing apparatus according to any one of claims 1 to 6,
The input image storage unit includes a line memory that holds pixel values of input pixels corresponding to the number of lines necessary to interpolate the pixel values of the output pixels.
Image processing device.

Images