JP3384875B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is applied to a facsimile machine, a digital copying machine, and the like, and performs an enlargement conversion process on image data obtained by an image scanner or the like. It relates to a processing device. 2. Description of the Related Art There are two types of image enlargement processing in facsimile apparatuses and digital copiers: optical processing and data processing. Data processing is more advantageous in terms of cost. In general, image enlargement by data processing is performed by inserting pixel data obtained by linear interpolation at a rate corresponding to an enlargement ratio between a large number of pixel data obtained by an image sensor. That is, for example, if the output of the image sensor shown in FIG. 6B when the image shown in FIG. 6A is read with the position indicated by L in the figure as the main scanning position is that shown in FIG. When the image is enlarged by 4/3 (approximately 133%), as shown in FIG. 6 (c), linearly with respect to three pixels (hereinafter referred to as read pixels) indicated by pixel data generated by the image sensor. Pixel obtained by interpolation (hereinafter referred to as interpolation pixel)
In the case of inserting one and enlarging at an enlargement ratio of 7/3 (about 233%), as shown in FIG. 6D, four interpolated pixels are inserted for three read pixels. The above is the enlargement process in the main scanning direction. The enlargement process in the sub-scanning direction is easily performed by using a line sensor as an image sensor and adjusting the reading timing during a period in which the original is conveyed by a fixed amount. be able to. When an image is handled in binary of white / black as in a facsimile machine or a digital copying machine, contrast enhancement processing is performed in order to reproduce fine white and black fine lines. In this contrast enhancement processing, all the pixels are sequentially set as a pixel of interest, and the level R of the pixel of interest is calculated as shown in FIG.
As shown in ( ₁₎ , each pixel located on the upper, left, right and lower sides of the target pixel is referred to as a reference pixel (O ₁ , O ₂ , O ₃ , O ₄ ).
R ′ = 5 × R− (O ₁ + O ₂ + O ₃ + O ₄ ) This is a process of replacing with R ′, which is obtained by the formula: That is, contrast enhancement is performed by Laplacian operation. When such a contrast enhancement process is performed on an image enlarged as described above, when the enlargement ratio exceeds 200%, the ratio of the number of interpolated pixels becomes larger than the number of read pixels, and contrast enhancement is performed. Cannot be performed properly. For example, image data obtained by reading the image shown in FIG. 8A with the position of L as the main scanning position is shown in FIG.
Assuming that the image data is as shown in FIG. 8B, when the image data is enlarged at a rate of 400% by the above-described method, the image data becomes as shown in FIG.
As shown in (c), the data has three interpolation pixels between the read pixels. Therefore, when the above-described contrast enhancement processing is performed, the value of the interpolation pixel does not change at all, and only the read pixel is enhanced. Therefore, an image obtained by binarizing the image data of FIG. 8D at the slice level shown in FIG. 8 is as shown in FIG. 8E, and the widths of the black and white thin lines are shown in FIG. It will not be four times the original image shown. As described above, the black line becomes thinner, and the white line is broken. [0008] As described above, in the conventional image processing apparatus, the image is enlarged at an arbitrary enlargement ratio by the linear interpolation process, and then the contrast enhancement process is performed. When the enlargement ratio is increased and the ratio of the interpolation pixels is increased, the contrast enhancement processing is not properly performed, and the thinning of the black line and the collapse of the white line occur, resulting in a problem that the image quality is deteriorated. The present invention has been made in view of such circumstances, and a purpose thereof is to accurately generate image data faithful to an image obtained by enlarging an image to be processed at an arbitrary enlargement ratio. It is an object of the present invention to provide an image processing apparatus capable of performing the above. In order to achieve the above-mentioned object, the present invention provides, for example, a pixel unit of original image data such as multi-valued image data generated by a reading unit and an A / D conversion circuit. In the meantime, n-1 (n is a predetermined integer: for example, 4) pixel data is inserted by a predetermined interpolation process such as a linear interpolation into an image n times as large as the image corresponding to the original image data. From the corresponding enlargement means such as a quadruple circuit for generating n-times image data such as four-times image data, and from the n-times image data generated by this enlargement means, pixel data is generated at a frequency corresponding to the enlargement ratio. For example, a thinning unit including a thinning circuit and a thinning timing generating circuit, which