JP5287581B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing method for enlarging an image, and more particularly to an image processing apparatus and an image processing method for reducing the jagged feeling of oblique lines included in an enlarged image.

  There is a case where an input image signal having a small number of pixels is desired to be enlarged to the number of display pixels on a digital display provided in a personal computer or a TV receiver. In such a case, it is common to use either the linear interpolation method or the bicubic method. In these methods, filtering according to the output image position is performed, and pixels are interpolated between the original pixels. As a result, the image can be enlarged by interpolating the pixels.

  However, according to a conventional general method such as a linear interpolation method or a bicubic method, there is a problem that a diagonal line included in an output image (enlarged image) looks jagged as described later. In view of this, an image enlargement method for reducing the jagged feeling of diagonal lines has been proposed. For example, in Patent Document 1, a pixel based on oblique line information is generated in the middle between two scanning lines adjacent to the output pixel position, and the pixel on the scanning line adjacent to the output pixel position and the intermediate pixel are A method of converting scan lines using the method is disclosed. In Patent Document 2, the edge detection results in the horizontal direction and the vertical direction between the original images around the position where the pixel is interpolated are compared, and the pixels determined to be continuous are filtered by the continuous pixels. A method for pixel interpolation is disclosed.

  FIG. 11 is a diagram for explaining a difference between the above-described conventional techniques. FIG. 11A shows an input image (image before enlargement). As shown in this figure, it is assumed that a boundary line between a bright part (white) and a dark part (black) exists in an oblique direction in the input image. FIG. 11B shows an image after the input image is enlarged twice in the vertical direction by a conventional general method. In the conventional general method, an average value of two original pixels adjacent in the vertical direction is obtained, and pixels having a pixel value of this average value are interpolated. For this reason, as shown in FIG. 11B, there is a problem that gray pixels are interpolated in the vicinity of the boundary line, and the oblique line looks jagged. FIG. 11C shows an image after the input image is enlarged twice in the vertical direction by the conventional method disclosed in Patent Document 1 or 2. In the conventional method disclosed in Patent Document 1 or 2, an oblique line included in an input image is detected, and an interpolation process using the correlation between pixel pairs on the detected oblique line is performed. Therefore, as shown in FIG. 11C, black pixels are interpolated on the diagonal lines, and white pixels are interpolated on the areas other than the diagonal lines. As a result, the diagonal lines are compared with the conventional general method. Can reduce the jagged feeling.

JP 2005-12601 A JP 2008-271085 A

  However, the device disclosed in Patent Document 1 is a scanning line conversion device that converts the number of pixels in the vertical direction. That is, Patent Document 1 does not disclose a configuration for enlarging an image in the horizontal direction in addition to the vertical direction.

  Further, the conventional method disclosed in Patent Document 2 has a problem in terms of the image quality of the enlarged image. That is, when the image is enlarged in the vertical direction and the horizontal direction, it is preferable to reflect the result of one enlargement process in the other enlargement process (described later). However, in the conventional method disclosed in Patent Document 2, any one of the horizontal enlargement filtering result, the vertical enlargement filtering result, and the oblique direction enlargement filtering result is selected. Cannot be reflected in the enlargement process.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an image processing apparatus and an image that can enlarge an image in the vertical and horizontal directions with good image quality while reducing the jagged feeling of diagonal lines. It is to provide a processing method.

A first feature according to the embodiment of the present invention is an image processing apparatus that realizes enlargement processing with diagonal line processing by performing interpolation processing using correlation between pixel pairs on diagonal lines included in an input image. A determination unit that determines an output pixel position that is a position of a pixel included in the enlarged image, an extraction unit that extracts a predetermined number of input pixel groups around the output pixel position, the output pixel position and the predetermined number A first enlargement unit that performs the enlargement process with the diagonal line processing in one of the vertical direction and the horizontal direction based on the input pixel group, and the output pixel position and a predetermined number of extractions by the extraction unit By detecting a second diagonal line based on the input pixel group and generating a post-enlargement pixel based on the output pixel position, the second diagonal line, and the enlargement processing result of the first enlargement unit In the other of the vertical and horizontal directions It is to have a second expansion section for performing enlargement processing with serial diagonal process.

