JP4470775B2 - Pixel interpolation device and pixel interpolation program - Google Patents

Description

  The present invention relates to a pixel interpolation device that compensates pixels between pixels of an image to increase the resolution of the image, and a pixel interpolation program that causes a computer to operate as such a pixel interpolation device.

  As is well known, when an image is enlarged by increasing the resolution of the image, it is necessary to supplement some new pixels between the original pixels constituting the original image. When a new pixel is compensated between the original pixels, there is a method in which the pixel value of the new pixel is a copy of the pixel value of the original pixel at the closest coordinates. Will look like a mosaic when viewed with a magnifying glass. Therefore, in order to make the enlarged image smooth, normally, the pixel value at each coordinate (this is called an interpolation point) defined as that of a new pixel in the gap between the original pixels constituting the original image is interpolated. A method of mathematically estimating from the pixel values of several original pixels around the point is adopted.

  As a method for mathematically estimating the pixel value at the interpolation point, a bicubic interpolation method (bicubic method) and a bilinear interpolation method (bilinear method) are well known. In any of the interpolation methods, basically, a weight corresponding to the distance from the original pixel to the interpolation point is assigned to each of the pixel values of several original pixels arranged around one interpolation point. The sum of the weighted pixel values is used as the pixel value of the interpolation point.

  Specifically, the bicubic interpolation method is obtained by substituting a distance value into a variable z of a cubic expression ignoring the fourth and subsequent terms in the Taylor expansion equation of the interpolation function (sinπz) / πz. An operation for calculating a weight for a pixel value of a pixel is performed in the vertical direction and the horizontal direction, and a value obtained by multiplying the vertical weight and the horizontal weight by each other is multiplied by the pixel value of the original pixel. The final weight. In this bicubic interpolation method, the weighting targets are 16 original pixels around the interpolation point, and for each of the 16 original pixels, the vertical calculation and horizontal calculation are performed. It is designed to perform the third order calculation of components. For this reason, according to this bicubic interpolation method, although calculation takes a very long time, a smooth image with high interpolation accuracy can be obtained.

  On the other hand, the bilinear interpolation method normalizes the vertical direction and the horizontal method distance of the rectangular area where the interpolation point is surrounded by four original pixels to 1 respectively, and the inverse distance from the original pixel to the interpolation point (the original distance). The value obtained by calculating the distance between the original pixel on the diagonal side of the pixel and its interpolation point in the vertical and horizontal directions and multiplying the vertical and horizontal weights together. Is the final weight multiplied by the pixel value of the original pixel. That is, this bilinear interpolation method is a method of calculating a so-called inverse distance weighted average at the interpolation point for the pixel values of four original pixels surrounding the interpolation point. As described above, in this bilinear interpolation method, the weighting target is four original pixels around the interpolation point, and the vertical component of each of the four original pixels is obtained. The primary calculation and the horizontal calculation are performed. For this reason, according to this bilinear interpolation method, although the interpolation accuracy is not so high, the calculation can be completed relatively quickly.

In addition, when a new pixel is supplemented between original pixels, the original pixel may not be arranged at a position adjacent to the boundary (edge) of the enlarged image depending on the way of taking the coordinates of the interpolation point in the original image. . In this case, there are some interpolation points between the original pixel and the boundary. Even if the pixel value is calculated using the above bilinear interpolation method for such an interpolation point, the interpolation point is used. Since the four original pixels are not arranged around, the pixel value cannot be calculated. Therefore, regarding the interpolation point adjacent between the original pixel and the boundary in this way, it is assumed that the original pixel having the same pixel value as that of the surrounding original pixel is outside the boundary. The inverse distance weighted average at the interpolation point for the pixel value of the pixel is calculated. Such processing is called boundary value processing.
Japanese Patent No. 3045134

  Incidentally, when an image is enlarged in an information processing apparatus such as a printer that does not have sufficient memory capacity, pixel interpolation processing may not be performed on image data of an original image at a time. In this case, the information processing apparatus may divide the image data of the original image into several parts and sequentially perform the above-described pixel interpolation processing on each of the divided image data.

  However, if pixel interpolation processing is performed separately for each of the divided image data, when the images based on the processed image data are combined as one enlarged image, the images are connected to the enlarged image. Eyes may appear.

  For example, an image composed of horizontal stripes and vertical stripes as shown in FIG. 18A is divided into two along the line EE, and the divided images are separately enlarged using bilinear interpolation, When the images after enlargement are combined, as shown in FIG. 18B, the boundary between the white stripes and the black stripes on the EE line in the vertical stripe pattern in the enlarged image after combination is light and shade at other boundaries. It will have a change of light and shade different from the change of the seam, and the joint will be conspicuous. The phenomenon in which the joint appears is generated when the image is divided at the boundary between the white stripe and the black stripe in the vertical stripe, and the boundary value processing described above is performed at the interpolation point adjacent to the division line. . Therefore, this phenomenon can occur not only in the checkered pattern as shown in FIG. 18A but also in the checkered pattern as shown in FIG.

  The present invention has been made in view of the problems of the prior art as described above, and the problem is that the image is divided at the boundary portion of the light and shade in an image having a sharp change in light and shade, and the divided image is divided. Even when the images are enlarged using a bilinear interpolation and then combined with each other, it is to make the joints of the images as inconspicuous as possible in the final enlarged image.

A pixel interpolation device invented to solve the above-described problem is an image data acquisition unit that acquires, as a processing target, image data of an original image composed of a large number of pixels arranged two-dimensionally, the image data acquisition A magnification acquisition unit that acquires an enlargement magnification for an original image of image data acquired by the unit, and the magnification acquisition unit acquires a coordinate value of an interpolation point in the coordinates of an output image determined by the enlargement magnification acquired by the magnification acquisition unit Using the inverse transformation magnification calculated based on the enlargement magnification, it is converted to the coordinate value in the coordinates of the original image, and the interpolation point is defined on the coordinates of the original image by shifting it to the left and the upper side by the half pixel of the original image Interpolation point definition unit to be used, and for each of all the interpolation points defined by the interpolation point definition unit, the vertical distance of the rectangular area defined by the four original pixels around the interpolation point And a reverse distance value generation unit for performing a process of calculating a vertical reverse distance value and a horizontal reverse distance value from the four original pixels to the interpolation point based on the horizontal distance and the horizontal distance, Out of all the interpolation points defined by the interpolation point definition unit, the pixel values of the surrounding four original pixels are outside the area occupied by the original pixels constituting the original image, and there is a predetermined density difference. For a certain interpolation point, a corrected reverse distance value corresponding to the reverse distance value calculated by the reverse distance value generation unit for the interpolation point from a storage device that stores the reverse distance value and the corrected reverse distance value in association with each other. The reverse distance value correction unit that replaces the reverse distance value with the corrected reverse distance value, and for each of all the interpolation points defined by the interpolation point definition unit, the reverse distance value generation unit for the interpolation point The reverse distance value calculated by or replaced by the reverse distance value correction unit A pixel value generation unit that performs processing for calculating a reverse distance weighted average of the interpolation point based on the corrected reverse distance value after the correction and the pixel values of the four original pixels around the interpolation point; An image data output unit that outputs image data including pixel values generated by the pixel value generation unit for each interpolation point as image data of an enlarged image is provided.