generates enlarged image data by thinning, and an image corresponding to the n-times image data with respect to arbitrary pixel data of interest. The process of obtaining the number m of pixels to be thinned out by the thinning means among the pixels located between the pixel of interest corresponding to the pixel of interest data and a pixel located n pixels away from the pixel of interest, A thinning-out pixel number calculating means such as a pixel interval calculating circuit for performing each pixel data of the n-times image data as a pixel of interest, and for any pixel data of interest, at least the level of the pixel data of interest and the pixel corresponding to this pixel of interest performing a process of performing a predetermined contrast enhancement based on the level of the pixel data corresponding to the pixel located at a position apart by nm pixels from each pixel data of the enlarged image data generated by the thinning unit as a pixel of interest; For example, an emphasis unit including a memory unit, an address generation circuit, a multiplication circuit, a coefficient generation circuit, and an addition circuit is provided. By taking such means, the original image data is temporarily enlarged by the enlargement means at the maximum enlargement ratio (n times) to become n times image data, and the n times image data is By subtracting pixel data from the image data at a frequency corresponding to an arbitrary enlargement ratio, enlarged image data corresponding to an image obtained by enlarging the image corresponding to the original image data to an arbitrary enlargement ratio is generated. At this time, for any pixel data of interest,
Number of pixels to be thinned out by thinning out of pixels located between a target pixel corresponding to the target pixel data and a pixel located n pixels away from the target pixel in an image corresponding to the n-times image data The process of obtaining m is performed using each pixel data of the n-times image data as a pixel of interest. Further, for any pixel data of interest, predetermined contrast enhancement is performed based on at least the level of the pixel data of interest and the level of pixel data corresponding to a pixel located at a position n-m pixels away from the pixel corresponding to this pixel of interest. The process to be performed is performed using each pixel data of the enlarged image data generated by the thinning unit as a pixel of interest. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a functional block diagram illustrating a main configuration of an image reading apparatus configured by applying the image processing apparatus according to the present embodiment. This image reading apparatus has a reading section 1, an A / D conversion circuit 2, a line memory 3, a quadruple circuit 4, a thinning circuit 5, a thinning timing generating circuit 6, a pixel interval calculating circuit 7, a memory section 8, an address, A generating circuit 9, a multiplying circuit 10,
It comprises a coefficient generating circuit 11 and an adding circuit 12. The reading section 1 further includes a light source 1a, a line sensor 1b, and document feed rollers 1c, 1d, 1e, 1f, and a document conveyed at a constant speed by the document feed rollers 1c, 1d, 1e, 1f. Light source 1a for 20
, And the reflected light from the original 20 at this time is received by the line sensor 1b, and an electric signal (analog image signal) corresponding to the reflected light image is repeatedly generated for each line. The A / D conversion circuit 2 converts an analog image signal output from the line sensor 1b into 8-bit multi-value image data. The line memory 3 holds the multi-valued image data supplied from the A / D conversion circuit 2 for one line and supplies the same to the quadruple circuit 4. The quadruple circuit 4 performs a process of enlarging an image corresponding to the multivalued image data four times by interpolation for each line of multivalued image data held in the line memory 3. When the thinning-out timing signal generated by the thinning-out timing generating circuit 6 is ON, the thinning-out circuit 5 does not output the data output from the quadruple circuit 4, thereby performing the thinning-out. The thinning-out timing generating circuit 6 generates a thinning-out timing signal which is turned on at a ratio corresponding to the image enlargement ratio designated by the operator, and supplies it to the thinning-out circuit 5. The pixel interval calculation circuit 7 sets a pixel corresponding to each pixel data in the multi-valued image data output from the quadruple circuit 4 as a pixel of interest, and the pixel at a position separated from the pixel of interest by 4 pixels Is determined based on the image enlargement ratio specified by the operator, and is notified to the address generation circuit 9. The memory unit 8 is composed of three line memories 8a, 8b and 8c each having a capacity for holding one line of multi-valued image data expanded four times, and connected in series. The output multi-valued image data is held for three consecutive lines. Then, the memory unit 8 calculates the address based on the address generated in consideration of the pixel interval obtained in the pixel interval calculation circuit 7 by the address generation circuit 9.