  According to a second feature of the present invention, in the image processing apparatus having the first feature, the first enlargement unit performs an enlargement process with the oblique line process in a vertical direction, A 2nd expansion part is performing the expansion process with the said diagonal line process in a horizontal direction.

According to a third feature of an embodiment of the present invention, in the image processing apparatus having the second feature, the first enlargement unit is a diagonal line included in the predetermined number of input pixel groups. A vertical enlargement oblique line detection unit that detects one oblique line, and a vertical enlargement oblique line that generates, as a preprocessing pixel, a pixel on a line adjacent to the output pixel position in the vertical direction based on the first oblique line. A line processing unit and a vertical expansion filter unit that generates a pixel after vertical expansion based on pixels on the first diagonal line, and the second expansion unit is included in the predetermined number of input pixel groups A horizontal enlargement oblique line detection unit for detecting a second oblique line that is a diagonal line, and the output pixel position and the horizontal direction based on the second oblique line and the enlargement processing result of the first enlargement unit water generates a pixel on an adjacent line as preprocessing pixels And enlarging oblique line processing unit, it is to have a horizontal expansion filter portion for generating a pixel after horizontal expansion based on pixels of the second diagonal line.

  According to a fourth feature of the present invention, in the image processing apparatus having the third feature, the vertical enlargement oblique line detection unit is configured to output the output pixel position in the predetermined number of input pixel groups. A diagonal line passing through the pair of input pixels on the v-th line and the v + 1-th line adjacent to each other in the vertical direction is detected as the first diagonal line, and the horizontal enlargement diagonal line detection unit detects the predetermined number of input pixel groups. Of the second diagonal line passing through the input pixel pair on the h pixel and the h + 1 pixel adjacent to the output pixel position in the horizontal direction and passing between the two consecutive pixels in the vertical direction after expansion. It is to detect as a line.

  According to a fifth feature of the present invention, in the image processing apparatus having the fourth feature, the vertical enlargement oblique line detection unit includes a pixel of the plurality of first oblique lines. A correlation that minimizes the absolute value of the difference between the pair is detected, and the horizontal enlargement oblique line detection unit detects the pixel pair from among the plurality of second oblique lines and the two consecutive vertically enlarged pixel pairs. Is to detect the correlation that minimizes the difference absolute value.

According to a sixth feature of the present invention, in the image processing apparatus having the fifth feature, the vertical enlargement oblique line processing unit weights the output pixel position according to a vertical distance. and, wherein the horizontal expansion for oblique line processing unit is to Rukoto weighted according to the horizontal distance of the output pixel position.

  According to a seventh feature of the embodiment of the present invention, in the image processing apparatus having the third feature, the vertical enlargement filter unit weights the output pixel position according to a vertical distance. The vertically expanded pixel is generated, and the horizontal expansion filter unit weights the output pixel position according to a horizontal distance to generate the horizontally expanded pixel.

An eighth feature according to the embodiment of the present invention is an image processing method for realizing enlargement processing with diagonal line processing by performing interpolation processing using correlation between pixel pairs on diagonal lines included in an input image. A determination step of determining an output pixel position that is a position of a pixel included in the enlarged image, an extraction step of extracting a predetermined number of input pixel groups around the output pixel position, the output pixel position and the predetermined number A first enlargement step for performing the enlargement process with the oblique line process in one of the vertical direction and the horizontal direction based on the input pixel group, and the output pixel position and a predetermined number of extractions performed in the extraction step By detecting a second diagonal line based on the input pixel group and generating a post-enlargement pixel based on the output pixel position, the second diagonal line, and the enlargement processing result of the first enlargement step Vertical direction A second expansion step of performing enlargement processing with the oblique line processing to the other of fine horizontal is that having a.

  According to the present invention, it is possible to provide an image processing apparatus and an image processing method capable of enlarging an image in the vertical and horizontal directions with good image quality while reducing the jagged feeling of diagonal lines.