  In this way, when generating a pixel value for one interpolation point, the interpolation point is outside the area occupied by the original pixel constituting the original image, and the four original pixels around it are also located. If the pixel value has a certain density difference, the density of the enlarged image will change if the corrected inverse distance value is used instead of the inverse distance value obtained by a general bilinear interpolation method. The change in shading in the portion where the boundary between the image and the dividing line of the image overlaps is alleviated, and the joint of the image in the final enlarged image becomes less noticeable.

In addition, a pixel interpolation program invented to solve the above-described problem is an image data acquisition unit that acquires, as a processing target, image data of an original image composed of a large number of pixels arranged two-dimensionally. , A magnification acquisition means for acquiring an enlargement magnification for the original image of the image data acquired by the image data acquisition means, a coordinate value of an interpolation point in the coordinates of the output image determined by the enlargement magnification acquired by the magnification acquisition means , Using the inverse conversion magnification calculated based on the enlargement magnification acquired by the magnification acquisition means, the coordinate value in the coordinates of the original image is converted, and the interpolation point is shifted to the left and the upper side by half a pixel of the original image For each of the interpolation points defined by the interpolation point definition means defined on the coordinates of the original image and the interpolation point definition means, Based on the vertical distance and the horizontal distance of the rectangular area defined by the four original pixels around the interpolation point, the vertical reverse distance from the four original pixels to the interpolation point A reverse distance value generating means for performing a process of calculating a value of and a reverse distance value in the horizontal direction;
Of all the interpolation points defined by the interpolation point defining means, outside the area occupied by the original pixels constituting the original image, and a predetermined density difference between the pixel values of the four surrounding original pixels For a certain interpolation point, a corrected inverse corresponding to the inverse distance value calculated by the inverse distance value generating means for the interpolation point from the storage device storing the inverse distance value and the corrected inverse distance value in association with each other. A reverse distance value correcting means for reading out the distance value and replacing the reverse distance value with a corrected reverse distance value;
After all of the interpolation points defined by the interpolation point defining means are replaced by the inverse distance value calculated by the inverse distance value generating means for the interpolation point or by the inverse distance value correcting means. A pixel value generating means for performing a process of calculating a reverse distance weighted average of the interpolation point based on the corrected reverse distance value of the interpolation point and the pixel values of the four original pixels around the interpolation point; and the pixel value It is characterized by functioning as image data output means for outputting image data consisting of pixel values generated for each interpolation point by the generating means as image data of an enlarged image.

  Therefore, the pixel interpolation program can operate the computer as a device that functions in the same manner as the pixel interpolation device according to the present invention described above.

  For this reason, according to the present invention, an image is divided at the boundary portion of the light and shade in an image including a pattern with sharp changes in light and shade, and each of the divided images is enlarged using a bilinear interpolation method. Even if the images are combined with each other, the joints of the images become less noticeable in the final enlarged image.

  Hereinafter, a printing apparatus according to one embodiment for carrying out the present invention will be described with reference to the accompanying drawings.

  FIG. 1 is a configuration diagram of a printing apparatus 10 according to the present embodiment. The printing device 10 is a device that performs printing by receiving a print request from a host computer (not shown) or image data from an external device (not shown).

  Note that external devices (not shown) include a camera device and a scanner. Among these, the camera device is a device with a USB port that can acquire image data of a still image, and specifically, a digital still camera, a digital video camera, a mobile phone with a camera, and the like. The scanner is a device that reads an image from a document placed on a document table and converts it into image data. When connected to a scanner (not shown), the printing apparatus 10 functions as a unit that performs printing in a copying machine or a multifunction machine.

  The printing apparatus 10 includes an external I / F device 11, an operation display board 12, a paper feeding mechanism 13, a printing mechanism 14, a DRAM 15, a ROM 16, a CPU 17, an EEPROM 18, and an ASIC 19 as main components.

  The external I / F device 11 is a so-called communication interface port for receiving a print request from a host computer (not shown) or receiving image data from an external device (not shown). Specifically, the external I / F device 11 includes a parallel port for performing communication in accordance with the IEEE 1284 standard, a parallel port of Centronics specification, and Ethernet (trademark of Xerox Corporation, Intel Corporation, USA, Compaq Computer Corporation) specifications. A LAN board, a serial port for performing communication according to the IEEE 1394 standard, a USB port for performing communication according to the USB standard, and the like.

  The operation display panel 12 is a device for receiving various instructions from the user and displaying various screens, and includes a push button and a touch screen.

  The paper feed mechanism 13 is a mechanism that supplies one or more types of paper to the printing mechanism 14 and includes several slots in which paper cassettes are detachably mounted. The printing mechanism 14 is a mechanism that actually performs printing on the paper supplied from the paper feeding mechanism 13.

  A DRAM (Dynamic Random Access Memory) 15 serves as a temporary storage area for print requests from a host computer (not shown) or image data from an external device (not shown), and various processes are performed on the image data. It is a volatile memory that is used as a work area when being read.

  A ROM [Read Only Memory] 16 is a non-volatile memory that stores a program for controlling the printing apparatus 10 and font data used when generating image data based on print data requested to be printed.

  The CPU [Central Processing Unit] 17 is a control circuit that controls each part in an integrated manner in accordance with a program in the ROM 16. Specifically, the CPU 17 obtains an instruction about a process to be executed from the user through an operation on a button or a touch screen of the operation display panel 12 provided on the upper surface of the printing apparatus 10, not shown. A process for generating image data to be delivered to the printing mechanism 14 based on a print request from a host computer or image data from an external device (not shown), and the printing mechanism 14 is controlled so that an image is printed based on the image data. The print control process is performed.

  An EEPROM [Electrically Erasable and Programmable Read Only Memory] 18 is a so-called flash memory for holding various settings relating to printing even after the power is turned off.