The pixel data of predetermined five pixels out of the held multi-valued image data of three lines is output to the multiplication circuit 10. The multiplication circuit 10 multiplies each of the five pixel data supplied from the memory unit 8 by a coefficient generated by the coefficient generation circuit 11 in synchronization with the main scanning synchronization signal. Then, the multiplication circuit 10 outputs each pixel data after multiplication by the coefficient to the addition circuit 12. The addition circuit 12 adds the values of the respective pixel data supplied from the multiplication circuit 10. Next, the operation of the image reading apparatus configured as described above will be described. First, the reading unit 1
In this case, the light source 1a emits light to the original 20 while the original 20 is transported at a constant speed by the original feed rollers 1c, 1d, 1e, and 1f. The light radiated at this time is reflected by the original 20, and at a predetermined reading position, the original 2
Light reflected from 0 enters the line sensor 1b. The line sensor 1b repeatedly generates an analog image signal corresponding to an incident reflected light image from the document 20 at a predetermined cycle. The cycle at which the line sensor 1b generates an analog image signal generates an analog image signal each time the document 20 is conveyed by a predetermined sub-scanning line pitch when 100% is specified as the image enlargement ratio. The cycle is set in advance. The cycle when an image magnification other than 100% is designated is changed and set in accordance with the image magnification based on the cycle when the image magnification is 100%. That is, for example, if the image enlargement ratio is 200
%, The cycle in which the line sensor 1b generates the analog image signal is 1 of the cycle when the image enlargement ratio is 100%.
/ 2. Thus, the line sensor 1
By changing the reading cycle in b, enlargement processing in the sub-scanning direction is performed. The enlargement process in the sub-scanning direction may be performed by changing the transport speed of the original 20 while keeping the read cycle of the line sensor 1b constant, or by changing both the read interval of the line sensor 1b and the transport speed of the original 20. The same can be done. The analog image signal generated by the line sensor 1b is a signal obtained by serially connecting the outputs of a large number of photoelectric conversion elements constituting the line sensor 1b. The analog image signal is converted into digital multi-valued image data by converting the output level of each photoelectric conversion element into 8-bit pixel data in the A / D conversion circuit 2 and temporarily stored in the line memory 3. Is done. Subsequently, the multi-valued image data stored in the line memory 3 is enlarged by the quadruple circuit 4 at an enlargement ratio of 400% (4 times) regardless of the designated image enlargement ratio. This enlargement process is performed by using pixel data (read pixel data) included in the multi-valued image data stored in the line memory 3.
This is a process of inserting three pieces of interpolated pixel data by linear interpolation between them. Specifically, for example, an image formed on the document 20 is shown in FIG. 2A, and image data obtained when the image of the document 20 is read with the position of L as the main scanning position is shown in FIG. 2), the multivalued image data after the enlargement processing is as shown in FIG. In FIG. 2C, for convenience, the multi-valued image data is shown in the form of an analog signal. However, actually, the read pixel data and the interpolated pixel in which each level of each read pixel and each interpolated pixel is indicated by 8 bits are shown. It consists of an array of data. The multi-valued image data (hereinafter, referred to as quadrupled image data) enlarged to 400% in this manner is subsequently outputted to a thinning-out circuit 5 by a thinning-out timing generating circuit 6.