1 is a configuration diagram of an image processing apparatus according to an embodiment of the present invention. It is a figure which shows operation | movement of the image processing apparatus in embodiment of this invention. It is a figure which shows the pixel position before and behind expansion in embodiment of this invention. It is a figure which shows the range of the input pixel group in embodiment of this invention. It is a figure which shows the combination of the diagonal line in embodiment of this invention. It is a figure which shows the diagonal line and vertical direction distance in embodiment of this invention. It is a figure which shows the combination of the diagonal line in embodiment of this invention. It is a figure which shows the diagonal line and horizontal direction distance in embodiment of this invention. It is a block diagram of the apparatus which developed the prior art currently disclosed by patent document 1. FIG. It is a figure which shows the difference with the prior art currently disclosed by patent document 2, and this invention. It is a figure for demonstrating the difference in a prior art.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a configuration diagram of an image processing apparatus according to an embodiment of the present invention. This image processing apparatus is an image processing apparatus that realizes enlargement processing with diagonal line processing by performing interpolation processing using correlation between pixel pairs on diagonal lines included in an input image. FIG. As shown, an output pixel position determination unit 101, an m × n pixel extraction unit 102, a vertical direction enlargement unit 301 with diagonal line processing, and a horizontal direction enlargement unit 302 with diagonal line processing are provided.

  The output pixel position determination unit 101 determines an output pixel position based on the input vertical synchronization signal and horizontal synchronization signal, the vertical direction enlargement ratio, and the horizontal direction enlargement ratio. The output pixel position is a position of a pixel included in the enlarged image, and is represented by, for example, coordinate values of x and y.

  The m × n pixel extraction unit 102 is a processing unit that extracts horizontal m pixels × vertical n pixels around the output pixel position from the input image signal, and specifically includes a line buffer of n lines or more. Is done. The pixels extracted by the m × n pixel extraction unit 102 are at least 2 lines × 2 pixels or more. Further, the pixel value of the image signal input to the m × n pixel extraction unit 102 is RGB or a luminance color difference value.

  The vertical enlargement unit 301 with diagonal line processing is a processing unit that performs enlargement processing with diagonal line processing in the vertical direction based on the output pixel position and the m × n input pixel group, and is shown in FIG. As described above, a vertical enlargement oblique line detection unit 103, a vertical enlargement oblique line processing unit 105, and a vertical direction enlargement filter unit 106 are provided. The vertical enlargement oblique line detection unit 103 detects an oblique line (first oblique line) included in the m × n input pixel group. The vertical enlargement oblique line processing unit 105 generates pixels (first preprocessed pixels) on the first oblique line. The vertical enlargement filter unit 106 generates a post-vertical enlargement pixel based on the pixels on the first diagonal line.

  The horizontal enlargement unit 302 with diagonal line processing is an enlargement process with diagonal line processing in the horizontal direction based on the output pixel position, the m × n input pixel group, and the enlargement processing result of the vertical direction enlargement unit 301 with diagonal line processing. As illustrated in FIG. 1B, the processing unit includes a horizontal enlargement oblique line detection unit 104, a horizontal enlargement oblique line processing unit 107, and a horizontal direction enlargement filter unit 108. The horizontal enlargement oblique line detection unit 104 detects an oblique line (second oblique line) included in the m × n input pixel group. The horizontal enlargement oblique line processing unit 107 generates pixels (second pre-processed pixels) on the second oblique line. The horizontal expansion filter unit 108 generates a pixel after horizontal expansion based on the pixels on the second diagonal line.

  FIG. 2 is a flowchart showing the operation of the image processing apparatus according to the embodiment of the present invention. Hereinafter, the configuration of the image processing apparatus will be described in more detail with reference to FIG.

  First, the output pixel position determining unit 101 determines an output pixel position (step S1). The output pixel position is a position of a pixel included in the enlarged image. The vertical position is determined based on the vertical magnification, and the horizontal position is determined based on the horizontal magnification. FIG. 3A shows the relationship between the pixel positions before and after enlargement in the vertical direction. When the interval between adjacent pixels before vertical expansion is 1, the interval dv between adjacent pixels after vertical expansion is the reciprocal of the vertical expansion rate. For example, when the vertical enlargement ratio is twice, the interval dv between adjacent pixels after the vertical enlargement is 0.5. In this case, the vertical distance y of the output pixel position is a fractional part of the distance from the upper end in the vertical direction, such as “0”, “0.5”, “0”, “0.5”,. It is represented by Similarly, FIG. 3B shows a relationship between pixel positions before and after enlargement in the horizontal direction. When the interval between adjacent pixels before horizontal expansion is 1, the interval dh between adjacent pixels after horizontal expansion is the reciprocal of the horizontal expansion rate. For example, when the horizontal enlargement ratio is double, the interval dh between adjacent pixels after horizontal enlargement is 0.5. In this case, the horizontal distance x of the output pixel position is a fractional part of the distance from the left end in the horizontal direction, such as “0”, “0.5”, “0”, “0.5”,. It is represented by

  Next, the m × n pixel extraction unit 102 extracts an input pixel group around the output pixel position (step S2). FIG. 4 shows the range of the input pixel group extracted by the m × n pixel extraction unit 102. That is, the m × n pixel extraction unit 102 extracts m × n pixels around four input pixels surrounding the output pixel position, as indicated by a solid line frame in the drawing.