  An ASIC [Application Specific Integrated Circuit] 19 is a semiconductor integrated circuit that controls exchange of data and instructions between the hardware 11 to 18, and is connected to the hardware 11 to 18. The ASIC 19 includes a video controller 191, an IO controller 192, and a memory controller 193 as interfaces between the hardware units 11 to 18.

  The video controller 191 is a circuit module for converting image data to be supplied to the printing mechanism 14 into an electrical signal in a form that can be processed by a scanning unit (not shown) in the printing mechanism 14.

  The IO controller 192 is a circuit module for controlling the external I / F device 11, the operation display panel 12, and the paper feed mechanism 13. The IO controller 192 performs a process of directly transferring a print request transmitted from a host computer (not shown) and image data transmitted from an external device (not shown) from the external I / F device 11 to the DRAM 15. It is also a circuit module to perform.

  The memory controller 193 is a circuit module having a function of controlling reading and writing of programs and data with respect to the DRAM 15, a function of controlling reading of programs and data from the ROM 16, and a function of controlling a controller (not shown) in the printing mechanism 14. is there.

  Since the storage capacity of the DRAM 15 is not sufficient in the printing apparatus 10 according to the present embodiment configured with the hardware as described above, the processing contents are appropriately constructed so that the storage capacity can be dealt with. Specifically, when a print request including print data created in a page description language is sent from a host computer (not shown), image data for one page is generated and then output to the video controller 191. Instead, a process of generating a part of the image data and then outputting the generated image data to the video controller 191 is sequentially performed for each part of the image data.

  Similarly, when a print request including image data is sent from a host computer (not shown), a part of the image data is received and output to the video controller 191 for each part of the image data. It is like that. The same applies when image data is sent from an external device (not shown).

  Note that the processing for generating image data from a print request and the processing for fetching image data sent from a host computer (not shown) or an external device (not shown) into the DRAM 15 are performed by the above-described image data acquisition means. It corresponds. The CPU 17, the external I / F device 11, and the IO controller 121 that execute these processes correspond to the above-described image data acquisition unit.

  By the way, an image based on image data for which a print request from a host computer (not shown) or an image based on image data from an external device (not shown) is too low in resolution and suitable for output to the video controller 191. There may not be. In this case, the CPU 17 reads a predetermined program from the ROM 16 and performs a process for increasing the resolution of the image based on the image data. FIG. 2 is a diagram showing the flow of this process.

  After the start of processing, the CPU 17 indicates information indicating the size of the image whose resolution is to be increased (hereinafter referred to as an original image) (here, the number of horizontal pixels is defined as Ih and the number of vertical pixels is defined as Iv). Is acquired from the header of the image data of the original image (S101).

  Subsequently, the CPU 17 indicates information indicating an image size (here, the horizontal pixel number is defined as Oh and the vertical pixel number is defined as Ov) set to be optimal for output to the video controller 191. Is obtained from the EEPROM 18 or the like (S102).

  After that, the CPU 17 calculates the vertical and horizontal magnifications based on the acquired size information, and calculates the inverse conversion magnification based on the calculated magnifications (S103).

  Here, the inverse conversion magnification will be described. In the description here, the coordinates of each pixel of the original image are expressed as (x, y), and the coordinates of each pixel of the enlarged output image are expressed as (X, Y). In addition, the X and Y coordinate values of the output image both take an integer having a minimum unit of 1 (meaning that the distance between adjacent pixels is 1).

  When enlarging the original image, the corresponding x coordinate value in the original image is usually obtained by multiplying the X coordinate value and Y coordinate value of each pixel of the output image by the reciprocal of the enlargement magnification (Oh / Ih, Ov / Iv), respectively. And the y coordinate value are calculated, and when both the x coordinate value and the y coordinate value are integers, the pixel value of the coordinate (x, y) in the original image is the coordinate (X, Y) of the output image. If either or both of these x-coordinate values and y-coordinate values are decimal numbers, the coordinates (x, y) are used as interpolation points, and the pixel at the coordinates (x, y). The value is estimated as described later, and the estimated pixel value is determined as the pixel value of the coordinates (X, Y) of the output image.

  However, if the minimum unit of the X and Y coordinate values of the output image is 1, the x and y coordinate values may be decimals with a value smaller than the decimal point. Then, when calculating the weight by the bilinear interpolation method based on these x and y coordinate values, the processing is burdened. For this reason, in this embodiment, a number obtained by multiplying the reciprocal of the magnification described above by a numerical value of 256 is defined as the inverse conversion magnification.

  This numerical value of 256 substantially functions as expressing the length between adjacent pixels in the original image as 256 instead of 1. In this way, when the length of one section in the original image is expressed as 256, when the enlargement magnification is 16 or less, the coordinates (x, y) of the original image corresponding to the coordinates (X, Y) of the output image The value above the decimal point in the coordinate value of can be increased, and the value after the decimal point can be rounded down for calculation. However, if the x and y coordinate values in the original image are expressed as a cumulative value of 256, the x and y coordinate values become extremely large. Therefore, in the following processing, whenever the x and y coordinate values reach 256, the values are reset to 0, and the value 256 is limited to represent only the length of one section between adjacent pixels. (Not used to represent a length of two or more intervals).

  Steps S101 to S103 correspond to the above-described magnification acquisition unit, and the CPU 17 that executes these steps S101 to S103 corresponds to the above-described magnification acquisition unit.

  After calculating the inverse transformation magnification ((Ih / Oh) × 256, (Iv / Ov) × 256) described above (S103), the CPU 17 divides the image data of the original image into predetermined sizes. A storage area for storing one of them (hereinafter referred to as divided original image data) is secured in the DRAM 15 (S104), and the divided original image data is taken into the storage area (S105).

  Subsequently, the CPU 17 stores a storage area for storing image data (hereinafter referred to as divided output image data) created as described later based on the divided original image data taken into the storage area in the DRAM 15. Is secured (S106).

  Thereafter, the CPU 17 performs pixel interpolation processing described later on the divided original image data captured in the DRAM 15 (S107).

  When the execution of the pixel interpolation process is completed, the CPU 17 executes a process of outputting the completed divided output image data to the video controller 191 (S108), and determines whether there is the next divided original image data (S109). . If there is next divided original image data (S109; NO), the CPU 17 again secures the storage area of the divided original image data (S104), fetches the divided original image data (S105), and the divided output image. A data storage area is secured (S106), and pixel interpolation processing described later is executed for the divided original image data (S107). Then, by repeating such processing, when there is no divided original image data to be processed (S109; YES), the CPU 17 ends the processing according to FIG.

  Step S108 corresponds to the above-described image data output means, and the CPU 17 that executes this step S108 corresponds to the above-described image data output unit.

