JP4700597B2 - Image binarization apparatus and image binarization method - Google Patents

Description

  The present invention relates to an image binarization apparatus and an image binarization method, and more specifically, an image that is binarized by removing luminance unevenness and shadows of a multi-valued image input by an image input apparatus or the like having a non-constant light source The present invention relates to a binarization apparatus and an image binarization method.

Conventionally, when a document is digitized and stored, a scanner or a scanner unit of a copying machine or a facsimile machine has been used exclusively as an image input device. Such a scanner unit is provided with a light source in the apparatus, the light emitted from the light source is applied to a document, the reflected light is read by a CCD or the like, and the image is binarized and stored as necessary. In such a case, since the light source and the optical system are constant, the occurrence of uneven brightness and shadows in the image due to the light source and the optical system are constant, so correction is easy and a high-quality image is output. be able to. Therefore, it was easy to binarize such a digital image with high quality.
On the other hand, in recent years, video cameras and digital cameras have been developed, and there is an increasing demand for character recognition of images input from such devices. In particular, digital cameras are remarkably high in pixel size and downsized, and have begun to be used for various purposes as portable information collection tools. For example, text information such as documents, signboards, and advertisements can be stored as a binary image because sufficient information can be obtained with a binary image and the storage capacity required for storage is smaller than that of a multi-value image. It is. In addition, the binarized image can be transmitted by facsimile, character recognition, and reuse.

In order to appropriately binarize such a digital image, Patent Document 1 discloses that for each pixel in the window, the difference between the brightness of each pixel and the brightness of a specific pixel is proportional to the contrast of the focused image portion. A binarization circuit unit that binarizes the image in the window by comparing the parameters, and a determination circuit that determines whether the binary pattern obtained by the binarization circuit unit is appropriate as an image pattern of the contour portion; A technique for obtaining a high-quality binarized image by providing the above is disclosed.
In Patent Document 2, a histogram of luminance values is created from a multi-valued image, a white pixel representative value and a black pixel representative value are determined from the histogram, a binarization threshold is determined from the average, and binarization is performed. A technique for binarizing a multi-valued image based on a threshold value, and dividing the image into blocks, obtaining a binarization threshold value or white pixel representative value / black pixel representative value for each block, and interpolating blocks without characters from surroundings However, a technique for removing luminance unevenness and shadows in an image by using a binarization threshold value of a block as a threshold value for each pixel is disclosed, and high-quality binarization can be performed.
Furthermore, in Patent Document 3, it is possible to obtain a binary image with higher quality than the above-described conventional technique at a lower cost by performing binarization of an image using a threshold value corresponding to the local brightness of the image. A possible image binarization apparatus is disclosed.
JP-A-3-237571 Japanese Patent Laid-Open No. 7-212591 JP 2002-354248 A

However, the conventional technique has the following problems. In other words, by obtaining a histogram for each local area of the image, calculating the average brightness of the image, and further calculating the binarization threshold locally, that is, by calculating a different value for each pixel. High-quality binarized images are obtained, however, the binarization is costly or time-consuming compared to the original binarization method. Especially when scanning a document centered on characters and inputting images, there are many partial areas that do not have information only in the background, and it is useless to binarize these areas as well, and at low cost and high speed. The binarization process could not be performed.
The present invention has been made in view of such problems, and provides an image binarization apparatus and an image binarization method that realize high-quality binarization processing equivalent to that of the prior art at higher speed and lower cost. The purpose is to do.