  Next, the vertical enlargement oblique line detection unit 103 detects an oblique line passing through the pair of input pixels adjacent to the output pixel position in the vertical direction (step S3). FIG. 5 shows combinations of diagonal lines detected from the input pixel group extracted by the m × n pixel extraction unit 102. That is, the vertical enlargement oblique line detection unit 103 is a line segment connecting a pixel pair on the v-th line and the v + 1-th line passing between the pixels P (h, v) and P (h, v + 1), specifically, , Line segment P (h, v) -P (h, v + 1), line segment P (h-1, v) -P (h + 1, v + 1), line segment P (h + 1, v) -P (h-1, v + 1), line segment P (h−2, v) −P (h + 2, v + 1), line segment P (h + 2, v) −P (h−2, v + 1),. In order to select one of the detected plurality of diagonal lines, first, an absolute difference value of a pixel pair of each diagonal line is calculated.

Correlation Cva = | P (h, v) −P (h, v + 1) |
Correlation Cvb = | P (h-1, v) −P (h + 1, v + 1) |
Correlation Cvc = | P (h + 1, v) −P (h-1, v + 1) |
Correlation Cvd = | P (h-2, v) −P (h + 2, v + 1) |
Correlation Cve = | P (h + 2, v) −P (h-2, v + 1) |
...
Next, the vertical enlargement oblique line detection unit 103 detects a correlation that is the minimum value among the calculated absolute difference values (correlation), and outputs oblique line information (pixel pairs) constituting the detected correlation. For example, when the correlation Cvc is the minimum value, diagonal line information (Dv1, Dv2) = (P (h + 1, v), P (h-1, v + 1)) is output.

  Next, the vertical enlargement diagonal line processing unit 105 generates a pixel (pre-processing pixel) to be used by the vertical direction enlargement filter unit 106 (step S4). First, the oblique line information Dv1 and P (h, v) output from the vertical enlargement oblique line detection unit 103 are interpolated according to the vertical distance y between P (h, v) and the output pixel position. In other words, the diagonal line information Dv1 and P (h, v) are weighted according to the vertical distance y.

Preprocess pixel Tv1 = P (h, v) × (0.5−y) × 2 + Dv1 × y × 2 (y ≦ 0.5)
Tv1 = Dv1 (y> 0.5)
Similarly, the oblique line information Dv2 and P (h, v + 1) are interpolated according to the vertical distance y between P (h, v) and the output pixel position. In other words, the diagonal line information Dv2 and P (h, v + 1) are weighted according to the vertical direction distance y.

Pre-processing pixel Tv2 = Dv2 (y <0.5)
Tv2 = P (h, v + 1) × (y−0.5) × 2 + Dv2 × (1.0−y) × 2 (y ≧ 0.5)
FIG. 6 shows the diagonal line and the vertical distance y when the diagonal line information (Dv1, Dv2) is (P (h + 1, v), P (h-1, v + 1)). The preprocessing pixel Tv1 is calculated from P (h, v) and P (h + 1, v), and the preprocessing pixel Tv2 is calculated from P (h, v + 1) and P (h-1, v + 1). If such pre-processing is performed, weighting according to the vertical distance y is performed, so that even when the output pixel position matches the input pixel position, the occurrence of discontinuous phenomena such as “pixel skip” is suppressed. Is possible.

  Next, the vertical expansion filter 106 performs vertical expansion interpolation (step S5). That is, interpolation is performed using the preprocessed pixels Tv1 and Tv2 generated by the vertical enlargement diagonal line processing unit 105 according to the distance y from the output pixel position. In other words, the preprocessing pixels Tv1 and Tv2 are weighted according to the vertical distance y.