  Next, pixel interpolation processing will be described. 3 to 8 are diagrams showing the flow of pixel interpolation processing.

  First, as shown in FIG. 3, the CPU 17 assigns zero to the variable X and assigns zero to the variable Y (S201).

  Subsequently, the CPU 17 calculates the y coordinate value in the divided original image corresponding to the coordinate value Y in the divided output image (S202). Specifically, the CPU 17 multiplies the coordinate value X by the horizontal reverse conversion magnification ((Ih / Oh) × 256), and further subtracts 128 from the result to obtain the y coordinate value.

  The value of 128 is a value obtained by multiplying 256 by 0.5, that is, a half (half pixel) of one section between adjacent original pixels. That is, by subtracting 128 from the coordinate value after the inverse transformation, the coordinates of the original pixel referred to the coordinates (X, Y) of the output image are shifted to the negative side (left side) by half a pixel.

  Here, the reason for shifting to the negative side by half a pixel will be described. Here, in order to facilitate understanding, the coordinate value of the original image is not multiplied by 256. For example, assuming that the original image is composed of 3 × 3 pixels as shown in FIG. 9A, when the original image is enlarged three times vertically and three times horizontally, as shown in FIG. 9B. Suppose that the coordinates in the original image corresponding to the coordinates (0,0) of the output image are defined as (0,0) (a black circle is placed at the coordinates (0,0) in FIG. 9B). That means this provision). In this case, attention is paid to a pixel A at coordinates (0, 1) and a pixel B at coordinates (2, 1) in the original image. In the bilinear interpolation method, the interpolation point indicated by a white circle in FIG. 9B is obtained by multiplying the pixel value of the original pixel by a weight inversely proportional to the distance from the original pixel (black circle) on the original image. The value is obtained as part of its own pixel value. Therefore, conversely, for example, the pixel B has its own pixel value at the interpolation point in the upper right, lower right, upper left, and lower left regions (interpolation point surrounded by a one-dot chain line in FIG. 9B). Will be distributed. On the other hand, the pixel A distributes its own pixel value only to the interpolation points (interpolation points surrounded by a broken line in FIG. 9B) in the upper right and lower right regions of the pixel A. That is, the pixel A uses only half of its own pixel value for accumulating the pixel value at the interpolation point. For this reason, in this embodiment, as shown in FIG. 9C, the coordinates of the original pixel corresponding to the coordinates (X, Y) of the output image are shifted by the negative side (left side and upper side) by half a pixel. Also for the pixel A, the upper left and lower left regions having the interpolation points to which the pixel values are to be distributed are given. As a result, the coordinates in the original image corresponding to the coordinates (0, 0) of the output image are (−0.5, −0.5), and the interpolation points on the output image are x, The y coordinate value does not necessarily have a corresponding coordinate on a coordinate whose integers are both integers.

  Subsequently, the CPU 17 calculates an x coordinate value corresponding to the coordinate value Y in the divided output image (S203). Specifically, the CPU 17 multiplies the coordinate value X by the vertical reverse conversion magnification ((Iv / Ov) × 256), and further subtracts 128 from the result to obtain an x coordinate value.

  Note that these steps S202 and S203 correspond to the above-described interpolation point defining means, and the CPU 17 executing these steps S202 and S203 corresponds to the above-described interpolation point defining unit.

  Subsequently, the CPU 17 is a subroutine for specifying the coordinates of the original pixel to be referred to when the pixel value of the coordinate (x, y) corresponding to the coordinate (X, Y) in the original image is calculated by the bilinear interpolation method. (S204). FIG. 10 shows the flow of this subroutine.

  After starting this subroutine, the CPU 17 specifies four coordinates surrounding the coordinate (x, y) among the coordinates having integer (including negative) coordinate values on the original image, and then, among these four coordinates. The x coordinate values of the coordinates at the upper left and lower left of the coordinates (x, y) are substituted into the variable xl indicating the left end position of the processing target range, and the x coordinates of the coordinates at the upper right and lower right of the coordinates (x, y) The value is substituted into the variable xr indicating the right end position of the processing target range, the y coordinate value of the coordinates at the upper left and upper right of the coordinates (x, y) is substituted into the variable yt indicating the upper end position of the processing target range, and the coordinates ( The y-coordinate values of the coordinates at the lower left and lower right of x, y) are substituted into a variable yb indicating the lower end position of the processing target range (S301).

  Subsequently, the CPU 17 determines whether or not the variable xl indicating the left end position of the processing target range is smaller than 0 (S302). When the variable xl is smaller than 0 (S302; YES), the CPU 17 increases the substitution value of the variable xl indicating the left end position of the processing target range and the substitution value of the variable xr indicating the right end position of the processing target range by one. (S303). On the other hand, if the variable xl is not smaller than 0 (S302; NO), the CPU 17 determines whether or not the variable xr indicating the right end position of the processing target range is larger than the maximum value xmax in the x direction of the divided original image. (S304). When the variable xr indicating the processing target range right end position is larger than the maximum value xmax (S304; YES), the CPU 17 assigns the variable xl indicating the processing target range left end position and the variable indicating the processing target range right end position. If the variable xr indicating the right end position of the processing target range is not larger than the maximum value xmax (S304; NO), the substituting value of the variable xl indicating the left end position of the processing target range is decreased by 1 (S305). And the substitution value of the variable xr indicating the right end position of the processing target range are left as they are.

  Subsequently, the CPU 17 determines whether or not the variable yt indicating the processing target range upper end position is smaller than 0 (S306). When the variable yt is smaller than 0 (S306; YES), the CPU 17 increases the substitution value of the variable yt indicating the processing target range upper end position and the substitution value of the variable yb indicating the processing target range lower end position by one. (S307). On the other hand, if the variable yt is not smaller than 0 (S306; NO), the CPU 17 determines whether or not the variable yb indicating the lower end position of the processing target range is larger than the maximum value ymax in the y direction of the divided original image. (S308). When the variable yb indicating the processing target range lower end position is greater than the maximum value xmax (S308; YES), the CPU 17 assigns the variable yt indicating the processing target range upper end position and the variable indicating the processing target range lower end position. When the variable yb indicating the lower end position of the processing target range is not larger than the maximum value ymax (S308; NO), the substituting value of the variable yt indicating the upper end position of the processing target range is decremented by 1 (S309). And the substitution value of the variable yb indicating the lower end position of the processing target range are left as they are.

  As described above, the CPU 17 specifies substitution values for the variables xl, xr, yt, and yb, and ends the processing according to FIG. Note that the reference original pixel specifying subroutine shown in FIG. 10 corresponds to the above-described reverse distance generation means, and the CPU 17 that executes this subroutine corresponds to the above-described reverse distance generation unit.