In order to solve this problem, the present invention provides a first block setting means for inputting a multi-value image and setting a block including a plurality of pixels for the multi-value image, and the first block. Block binarization threshold value calculation means for calculating a block binarization threshold value for each block in the first block set by the setting means based on the luminance average value, and the multivalued image is less than the first block compares the second block setting means for setting a second block including the pixel, the luminance values of the pixels of the second block set by said second block setting means and the said block binarization threshold , when the luminance value is small, attention to which the second block and attention to that the present binary to a plurality of second blocks adjacent to the second block is set to the binarization performed block Execution block And setting means, in the present binarization execution block set by the main binarization execution block setting means, and the luminance value of each pixel, block binarization of the plurality of first blocks surrounding the pixel And a binarization unit that performs binarization by comparing with a pixel threshold calculated based on the threshold.
According to a second aspect of the present invention, the block binarization threshold value calculation means is a sampling means for sampling a pixel from the input first block, and for excluding the pixel from a luminance average calculation target when the pixel has a low luminance value. A low luminance exclusion unit; an average luminance calculation unit that calculates an average luminance; and a block binarization threshold calculation unit that calculates a block binarization threshold by a predetermined calculation from the average luminance. .

Claim 3, wherein the binarization execution block setting means includes sampling means for sampling the pixels from the second block input block binarization of obtaining the pixel in the block binarized threshold calculating means Pre-binarization means for binarization using a threshold, and the second block of interest and the second of interest when there is a pixel determined to be black by the pre-binarization means And a binarization execution block determination unit that determines a plurality of second blocks adjacent to the block as a binarization execution block.
According to a fourth aspect of the present invention, the binarization unit selects a block that is determined to be the binarization execution block from the input second blocks, and a block binarization threshold value. A pixel threshold interpolating means for interpolating a pixel threshold at a position of a pixel to be binarized, and a binarizing means for binarizing all the pixels of the second block with the pixel threshold. .
Claim 5, wherein the binarizing means receives the second block, the two and binarization block selection means for the value of the execution block selectively outputs the selected second block a pixel threshold interpolation means for calculating a binarization threshold value for the pixel block from the block binarization thresholds corresponding to the plurality of first blocks of the peripheral pixel blocks comprising a plurality of pixels of all pixels of the second block And binarizing means for binarizing with a binarization threshold value calculated with the pixel block as a unit.

The present invention provides a first block setting step for inputting a multi-valued image and setting a block including a plurality of pixels for the multi-valued image, and a first block setting step set in the first block setting step. A block binarization threshold value calculating step for calculating a block binarization threshold value for each block based on a luminance average value; and a second block including fewer pixels than the first block for the multi-valued image. The second block setting step and the luminance value of the pixel in the second block set by the second block setting step are compared with the block binarization threshold, and attention is paid when the luminance value is small. and that said second block and attention to that the second of the binarization execution block setting step of setting a plurality of second blocks in the present binarization execution blocks adjacent to the block and This in binarization execution block set by the main binarization execution block setting means, and the luminance value of each pixel, based on the block binarization threshold of the plurality of first blocks surrounding the pixel And a binarization step for performing binarization by comparing with the calculated pixel threshold value.
According to a seventh aspect of the present invention, the block binarization threshold value calculating step includes: a sampling step of sampling pixels from the input first block; and a case where the pixels have a low luminance value, so as to be excluded from a luminance average calculation target A low luminance exclusion step; a step of calculating an average luminance; and a step of calculating a block binarization threshold value by a predetermined calculation from the average luminance.

Claim 8, wherein the binarization execution block setting step includes a sampling step of sampling the pixels from the second block input block binarization of obtaining the pixel in the block binarized threshold calculating step A pre-binarization step for binarization using a threshold value, and a second block of interest and a second of interest when there are pixels determined to be black by the pre-binarization step A binarization execution block determination step for determining a plurality of second blocks adjacent to the block as a binarization execution block.
The present binarization step includes a binarization block selection step of selecting only a block determined to be the present binarization execution block from the input second block, and a block binarization threshold value A pixel threshold interpolation step for interpolating a pixel threshold at a position of a pixel to be binarized, and a binarization step for binarizing all the pixels of the second block with the pixel threshold.
Claim 10, wherein the binarization step is to enter the second block, the two and binarization block selection step the value of the execution block selectively outputs the selected second block a pixel threshold interpolation calculating a binarization threshold value for the pixel block from the block binarization thresholds corresponding to the plurality of first blocks of the peripheral pixel blocks comprising a plurality of pixels of all pixels of the second block And binarizing means for binarizing with a binarization threshold value calculated with the pixel block as a unit.