Vertically expanded pixel SV = Tv1 × (1−y) + Tv2 × y
Thereby, the vertical enlargement process with the diagonal line process is completed, and the vertically enlarged pixel SV is output. The contents of the vertical enlargement process with the diagonal line process are not limited to the contents exemplified here. For example, as disclosed in Patent Document 1, vertical enlargement processing with diagonal line processing can also be realized by interpolation between an intermediate value of a pixel pair of diagonal lines and a pixel adjacent to the output pixel position.

  On the other hand, horizontal enlargement oblique line detection processing is performed in parallel with the vertical enlargement oblique line detection processing (step S3). That is, the horizontal enlargement oblique line detection unit 104 detects an oblique line passing between the output pixel position and the vertically enlarged pixels adjacent in the horizontal direction (step S6). FIG. 7 shows combinations of diagonal lines detected from the input pixel group extracted by the m × n pixel extraction unit 102. Here, SV (h, v) and SV (h + 1, v) in the figure are two consecutive pixels after enlargement in the vertical direction. “Consecutive” means that SV (h + 1, v) is output from the vertical expansion filter 106 following SV (h, v). The horizontal enlargement diagonal line detection unit 104 passes between SV (h, v) and SV (h + 1, v) and is a line segment of a pixel pair on the h-th pixel and the h + 1-th pixel, specifically, a line Min SV (h, v) -SV (h + 1, v), line segment P (h, v) -P (h + 1, v + 1), line segment P (h + 1, v) -P (h, v + 1), line segment P (H, v-1) -P (h + 1, v + 2), line segment P (h + 1, v-1) -P (h, v + 2),... Are detected. In order to select one of the detected plurality of diagonal lines, first, an absolute difference value of a pixel pair of each diagonal line is calculated.

Correlation Cha = | SV (h, v) −SV (h, v + 1) |
Correlation Chb = | P (h, v) −P (h + 1, v + 1) |
Correlation Chc = | P (h + 1, v) −P (h, v + 1) |
Correlation Chd = | P (h, v-1) −P (h + 1, v + 2) |
Correlation Che = | P (h + 1, v-1) −P (h, v + 2) |
...
Next, the horizontal enlargement diagonal line detection unit 104 detects a correlation that is the minimum value among the calculated absolute difference values (correlation), and outputs diagonal line information (pixel pairs) that constitutes the detected correlation. For example, when the correlation Chc is the minimum value, diagonal line information (Dh1, Dh2) = (P (h, v + 1), P (h + 1, v)) is output.

  Next, the horizontal enlargement diagonal line processing unit 107 generates a pixel (pre-processing pixel) used in the horizontal direction enlargement filter unit 108 (step S7). First, the oblique line information Dh1 and SV (h, v) output from the horizontal enlargement oblique line detection unit 104 are interpolated according to the horizontal distance x between SV (h, v) and the output pixel position. In other words, the diagonal line information Dh1 and SV (h, v) are weighted according to the horizontal distance x.

Pre-processing pixel Th1 = SV (h, v) × (0.5−x) × 2 + Dh1 × x × 2 (x ≦ 0.5)
Th1 = Dh1 (x> 0.5)
Similarly, the diagonal line information Dh2 and SV (h + 1, v) are interpolated according to the horizontal distance x between SV (h, v) and the output pixel position. In other words, the diagonal line information Dh2 and SV (h + 1, v) are weighted according to the horizontal distance x.

Preprocessing pixel Th2 = Dh2 (x <0.5)
Th2 = SV (h + 1, v) × (x−0.5) × 2 + Dh2 × (1.0−x) × 2 (x ≧ 0.5)
FIG. 8 shows the diagonal line and the horizontal distance x when the diagonal line information (Dh1, Dh2) is (P (h, v + 1), P (h + 1, v)). The preprocessing pixel Th1 is calculated from SV (h, v) and P (h, v + 1), and the preprocessing pixel Th2 is calculated from SV (h + 1, v) and P (h + 1, v). If such pre-processing is performed, weighting according to the horizontal distance x is performed, so that even when the output pixel position coincides with the pixel position after enlargement in the vertical direction, a discontinuous phenomenon such as “pixel skip” occurs. Can be suppressed.