When the output of the thinning-out timing signal generated in step (1) is ON, the output is stopped, thereby thinning out the pixel data. Since the thinning-out timing signal is a signal that is turned ON at a ratio corresponding to the image enlargement ratio specified by the operator, the pixel data is thinned out from the quadrupled image data at a ratio according to the specified image enlargement ratio. Specifically, for example, if 200% is specified as the image enlargement ratio, pixel data is thinned out at a ratio of one pixel to two pixels. Thus, the output of the thinning circuit 5 includes the image data corresponding to the image obtained by enlarging the image corresponding to the multi-valued image data output from the A / D conversion circuit 2 at the image magnification specified by the operator. Become. The image data which has been subjected to the thinning processing in the thinning circuit 5 in this manner is sequentially transferred to the line memories 8a, 8b, 8c of the memory unit 8, and is transferred to the memory unit 8 for a reading period of three lines. Is held. Thus, the memory unit 8 holds three consecutive lines of image data. The address generation circuit 9 generates an address for outputting a pixel of interest so that each pixel data stored in the line memory 8b is sequentially output as a pixel of interest.
Further, the address generation circuit 9 outputs a value of 4 to the pixel indicated by the pixel data output by the target pixel output address.
An address for reference pixel output is generated such that pixel data of one reference pixel is output as reference pixel data. Here, the reference pixel corresponds to FIG.
The pixel at the position shown in (a), that is, the pixel at the same position in the main scanning direction of the previous line, the two pixels at a position separated by 4 pixels in the main scanning direction, and the main scanning of the next line The pixels are located at the same position in the direction. The two reference pixels located in the main scanning direction are pixels located four pixels apart from each other in the quadruple image because an adjacent pixel is an interpolation pixel and has accurate information for performing contrast enhancement. This is because pixels adjacent to the image data before quadrupling (having optimal information for performing contrast enhancement) are used as reference pixels. However, the image data supplied to the memory unit 8 is thinned out. If there is a thinned pixel as shown in FIG. In the state shown in FIG. 3B, the two reference pixels located in the main scanning direction are not 4 pixels away from the target pixel. Therefore, the pixel interval calculation circuit 7 calculates the pixel interval between the pixel of interest and the pixel to be used as the reference pixel in the image after the thinning-out, and the address generation circuit 9 determines the corresponding address based on the pixel interval. Is to occur. Specifically, the pixel interval calculation circuit 7 includes:
The pixel interval between the pixel to be a reference pixel and the pixel of interest in the thinned image is set as a table as shown in FIG. 4 in association with the image enlargement ratio, and the image enlargement specified by the operator is performed. The pixel interval is obtained by referring to this table according to the rate. Then, based on the pixel interval notified from the pixel interval calculation circuit 7, if the pixel interval "3" is notified, for example, as shown in FIG. One reference pixel is a pixel located three pixels away from the pixel of interest, and an address corresponding to this pixel is generated. When the reference pixel corresponds to the thinned pixel, an address corresponding to a pixel located next to the reference pixel is generated. The memory section 8 outputs five pixel data corresponding to the five addresses generated by the address generating circuit 9 to the multiplying circuit 10. Each pixel data of the image data is set in the pixel of interest in order,
The one reference pixel is also sequentially changed with respect to the pixel set as the target pixel. The 5 extracted in the memory unit 8 in this way
The two pixel data are output to a coefficient generation circuit 11 by a multiplication circuit 10.
Are multiplied. Here, the coefficient generation circuit 11 outputs “5” to the pixel of interest and “−1” to the reference pixel as coefficients.