  According to the present invention, a first block setting unit that inputs a multi-value image and sets a block including a plurality of pixels for the multi-value image, and a first block setting unit that is set by the first block setting unit. Block binarization threshold value calculation means for calculating a block binarization threshold value for each block based on the luminance average value, and second block setting for setting a block including fewer pixels than the first block for a multi-valued image And the luminance value of the pixel in the second block set by the second block setting means and the block binarization threshold value, and if the luminance value is small, the second block and the relevant attention The binarization execution block setting means for setting a block in the same group as the block as the binarization execution block, and the binarization execution block set by the binarization execution block setting means. The binarization means for performing binarization by comparing the luminance value of each pixel with the pixel threshold obtained by interpolating the block binarization threshold in the image, so that high-quality image binarization is provided. Can be executed at high speed and low cost.

Hereinafter, the present invention will be described in detail with reference to embodiments shown in the drawings. However, the components, types, combinations, shapes, relative arrangements, and the like described in this embodiment are merely illustrative examples and not intended to limit the scope of the present invention only unless otherwise specified. .
FIG. 1 is a block diagram showing an image binarization apparatus of the present invention. This image binarization apparatus 100 receives a multi-value image 1 and sets a first block setting device (first block setting means) 2 for setting a block including a plurality of pixels for the multi-value image 1; A first block binarization threshold value calculator (block binarization threshold value calculation means) 3 for calculating a block binarization threshold value for each block in the first block set by the block setting device 2 based on the average luminance value; A second block setting device (second block setting means) 4 for setting a block including fewer pixels than the first block for the multi-valued image, and the second block setting device 4 When the luminance value is small, the second block and a block in the vicinity of the same group as the target block are set as the binarization execution block. Book In the binarization execution block set by the binarization execution block setting device (the present binarization execution block setting means) 5 and the binarization execution block setting device 5, the luminance value of each pixel and the block binarization This binarization processor (the present binarization means) 6 that executes binarization by comparing the pixel threshold value obtained by interpolating the threshold value is configured.

First, an 8-bit grayscale multilevel image 1 is input. When the input image is a color image, a luminance signal is extracted in advance and converted into an 8-bit grayscale image. This image is input to the first block setting device 2. The first block setting device 2 divides the image into blocks of 128 × 128 pixels as shown in FIG. The first block binarization threshold value calculator 3 receives each first block as an input, and first extracts a pixel position and a value for performing a subsequent calculation by the sampler.
As shown in FIG. 3, sampling is performed every 8 pixels in the horizontal and vertical directions. The sampled pixel value is compared with a low luminance exclusion threshold by a comparator.

FIG. 4 is a block diagram showing the internal configuration of the first block binarization threshold value calculator. The first block binarization threshold value calculator 3 excludes a sampling unit (sampling means) 11 that samples pixels from the input first block 10 and a luminance average calculation target when the pixel has a low luminance value. A low-luminance exclusion means comprising a comparator 12 and a low-luminance exclusion threshold memory 16, an average luminance calculator (average luminance calculation means) 13 for calculating average luminance, and a first block by a predetermined calculation from the average luminance Block binarization threshold value calculating means for calculating the binarization threshold value 15, and if the luminance value of the pixel is higher than the low luminance exclusion threshold value, the luminance value is output to the average luminance calculator 13 and average luminance Is not output if it is less than that. This is a process performed in order to binarize a substantially black solid area into black correctly in a later binarization process. The average luminance calculator 13 adds pixel values that are targets of average luminance calculation, and calculates the average value. If there is no target pixel for calculating the average luminance in the first block, the average luminance calculator 13 outputs a value obtained by dividing the low luminance exclusion threshold by the coefficient stored in the coefficient memory 18. This means that an output (= first block binarization threshold) of a multiplier 14 described later becomes a value equal to the low luminance exclusion threshold. The multiplier 14 multiplies the average luminance calculated by the average luminance calculator 13 by a positive coefficient less than 1 stored in the coefficient memory 18 to calculate the first block binarization threshold 15.
The second block setting device 4 divides the image into second blocks of 16 × 16 pixel units as shown in FIG. The reason why the division unit is smaller than the first block is that in a document with many characters, there are many portions that are fine and black, such as line spacing and characters, so only the portion with black is binarized by reducing the block size. This is because the efficiency is improved.