  Next, the horizontal enlargement filter unit 108 performs horizontal enlargement interpolation (step S8). In other words, the pre-processing pixels Th1 and Th2 generated by the horizontal enlargement oblique line processing unit 107 are used for interpolation according to the distance x from the output pixel position. In other words, the preprocessing pixels Th1 and Th2 are weighted according to the horizontal distance x.

Horizontally expanded pixel SH = Th1 × (1−x) + Th2 × x
As a result, the pixel SH expanded with the diagonal line processing in the vertical direction and the horizontal direction is output. The same process is repeated for all output pixels (step S9).

  As described above, in the present invention, the vertical type enlargement process 301 with the diagonal line processing and the horizontal direction enlargement part 302 with the diagonal line processing are employed in series, so that the vertical enlargement process is performed. The result is reflected in the horizontal enlargement process. This makes it possible to enlarge the image in the vertical and horizontal directions with good image quality while reducing the jagged feeling of the diagonal lines. Hereinafter, in order to clarify the effect of the present invention, differences from the prior art disclosed in Patent Documents 1 and 2 will be described.

  FIG. 9 is a configuration diagram of an apparatus obtained by developing the prior art disclosed in Patent Document 1. In FIG. As already described, the apparatus disclosed in Patent Document 1 is a scanning line conversion apparatus. Therefore, when the number of horizontal pixels of the input image is different from the number of horizontal display pixels of the digital display, the image cannot be displayed as it is. Therefore, as shown in FIG. 9, two devices disclosed in Patent Document 1 are connected in series. If the former apparatus is a vertical pixel number conversion apparatus and the latter apparatus is a horizontal pixel number conversion apparatus by rotating an image by 90 degrees, it is estimated that an image can be displayed. However, in this case, as shown in FIG. 9, both the vertical pixel number conversion device and the horizontal pixel number conversion device require m × n pixel extraction units 304 and 307 used for oblique line detection and the like. The m × n pixel extraction units 304 and 307 are configured by line buffers. For an image with a large number of horizontal pixels such as a high-definition image, the memory capacity required for the line buffer becomes very large, which is a major increase in the cost of the apparatus. There is a problem that becomes a cause. In addition, since diagonal lines are detected using pixels after vertical enlargement, the difference in pixel values between adjacent pixels after vertical enlargement tends to decrease as the vertical enlargement ratio increases. As a result, there is a problem that the detection accuracy of the oblique line in the m × n pixel extraction unit 307 of the horizontal direction pixel number conversion device is lowered. On the other hand, in the present invention, as shown in FIG. 1A, there is no processing unit corresponding to the m × n pixel extraction unit 307 that occupies a large proportion of the circuit scale of the entire apparatus. Therefore, the apparatus cost can be reduced by reducing the circuit scale.

  FIG. 10 is a conceptual diagram for briefly explaining the difference between the prior art disclosed in Patent Document 2 and the present invention. As already described, with the conventional method disclosed in Patent Document 2, the result of the vertical enlargement process cannot be reflected in the horizontal enlargement process. On the other hand, in the present invention, as shown in FIG. 10, after the gray enlarged pixel 14 is generated using the white original pixel 11 and the black original pixel 15, Can be used to generate the light gray post-enlargement pixel 13. That is, according to the present invention, the result of the vertical enlargement process is reflected in the horizontal enlargement process, so that it is possible to provide better image quality than the conventional technique disclosed in Patent Document 2. is there.

  Here, the horizontal enlargement process is performed after the vertical enlargement process, but the present invention is not limited to this. That is, when it is necessary to perform the vertical direction enlargement process after performing the horizontal direction enlargement process, the vertical direction enlargement unit 301 with diagonal line processing and the horizontal direction enlargement unit 302 with diagonal line processing can be interchanged. In this case, it goes without saying that the result of the horizontal enlargement process is reflected in the vertical enlargement process.