  After the process according to FIG. 10 is completed, the CPU 17 performs a table determination process for determining which of the three tables recorded in the ROM 16 is used for the subsequent processes as shown in FIG. S205 to S209).

  Note that each of the three tables recorded in the ROM 16 has a distance value obtained by subtracting the xl coordinate value from the x coordinate value (this is a distance value obtained by subtracting the xr coordinate value from the x coordinate value is 256. And the distance value obtained by subtracting the yt coordinate value from the y coordinate value (this is the distance value obtained by subtracting the yb coordinate value from the y coordinate value from 256. This is a table used when replacing a subtracted value (equal to a reverse distance value) with another value. As will be described later, these three tables are selectively used according to the difference in the position of the coordinates (x, y) in the original image.

  Specifically, the CPU 17 first determines whether one of the x-coordinate value and the y-coordinate value is a value smaller than 0 (S205).

  If either the x-coordinate value or the y-coordinate value is smaller than 0 (S205; YES), the CPU 17 selects the left edge or upper edge of the original image from the three tables described later recorded in the ROM 16. It is determined that the upper left edge table for use with respect to the interpolation point at is used for the distance value determination process described later (S207). Note that the left edge of the original image refers to an area below the minimum value 0 in the x direction of the original image. In the example of FIG. 9C, coordinates (0,0) to coordinates (0,8) This is an area with coordinates (1, 0) to (1, 8). Further, the upper edge of the original image refers to an area below the minimum value 0 in the y direction of the original image. In the example of FIG. 9C, coordinates (0,0) to coordinates (8,0) This is an area having coordinates (0, 1) to (8, 1).

  On the other hand, if neither the x coordinate value nor the y coordinate value is smaller than 0 (S205; NO), the CPU 17 determines whether the x coordinate value is larger than xmax or the y coordinate value is larger than ymax. It is determined whether or not (S206).

  If the x-coordinate value is larger than xmax or the y-coordinate value is larger than ymax (S206; YES), the CPU 17 selects the original image from the three tables described later recorded in the ROM 16. It is determined that the table for the lower right edge to be used for the interpolation point at the right edge or the lower edge is used for the distance value determination process described later (S208). The right edge of the original image means a region exceeding the maximum value xmax in the x direction of the original image. In the example of FIG. 9C, the coordinates (8, 2) to the coordinates (8, 8) are It is a certain area. Further, the lower edge of the original image means a region exceeding the maximum value ymax in the y direction of the original image. In the example of FIG. 9C, the coordinates (2, 8) to the coordinates (8, 8) are It is a certain area.

  On the other hand, when the x-coordinate value is not larger than xmax and the y-coordinate value is not larger than ymax (S206; NO), the CPU 17 selects the following three tables recorded in the ROM 16: Then, it is determined that an internal table for use with the interpolation points in the original image is used for the distance replacement process described later (S209). The inside of the original image is a region in the range from the minimum value 0 to the maximum value xmax in the x direction of the original image, and in the range from the minimum value 0 to the maximum value ymax in the y direction of the original image. Say. In the example of FIG. 9C, it refers to an area having coordinates (2, 2) to (7, 7).

  After determining the table used for the distance value determination process described later as described above, the CPU 17 coordinates (xl, yt), (xr, yt), (xl, yb) as shown in FIG. , (Xr, yb), the pixel values of the respective RGB color components are read from the storage area of the divided original pixel data (S210).

  Subsequently, the CPU 17 records the x-direction edge flag information and the y-direction edge flag information in the DRAM 15 in a state indicating OFF (S211). The edge flag information is flag information for storing determination results in subsequent determination processes (S212 to S235). The determination process will be described in detail later, but will be briefly described. As shown in FIG. 11, the density difference of pixel values in a minute image composed of four original pixels around the coordinates (x, y) is shown. A process for determining whether or not any of the patterns is present.

  After the edge flag information in the off state is recorded in the DRAM 15 as described above (S211), the CPU 17 determines the sum of the R pixel value and G pixel value of the original pixel at the upper left coordinates (xl, yt) and the upper right coordinates (xr , Yt), the absolute value of the difference is calculated for the sum of the R pixel value and G pixel value of the original pixel (S212).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S213). If the absolute value exceeds the predetermined threshold (S213; YES), the CPU 17 switches the x-direction edge flag information from the off state to the on state (S214), and the absolute value exceeds the predetermined threshold. If not (S213; NO), the x-direction edge flag information remains in the OFF state.

  Subsequently, the CPU 17 sums the R pixel value and the G pixel value of the original pixel at the upper left coordinate (xl, yt) and the sum of the R pixel value and the G pixel value of the original pixel at the lower left coordinate (xl, yb). The absolute value of the difference is calculated with respect to (S215).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S216). If the absolute value exceeds the predetermined threshold (S216; YES), the CPU 17 switches the y-direction edge flag information from the off state to the on state (S217), and the absolute value exceeds the predetermined threshold. If not (S216; NO), the y-direction edge flag information remains in the OFF state.

  Next, as shown in FIG. 5, the CPU 17 sums the R pixel value and G pixel value of the original pixel at the lower left coordinate (xl, yb) and the original pixel at the lower right coordinate (xr, yb). The absolute value of the difference is calculated for the sum of the R pixel value and the G pixel value (S218).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S219). When the absolute value exceeds the predetermined threshold (S219; YES), the CPU 17 switches the state of the x-direction edge flag information to the on state (S220), and the absolute value does not exceed the predetermined threshold. If this happens (S219; NO), the state of the x-direction edge flag information is not changed.

  Subsequently, the CPU 17 calculates the sum of the R pixel value and the G pixel value of the original pixel located at the upper right coordinate (xr, yt) and the R pixel value and the G pixel value of the original pixel located at the lower right coordinate (xr, yb). The absolute value of the difference is calculated for the sum (S221).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S222). If the absolute value exceeds the predetermined threshold (S222; YES), the CPU 17 switches the state of the y-direction edge flag information to the on state (S223), and the absolute value does not exceed the predetermined threshold. If this happens (S222; NO), the state of the y-direction edge flag information is not changed.

  Subsequently, the CPU 17 sums the G pixel value and the B pixel value of the original pixel at the upper left coordinate (xl, yt) and the sum of the G pixel value and the B pixel value of the original pixel at the upper right coordinate (xr, yt). Then, the absolute value of the difference is calculated (S224).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S225). When the absolute value exceeds the predetermined threshold (S225; YES), the CPU 17 switches the state of the x-direction edge flag information to the on state (S226), and the absolute value does not exceed the predetermined threshold. If it is detected (S225; NO), the state of the x-direction edge flag information is not changed.