  FIG. 6 is a block diagram showing an internal configuration of the binarization execution block setting device. This binarization execution block setting device includes a sampling device (sampling means) 21 that samples pixels from the input second block 20, and a block binary value obtained by the pixels using the first block binarization threshold calculator 3. A pre-binarizer (pre-binarization means) 22 for binarization using a binarization threshold, and a second block and its neighborhood second when there are pixels determined to be black by the pre-binarizer 22 A binarization execution block determination unit (main binarization execution block determination means) 23 that determines a block as a main binarization execution block, and inputs a second block of 16 × 16 pixel units. It is determined whether or not there is a pixel that becomes black after binarization. If there is a pixel that becomes black, the second block and the surrounding second block are set as the binarization execution block. This will be specifically described with reference to FIG. First, a 16 × 16 second block is input, and a pixel is sampled by a sampler. This works in the same way as the sampler 11 of FIG. Next, the pre-binarizer 22 binarizes the sampled pixels. The pixel is binarized by comparing the first block binarization threshold of the first block including the second block with the luminance value of the sampled pixel. If the luminance value of the pixel is smaller than the first block binarization threshold, the pixel is black, otherwise it is white. Since there are four sampled pixels in the second block, the pre-binarizer 22 outputs the black and white of these four pixels. The binarization execution block decision unit 23 receives the black and white of the four pixels, and if one or more of them are present, the second block b (i, j) of interest as shown in FIG. To perform this binarization in the three neighboring second blocks (upper left b (i-1, j-1), upper b (i, j-1), left b (i-1, j)) The flag of the binarization execution block is set and stored in the binarization execution block position memory.

  FIG. 8 is a block diagram showing the internal configuration of the binarization processor. The present binarization processor 6 selects a binarization block selector (binarization block selection means) 34 for selecting only a block determined to be the present binarization execution block from the input second block 33. A pixel threshold interpolator (pixel threshold interpolation means) 32 for interpolating the pixel threshold at the position of the pixel to be binarized from the block binarization threshold, and binarizing all the pixels of the second block 33 with the pixel threshold. The binarization processing unit 6 executes binarization processing only on the second block for which the flag of the binarization execution block is set. The binarization block selector 34 determines whether or not the input second block is the binarization execution block based on the flag of the binarization execution block. When it is determined that this is the binarization execution block, the binarizer 35 binarizes each pixel of the second block using the binarization threshold of the pixel unit interpolated by the pixel threshold interpolator 32. Turn into. If the luminance value of each pixel is smaller than the interpolated threshold, it is determined to be black (1), otherwise it is determined to be white (0), and black information is stored in the binary image memory that has been set to all white (all 0) in advance. Write out.

FIG. 9 is a diagram for explaining the detailed operation of the pixel threshold interpolator. FIG. 9A is a diagram showing four first blocks and a region in which a binarization threshold value in units of pixels is interpolated. The binarized threshold value is calculated from the first block binarization threshold values a, b, c, and d of each block in the area where the hatched portion 37 interpolates the binarization threshold value in pixel units.
Next, an interpolation method will be described based on FIG. The size (pixel unit) of the interpolation area is defined as a horizontal direction xbnum and a vertical direction ybnum. If the pixel position is defined as (m, l) with the upper left in the shaded area as the origin a,
leftth = (a * (ybnum-l) + c * l) / ybnum
rightth = (b * (ybnum-l) + d * l) / ybnum
The binarization threshold th (m, l) of the pixel is determined from leftth and rightth.
th (m, l) = (leftth * (xbnum-m) + rightth * m) / xbnum
In the case of the edge of the input image, four block binarization threshold values cannot be obtained, and only one or two can be obtained. In this case, the above-described interpolation formula is used by substituting and using the obtained first block binarization threshold in a place where it has not been obtained. For example, when obtaining the binarization threshold value at the upper left corner of the image, the obtained first block binarization threshold value is only d, and a, b, and c are not obtained. Therefore, the values of a, b and c are substituted with the value of d.