Therefore, if the level of the target pixel is R and the levels of the reference pixels are O ₁ , O ₂ , O ₃ , and O ₄ , respectively, the level of the pixel data corresponding to the target pixel is “5R” and the level of the reference pixel is “5R”. The level of the pixel data is “−O ₁ ”,
_{"-O 2", "- O} 3", - is converted to "O _4". These pixel data are added to the adder 12
And added here. Accordingly, the level of the output data of the adder circuit 12 is 5 × R− (O ₁ + O ₂ + O ₃ + O ₄ ). That is, the Laplacian operation is performed on the target pixel, and the output data of the addition circuit 12 is
The result is that contrast enhancement is performed on the target pixel. Then, in the addition circuit 12, the above processing is repeated using each pixel data of the image data as the pixel data of interest. Therefore, the output of the addition circuit 12 is the data obtained by performing the contrast enhancement by the Laplacian operation on the image data. Become. Here, since the reference pixel located in the main scanning direction with respect to the target pixel is a pixel which is four pixels away from the target pixel in the quadrupled image, the density change of surrounding pixels also occurs with respect to the interpolation pixel. Is appropriately performed in accordance with. Specifically, for example, when the quadruple image data output from the quadruple circuit 4 is as shown in FIG. 2C, when the thinning circuit 5 does not perform the thinning (when the image enlargement ratio is 400%) 2), the image data output from the addition circuit 12 becomes data as shown in FIG. 2D. When this image data is binarized at the slice level shown in FIG. As a result, an image obtained by quadrupling the original image shown in FIG. 2A is obtained. In FIG. 2D, the image data is shown in the form of an analog signal for convenience, but actually, the read pixel data and the interpolated pixel data in which each level of each read pixel and each interpolated pixel is indicated by 8 bits. Consists of As described above, according to the present embodiment, the image is temporarily enlarged four times regardless of the designated image enlargement ratio, and then the pixel data is thinned out according to the designated image enlargement ratio to obtain a desired image. Corresponding image data is obtained, and contrast enhancement is performed on the thinned image data by using, as a reference pixel, a pixel located four pixels away from the pixel of interest in the main scanning direction in the state of the quadrupled image data. . As a result, appropriate contrast enhancement is applied to all pixel data according to the change in the density of surrounding pixels, and the original image can be enlarged to the specified image enlargement ratio without causing thinning of black lines or collapse of white lines. Thus, image data corresponding to an image faithfully enlarged is obtained. By the way, in the above embodiment, the pixel interval calculation circuit 7 determines the pixel interval uniformly according to the image magnification. However, depending on the image enlargement ratio, the number of thinned pixels between the target pixel and the reference pixel in the state of the quadrupled image may differ from part to part, and in such a case, the pixel interval is reduced as in the above-described embodiment. If you decide uniformly,
In some cases, the reference pixel used for contrast enhancement is shifted from a pixel located four pixels away from the pixel of interest in the main scanning direction in the quadruple image. Therefore, as shown in FIG. 5 (a), for each pixel of interest in the quadruple image, a section A up to the reference pixel on the front side (left side in the figure) and the rear side in the main scanning direction (left side in the figure) (Right side of), the number m of reference pixels in each of the sections B up to the reference pixels is determined, and the pixel interval is calculated separately in each of the sections A and B by calculating [4-m]. . Then, for example, FIG.
If only one thinned pixel exists in the section B as shown in FIG. 5, "4" is obtained as the pixel interval in the section A, and "3" is obtained as the pixel interval in the section B. The target pixel and the reference pixel having the positional relationship shown in FIG. When the number of thinned pixels differs in each section as described above, the thinning circuit 5 performs thinning in a predetermined pattern. Therefore, thinning in each section is performed based on this thinning pattern. The number of pixels can be easily obtained. When the processing of the pixel interval calculation circuit 7 is changed in this way, the processing of the pixel interval calculation circuit 7 becomes slightly complicated. However, even in the case of the following image enlargement ratio, the contrast using the optimal reference pixel is always used. Emphasis can be made. The present invention is not limited to the above embodiment. For example, in the above-described embodiment, the enlargement process in the sub-scanning direction is performed based on the reading cycle of the line sensor 1 b and the original 20.