  Subsequently, as shown in FIG. 6, the CPU 17 sums the G pixel value and the B pixel value of the original pixel at the upper left coordinate (xl, yt) and the G of the original pixel at the lower left coordinate (xl, yb). The absolute value of the difference is calculated for the sum of the pixel value and the B pixel value (S227).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S228). When the absolute value exceeds the predetermined threshold (S228; YES), the CPU 17 switches the state of the y-direction edge flag information to the on state (S229), and the absolute value does not exceed the predetermined threshold. If this is the case (S228; NO), the state of the y-direction edge flag information is not changed.

  Subsequently, the CPU 17 calculates the sum of the G pixel value and the B pixel value of the original pixel at the lower left coordinate (xl, yb) and the G pixel value and the B pixel value of the original pixel at the lower right coordinate (xr, yb). The absolute value of the difference is calculated for the sum (S230).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S231). If the absolute value exceeds the predetermined threshold (S231; YES), the CPU 17 switches the state of the x-direction edge flag information to the on state (S232), and the absolute value does not exceed the predetermined threshold. If it is detected (S231; NO), the state of the x-direction edge flag information is not changed.

  Subsequently, the CPU 17 calculates the sum of the G pixel value and the B pixel value of the original pixel at the upper right coordinate (xr, yt) and the G pixel value and the B pixel value of the original pixel at the lower right coordinate (xr, yb). The absolute value of the difference is calculated for the sum (S233).

  Thereafter, the CPU 17 determines whether or not the absolute value exceeds a predetermined threshold (S234). If the absolute value exceeds the predetermined threshold (S234; YES), the CPU 17 switches the state of the y-direction edge flag information to the on state (S235), and the absolute value does not exceed the predetermined threshold. If this occurs (S234; NO), the state of the y-direction edge flag information is not changed.

  After completing the determination processing (S212 to S235) described above, the CPU 17 determines the distance value from the coordinates (x, y) to the four original pixels around it as shown in FIG. Distance determination processing is performed (S236 to S243).

  Specifically, the CPU 17 first determines whether the x-direction edge flag information is in an on state or an off state (S236).

  If the x-direction edge flag information is in the off state (S236; YES), the CPU 17 subtracts the xl coordinate value from the x coordinate value, and divides the result by 256 to obtain the value p as the distance value. (S237).

  On the other hand, when the x-direction edge flag information is in the ON state (S236; NO), the CPU 17 among the three tables recorded in the ROM 16 is determined to be used at this time (S205 to S205). Referring to S209), the value corresponding to the value obtained by subtracting the xl coordinate value from the x coordinate value is read (S238), and the value obtained by dividing the read value by 256 is set as the distance value in the variable p. Substitute (S239).

  Subsequently, the CPU 17 determines whether the y-direction edge flag information is on or off (S240).

  If the y-direction edge flag information is in the off state (S240; YES), the CPU 17 subtracts the yt coordinate value from the y coordinate value, and divides the result by 256 to obtain a variable q (S241).

  On the other hand, when the y-direction edge flag information is in the ON state (S240; NO), the CPU 17 among the three tables recorded in the ROM 16 is determined to be used at this time (S205 to S205). Referring to S209), the value corresponding to the value obtained by subtracting the yt coordinate value from the y coordinate value is read (S242), and the value obtained by dividing the read value by 256 is set as a distance value to the variable q. Substitute (S243).

  Note that steps S237 and S241 correspond to the above-described reverse distance value generation unit, and steps S212 to S235, S238, S239, S242, and S243 correspond to the above-described reverse distance value correction unit. The CPU 17 that executes steps S237 and S241 corresponds to the above-described reverse distance value generation unit, and the CPU 17 that executes steps S212 to S235, S238, S239, S242, and S243 includes the above-described reverse distance value correction unit. It corresponds to.

  After determining the substituted values of the variables p and q as described above, the CPU 17 calculates the pixel value of the coordinates (x, y) for each component of RGB using these variables p and q (S244). . The calculation formula is a calculation formula used in a general bilinear interpolation method: “pixel value of interpolation point = (1−p) (1−9) × (pixel value of upper left original pixel) + p (1−q)” × (pixel value of the upper right original pixel) + (1-p) q × (pixel value of the lower left original pixel) + pq × (pixel value of the lower right original pixel) ”.

  The value p is the reverse distance value in the x direction between the upper right original pixel and the lower right original pixel, and the value (1-p) is the reverse distance value in the x direction between the upper left original pixel and the lower left original pixel. . The q value is the reverse distance value in the y direction between the lower left original pixel and the lower right original pixel, and the value (1-q) is the reverse distance value in the y direction between the upper left original pixel and the upper right original pixel. .

  Then, the CPU 17 stores the calculated RGB pixel values at the position of the coordinates (x, y) in the area secured in the DRAM 15 as the storage area for the divided output image data (S245).

  Note that steps S244 and S245 correspond to the above-described pixel value generation means, and the CPU 17 that executes steps S244 and S245 corresponds to the above-described pixel value generation unit.

  Thereafter, as shown in FIG. 8, the CPU 17 determines whether or not the variable X has reached the maximum value Xmax in the X direction of the divided output image (S246). At this time, if the variable X has not reached the maximum value Xmax (S246; NO), the CPU 17 increases the substitution value of the variable X by 1 (S247), and executes the processes of steps S203 to S245 again.

  When the variable X reaches the maximum value Xmax by repeating the processing of steps S203 to S247 (S246; YES), the CPU 17 substitutes 0 for the variable X (S248), and the variable Y is divided and output. It is determined whether or not the maximum value Ymax in the Y direction of the image has been reached (S249). At this time, if the variable Y has not reached the maximum value Ymax (S249; NO), the CPU 17 increases the substitution value of the variable Y by 1 (S250), and executes the processes of steps S202 to S248 again.

  When the variable Y reaches the maximum value Ymax by repeatedly performing the processes of steps S202 to S250 (S249; YES), the CPU 17 ends the processes according to FIGS.

  The printing apparatus 10 according to the present embodiment is configured to perform the processing as described above, and thus has the operations and effects described below.

  As described above, in the printing apparatus 10 of the present embodiment, when the divided original image data is taken into the DRAM 15, the pixel values of the pixels of the divided output image are sequentially set based on the pixel values of the pixels of the divided original image. It is generated (S201 to S250).