DESCRIPTION OF SYMBOLS 101 ... Output pixel position determination part 102 ... m * n pixel extraction part 103 ... Vertical expansion diagonal line detection part 104 ... Horizontal expansion diagonal line detection part 105 ... Vertical expansion diagonal line processing part 106 ... Vertical expansion filter part 107 ... An oblique line processing unit 108 for horizontal enlargement ... A horizontal direction enlargement filter unit 301 ... A vertical direction enlargement unit 302 with oblique line processing ... A horizontal direction enlargement unit with oblique line processing

Claims (8)

An image processing apparatus that realizes enlargement processing with diagonal line processing by performing interpolation processing using correlation between pixel pairs on diagonal lines included in an input image,
A determination unit that determines an output pixel position that is a position of a pixel included in the enlarged image;
An extraction unit for extracting a predetermined number of input pixel groups around the output pixel position;
A first enlargement unit that performs an enlargement process with oblique line processing in one of a vertical direction and a horizontal direction based on the output pixel position and the predetermined number of input pixel groups;
A second diagonal line is detected based on the output pixel position and a predetermined number of input pixel groups extracted by the extraction unit, and the output pixel position, the second diagonal line, and the first enlargement unit And a second enlargement unit that performs enlargement processing with diagonal line processing in the other of the vertical direction and the horizontal direction by generating post-enlargement pixels based on the enlargement processing result. Image processing device. The first enlargement unit performs an enlargement process with the oblique line process in a vertical direction,
The image processing apparatus according to claim 1, wherein the second enlargement unit performs the enlargement process with the diagonal line process in a horizontal direction. The first enlargement unit detects a first oblique line, which is an oblique line included in the predetermined number of input pixel groups, and an output based on the first oblique line and the output based on the first oblique line A vertical enlargement diagonal line processing unit that generates pixels on a line adjacent to the pixel position in the vertical direction as a preprocessing pixel, and a vertical enlargement that generates a post-vertical enlargement pixel based on the pixels on the first diagonal line With a filter section,
The second enlargement unit includes a horizontal enlargement oblique line detection unit that detects a second oblique line that is an oblique line included in the predetermined number of input pixel groups, the second oblique line, and the first oblique line. Based on a horizontal enlargement diagonal line processing unit that generates a pixel on a line adjacent to the output pixel position in the horizontal direction as a preprocessing pixel based on an enlargement processing result of the enlargement unit, and a pixel on the second diagonal line The image processing apparatus according to claim 2, further comprising a horizontal enlargement filter unit that generates pixels after horizontal enlargement. The vertical enlargement oblique line detection unit detects an oblique line passing through an input pixel pair on the v-th line and the (v + 1) -th line adjacent to the output pixel position in the vertical direction in the predetermined number of input pixel groups. Detected as diagonal lines,
The horizontal enlargement oblique line detection unit passes through the input pixel pair on the h pixel and the h + 1 pixel adjacent to the output pixel position in the horizontal direction in the predetermined number of input pixel groups, and two continuous pixels. The image processing apparatus according to claim 3, wherein an oblique line passing between the pixels after the vertical enlargement is detected as the second oblique line. The vertical enlargement diagonal line detection unit detects a correlation that minimizes the absolute value of the difference between the pixel pair from the plurality of first diagonal lines,
The horizontal enlargement oblique line detection unit detects a correlation that minimizes the absolute value of the difference between the pixel pair from the plurality of the second oblique lines and the two consecutive pixel enlargements in the vertical direction. The image processing apparatus according to claim 4, wherein: The vertical enlargement oblique line processing unit performs weighting according to a vertical distance of the output pixel position ,
The horizontal enlarging oblique line processing unit, an image processing apparatus according to claim 5, wherein the you weighted according to the horizontal distance of the output pixel position. The vertical enlargement filter unit generates the post-vertical enlargement pixel by weighting according to a vertical distance of the output pixel position,
The image processing apparatus according to claim 3, wherein the horizontal enlargement filter unit generates the post-horizontal enlargement pixel by performing weighting according to a horizontal distance of the output pixel position. An image processing method for realizing enlargement processing with diagonal line processing by performing interpolation processing using correlation between pixel pairs on diagonal lines included in an input image,
A determination step of determining an output pixel position that is a position of a pixel included in the image after enlargement;
An extraction step of extracting a predetermined number of input pixel groups around the output pixel position;
A first enlargement step for performing enlargement processing with diagonal line processing in one of a vertical direction and a horizontal direction based on the output pixel position and the predetermined number of input pixel groups;
A second diagonal line is detected based on the output pixel position and a predetermined number of input pixel groups extracted in the extraction step, and the output pixel position, the second diagonal line, and the first expansion step are detected. And a second enlargement step for performing enlargement processing with diagonal line processing in the other of the vertical direction and the horizontal direction by generating post-enlargement pixels based on the enlargement processing result. Image processing method.

Images