Is performed by changing the relative relationship with the transport speed, but the processing by linear interpolation can be used for the enlargement processing in the sub-scanning direction as in the main scanning direction. In such a case, the reference pixels arranged in the sub-scanning direction should be determined in the same manner as in the above embodiment. In the above embodiment, the image enlargement ratio by the enlargement means is four times, but this magnification may be arbitrary (however, an integer). When the image enlargement ratio by the enlargement unit is 4 times, the final image enlargement ratio is 100% to 400%.
%, The final maximum value of the image enlargement ratio can be increased by increasing the image enlargement ratio by the enlargement means. Conversely, when the final image enlargement ratio may be smaller than 400%, the image enlargement ratio by the enlargement unit may be reduced. When the image enlargement ratio is changed by the enlargement means, it is necessary to change the processing of the pixel interval calculation circuit 7 in accordance with the magnification. That is, if the image enlargement ratio by the enlargement unit is n times, the pixel interval becomes
[Nm (m is the number of thinned pixels)]. In the above embodiment, linear interpolation is performed for image enlargement, and Laplacian operation is performed for contrast enhancement. However, other specific methods may be used. In the above embodiment, the image processing apparatus according to this embodiment is applied to an image reading apparatus. However, the image processing apparatus can be realized as an independent image processing apparatus for processing image data generated by another apparatus. Needless to say, the present invention can be applied to an apparatus other than the image reading apparatus. In addition, various modifications can be made without departing from the spirit of the present invention. [0046] According to the present invention can provide an image processing apparatus which can accurately generate a faithful image data to an image obtained by enlarging the processed target image at an arbitrary magnification.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a functional block diagram showing a main configuration of an image reading apparatus configured by applying an image processing apparatus according to an embodiment of the present invention. FIG. 2 is a view for explaining an example of a change in data in the image reading apparatus shown in FIG. FIG. 3 is a diagram illustrating a positional relationship between a target pixel and a reference pixel when contrast enhancement is performed. FIG. 4 is a diagram schematically showing the contents of a table showing a relationship between an image enlargement ratio and a pixel interval. FIG. 5 is a diagram showing a modification of the positional relationship between a target pixel and a reference pixel when contrast enhancement is performed. FIG. 6 is a diagram illustrating a conventional image enlargement process. FIG. 7 is a diagram showing a positional relationship between a target pixel and a reference pixel in conventional contrast enhancement. FIG. 8 is a view for explaining an example of a change in data in the conventional image processing apparatus. [Description of Signs] 1 ... Reading unit 2 ... A / D converter 3 ... Line memory 4 ... 4 times circuit 5 ... Thinning circuit 6 ... Thinning timing generation circuit 7 ... Pixel interval calculation circuit 8 ... Memory unit 9 ... Address generation circuit 10 Multiplication circuit 11 Coefficient generation circuit 12 Addition circuit

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-63-311866 (JP, A) JP-A-3-64166 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 1/38-1/393 G06T 3/40

Claims (1)

(57) [Claims 1] Original image data composed of an array of a large number of pixel data in which the density of each corresponding pixel is represented by a multi-value is
An image processing apparatus for converting an image corresponding to the original image data into enlarged image data corresponding to an image enlarged at an arbitrary enlargement ratio, wherein a predetermined interpolation process is performed between each pixel data of the original image data. One (n is a predetermined integer) pixel data is inserted, and n of the image corresponding to the original image data is inserted.
Enlarging means for generating n-fold image data corresponding to a doubled image; thinning means for generating magnified image data by thinning pixel data from the n-fold image data generated by the enlarging means in accordance with the enlargement ratio And each pixel data in the n-times image data as a pixel of interest,
And line Cormorant thinned pixel number finding means a process for obtaining the number m of pixels to be thinned out by the thinning means among pixels located between the pixels in the target pixel and spaced n pixels from the target pixel position, Each image of the enlarged image data generated by the thinning means
The raw data and the target pixel, a predetermined contrast strong tone processing on the basis of the level of the target pixel data and the focused image iodination et n-m-level image containing data position Ru near distant pixels
The image processing apparatus characterized by was and a row intends enhancement means.