  At this time, if the divided original image includes a portion where the change in shading is gentle like a landscape image, the printing apparatus 10 does not have any boundary portion where the shading changes sharply in that portion. It is also determined that there is no density difference between the four original pixels that are referred to for creating the pixel value of the interpolation point (S213; NO, S216; NO, S219; NO, S222; NO, S225; NO, S228; NO, S231; NO, and S234; NO). For this reason, in this case, since both the x-direction edge flag information and the y-direction edge flag information remain in the off state, the pixel values of the four original pixels are used to create the pixel value of one interpolation point. As the weight multiplied by, an inverse distance value inversely proportional to the distance is used as in the general bilinear interpolation method (S237, S241).

  On the other hand, when the divided original image includes a portion such as a solid image such as a checkered pattern or a cartoon, since there is a boundary portion where the shading changes sharply in the portion, the printing apparatus 10 is any of the portions. It is determined that there is a density difference between the four original pixels referred to for creating the pixel value of the interpolation point (S213; YES, S216; YES, S219; YES, S222; YES, S225). YES, S228; YES, S231; YES, or S234; YES). For this reason, since either or both of the x-direction edge flag information and the y-direction edge flag information are turned on for the interpolation point determined in this manner, the pixel value of the interpolation point is created. The weight multiplied by the pixel values of the four original pixels is not the inverse distance value inversely proportional to the distance as in the general bilinear interpolation method, but the corrected inverse distance value is used. (S238, S239, S242, S243).

  12 and 13 show graphs obtained by plotting the values of the table used for correcting the reverse distance value on the coordinates. In these graphs, the horizontal axis represents the reverse distance value before correction, and the vertical axis represents the reverse distance value after correction.

  Among these, the broken line graph in FIG. 12 shows the relationship between the reverse distance value before the correction and the reverse distance value after the correction when the interpolation points in the original image are corrected. According to the internal table indicated by the broken line in FIG. 12, when the distance value between the original pixel and the interpolation point is less than about 80, the reverse distance value is corrected to be a smaller reverse distance value. When the reverse distance value exceeds about 175, the reverse distance value is corrected so as to become a larger reverse distance value. Note that, for comparison, the solid line graph in FIG. 12 shows the relationship between the reverse distance values when correction using the table is not performed (the reverse distance value before correction when the correction procedure is taken and the correction after correction). The relationship with the inverse distance value).

  When the inverse distance value for calculating the weight with respect to the pixel value of the original pixel is corrected in this way for the interpolation point inside the original image, the boundary portion (edge) in which the shading changes in the enlarged image is obtained. I won't make much money. For example, assuming that a checkered pattern as shown in FIG. 14A exists inside the original image, when the printing apparatus 10 of the present embodiment enlarges this original image, the enlarged image is shown in FIG. ). As is clear by comparing the image of FIG. 14B with the enlarged image of FIG. 16B showing the result of enlargement by a general bilinear interpolation method, the enlarged image of FIG. The boundary between the white lattice and the black lattice is slightly emphasized.

  FIG. 15 is a graph showing the pixel value distribution on the C line of the enlarged image shown in FIG. The pixel value distribution shown in FIG. 15 is compared with the pixel value distribution shown in FIG. 17 as the pixel value distribution on the D line of the enlarged image by the general bilinear interpolation method shown in FIG. According to the printing apparatus 10 of the embodiment, the pixel values of the pixels at and near the boundary between the white lattice and the black lattice are distributed in a sinusoidal shape. From this graph, it is clear that the degree of blur appearing at the boundary between the white lattice and the black lattice is reduced.

  The solid line graph in FIG. 13 shows the relationship between the reverse distance value before correction and the reverse distance value after correction when correction is performed on the interpolation point at the left edge or the upper edge of the original image. According to the upper left edge table shown in the solid line graph of FIG. 13, when the reverse distance value between the original pixel and the interpolation point is less than about 50, the reverse distance value is a larger reverse distance value. Even when the reverse distance value exceeds about 175, the reverse distance value is corrected so as to be a larger reverse distance value.

  Furthermore, the broken line graph of FIG. 13 shows the relationship between the reverse distance value before correction and the reverse distance value after correction when correction is performed for the interpolation point at the right edge or the lower edge of the original image. According to the table for the lower right edge indicated by the broken line in FIG. 13, when the reverse distance value between the original pixel and the interpolation point is less than about 80, the reverse distance value is smaller than that. Even when the reverse distance value exceeds about 210, the reverse distance value is corrected so as to be a reverse distance value smaller than that.

  For the interpolation points at the left and right edges and the top and bottom edges of the original image, when the inverse distance value for calculating the weight for the pixel value of the original pixel is corrected as described above, in an image including a picture with sharp changes in shading Even when the image is divided at the shading boundary portion, and the divided images are separately enlarged and then combined with each other, the joints of the images are not so noticeable in the final enlarged image. It becomes like this. For example, as shown in FIG. 18A, an original image composed of a horizontal stripe pattern and a vertical stripe pattern is divided along the EE line (vertical line), and the divided images are respectively general bilinear. When separately enlarged by the interpolation method and then combined with each other, as shown in FIG. 18 (b) and FIG. 19 (a) which is an enlarged view thereof, E in the vertical stripe pattern in the enlarged image after the combination is obtained. The boundary between the white stripe and the black stripe on the −E line has a light / dark change different from the light / dark change at the other boundary, and the joint becomes conspicuous. However, in the enlarged image obtained by performing the same processing (division, separate enlargement, and combination) on the same original image by the printing apparatus 10 according to the present embodiment, as shown in FIG. The region of the gray portion of the white stripe on the left side of the −E line is widened, and the darkness of the black stripe on the right side of the EE line is slightly lighter.

  FIG. 20 is a graph showing the pixel value distribution on the F line of the enlarged image shown in FIG. In FIG. 20, a solid line graph and a broken line graph are shown. The solid line graph is an enlarged image obtained by using a general bilinear interpolation method (that is, an enlarged image of FIG. 19A). ), And the broken line graph shows the pixel value distribution of the enlarged image (that is, the enlarged image of FIG. 19B) obtained by the printing apparatus 10 of the present embodiment. As shown in FIG. 20, according to the printing apparatus 10 of the present embodiment, when the white stripe is at the right edge of the original image, the brightness of the white stripe is reduced and the black stripe is at the left edge of the original image. In some cases, the darkness of the black stripe is reduced. Conversely, although not shown, when the black stripe is at the right edge of the original image, the darkness of the black stripe is reduced, and when the white stripe is at the right edge of the original image, the brightness of the white stripe is reduced. Reduced. In other words, when the upper left edge table and the lower right edge table are used, the change in shading of the portion in the vicinity of the dividing line in the images on both sides sandwiching the dividing line becomes gradual, and the joint by the dividing line becomes less conspicuous.

  The correction using the table for the upper left edge and the table for the lower right edge acts to loosen the change in shading at the joint, whereas the correction using the internal table described above is It works to make the change in light and dark stand out. Accordingly, the correction using the upper left edge table and the lower right edge table is opposite to the correction using the internal table.

Configuration diagram of the printing apparatus of the present embodiment The figure which shows the flow of the process which raises the resolution of the image based on image data Diagram showing the flow of pixel interpolation processing Diagram showing the flow of pixel interpolation processing Diagram showing the flow of pixel interpolation processing Diagram showing the flow of pixel interpolation processing Diagram showing the flow of pixel interpolation processing Diagram showing the flow of pixel interpolation processing Explanatory diagram of the reason for shifting the half-pixel shift after obtaining the coordinates of the interpolation point on the original image with the inverse transform magnification The figure which shows the flow of a reference original pixel specific process The figure which shows the pattern of the density difference of four original pixels The figure which shows the graph which plots the value of the table for internal use used when correcting the distance value about the interpolation point in the inside of an original image on a coordinate The figure which shows the graph formed by plotting the value of the table for upper left edges and the table for lower right edges used on correcting the distance value about the interpolation point in the upper left edge of an original image, and a lower right edge on a coordinate, respectively. The figure which shows the state before enlarging the image of a checkered pattern by the printing apparatus of this embodiment, and after enlarging. The figure which shows the graph of the pixel value distribution on the C line of the image of the enlarged checkered pattern The figure which shows the state before and after enlarging the checkerboard pattern image using a general bilinear interpolation method The figure which shows the graph of the pixel value distribution on D line of the checkered pattern image expanded using the general bilinear interpolation method The figure which shows the state before and after enlarging the image which consists of a horizontal stripe and a vertical stripe with the printing apparatus of this embodiment. Enlarged view of the boundary between white stripes and black stripes on the EE line when using a general bilinear interpolation method and when using the printing apparatus of the present embodiment The figure which shows the graph of the pixel value distribution on the F line of the enlarged image which consists of a vertical stripe and a horizontal stripe The figure which shows the state when the checkered pattern image is divided and separately enlarged and combined

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Printing apparatus 11 External I / F apparatus 12 Operation display panel 13 Paper feed mechanism 14 Printing mechanism 15 DRAM
16 ROM
17 CPU
18 EEPROM
19 ASIC

Claims (3)

An image data acquisition unit for acquiring, as a processing target, image data of an original image composed of a number of pixels arranged two-dimensionally;
A magnification acquisition unit for acquiring an enlargement magnification for the original image of the image data acquired by the image data acquisition unit;
The coordinates in the coordinates of the original image using the inverse transform magnification calculated based on the magnification of the magnification acquired by the magnification acquisition unit, the coordinate value of the interpolation point in the coordinates of the output image determined by the magnification obtained by the magnification acquisition unit An interpolation point defining unit that defines an interpolation point on the coordinates of the original image by converting it into a value and shifting it to the left and upper sides by half a pixel of the original image ;
For each of all the interpolation points defined by the interpolation point definition unit, based on the vertical distance and the horizontal distance of the rectangular area defined by the four original pixels around the interpolation point, A reverse distance value generation unit for performing processing for calculating a vertical reverse distance value and a horizontal reverse distance value from the four original pixels to the interpolation point;
Out of all the interpolation points defined by the interpolation point definition unit, the pixel values of the surrounding four original pixels are outside the area occupied by the original pixels constituting the original image, and there is a predetermined density difference. For a certain interpolation point, a corrected reverse distance value corresponding to the reverse distance value calculated by the reverse distance value generation unit for the interpolation point from a storage device that stores the reverse distance value and the corrected reverse distance value in association with each other. , And a reverse distance value correction unit that replaces the reverse distance value with a corrected reverse distance value,
For each of all the interpolation points defined by the interpolation point definition unit, the reverse distance value calculated by the reverse distance value generation unit for the interpolation point or the correction after being replaced by the reverse distance value correction unit A pixel value generation unit that performs a process of calculating a reverse distance weighted average of the interpolation point based on the reverse distance value and the pixel values of the four original pixels around the interpolation point; and
A pixel interpolation device comprising: an image data output unit that outputs image data including pixel values generated by the pixel value generation unit for each interpolation point as image data of an enlarged image. For the interpolation point outside the area occupied by the original pixel constituting the original image, the pixel value generation unit uses the interpolation point as the pixel value of the original pixel for calculating the inverse distance weighted average of the interpolation point. 2. The pixel interpolating apparatus according to claim 1, wherein a pixel value of an original pixel inside the pixel and a pixel value of an original pixel inside the pixel are further used. Computer
Image data acquisition means for acquiring, as a processing target, image data of an original image composed of a plurality of pixels arranged two-dimensionally;
A magnification acquisition means for acquiring an enlargement magnification for the original image of the image data acquired by the image data acquisition means;
The coordinate value of the interpolation point in the coordinates of the output image determined by the magnification obtained by the magnification obtaining unit is calculated using the inverse transformation magnification calculated based on the magnification obtained by the magnification obtaining unit in the coordinates of the original image. Interpolation point defining means for converting into coordinate values and defining an interpolation point on the coordinates of the original image by shifting to the left and upper sides by half a pixel of the original image ,
For each of all the interpolation points defined by the interpolation point defining means, based on the vertical distance and the horizontal distance of the rectangular area defined by the four original pixels around the interpolation point. Reverse distance value generating means for performing processing for calculating the value of the reverse distance in the vertical direction and the value of the reverse distance in the horizontal direction from the four original pixels to the interpolation point;
Of all the interpolation points defined by the interpolation point defining means, outside the area occupied by the original pixels constituting the original image, and a predetermined density difference between the pixel values of the four surrounding original pixels For a certain interpolation point, a corrected inverse corresponding to the inverse distance value calculated by the inverse distance value generating means for the interpolation point from the storage device storing the inverse distance value and the corrected inverse distance value in association with each other. A reverse distance value correcting means for reading out the distance value and replacing the reverse distance value with a corrected reverse distance value;
After all of the interpolation points defined by the interpolation point defining means are replaced by the inverse distance value calculated by the inverse distance value generating means for the interpolation point or by the inverse distance value correcting means. A pixel value generating means for performing a process of calculating a reverse distance weighted average of the interpolation point based on the corrected reverse distance value of the interpolation point and the pixel values of the four original pixels around the interpolation point; and the pixel value A pixel interpolation program that functions as image data output means for outputting image data composed of pixel values generated for each interpolation point by the generating means as image data of an enlarged image.