  FIG. 10 is a flowchart for realizing the binarization method by software. Details of the flowchart of FIG. 10 are shown in FIGS. 13 to 15 and will be described. First, in the first block setting step S1, the multi-valued image is divided into blocks of 128 × 128 pixels as shown in FIG.

FIG. 11 is a diagram showing another embodiment of the binarization execution block setting device of the present invention. The same components will be described with the same reference numerals as in FIG. The difference from this binarization execution block setting device is that the binarization threshold value used in the previous binarization device 22 is not the block binarization threshold value, but the pixel threshold value interpolator 26 interpolates pixel by pixel. This is the point. First, the second block 20 is input, and the sampler 21 performs sampling every 8 pixels in the horizontal and vertical directions. The pixel threshold interpolator 26 reads the binarization threshold values of the first block including the sampled pixels and the first block in the vicinity thereof from the first block binarization threshold value memory 25, and the pixel threshold value at the position of the sampled pixels Is interpolated. The previous binarizer 22 performs binarization by comparing the interpolated threshold value and the luminance value of the sampled pixel, and determines whether or not the second block is the binarization execution block. The execution block determination unit 23 makes the determination. The pixel interpolation procedure of the pixel threshold interpolator 26 used here is the same as that of the pixel threshold interpolator 32 shown in FIG.
Compared with FIG. 6, the cost is slightly increased because the pixel threshold value is obtained by interpolation, but the previous binarizer 22 can perform binarization with higher quality, and this binarization execution block determination Can be performed with higher accuracy.

  FIG. 12 is a flowchart for explaining the binarization execution block setting step. FIG. 12 differs from FIG. 14 in that the binarization threshold used in the previous binarization step S43 is not the first block binarization threshold but the binarization threshold interpolated in pixel units. Differences from FIG. 14 will be described. A pixel threshold interpolation step S42 is inserted immediately before the previous binarization step S43. The pixel threshold interpolation step S42 interpolates the pixel threshold at the position of the sampled pixel using the binarization threshold of the first block to which the sampled pixel belongs and the blocks in the vicinity thereof. Pre-binarization is performed by comparing the interpolated threshold value and the luminance value of the sampled pixel. The pixel threshold interpolation procedure used here interpolates the pixel threshold in the same procedure as the pixel threshold interpolation step S32 shown in FIG.

  FIG. 13 is a flowchart for explaining the first block binarization threshold value calculation step. In the first block selection step S11, a target block for calculating the first block binarization threshold is selected. The selected block extracts a pixel position and a value for performing calculation for obtaining an average value of the luminance values of the pixels in the sampling step S12. As shown in FIG. 3, sampling is performed every 8 pixels in the horizontal and vertical directions. The sampled pixel value is compared with a low luminance exclusion threshold value in a low luminance exclusion step S13. If the pixel value is higher than the low-luminance exclusion threshold, it is output to the next average luminance calculation step S14 to be the target for calculating the average luminance. In the average luminance calculation step S14, pixel values to be subjected to average luminance calculation are added, and the average value is calculated. If there is no target pixel for calculating the average luminance in the block, a predetermined value is output so that the output of the first block binarization threshold calculation step S15 described later is equal to the low luminance exclusion threshold. . The first block binarization threshold calculation step calculates the block binarization threshold by multiplying the average luminance calculated by the average luminance calculation step by a positive coefficient less than 1.

  FIG. 14 is a flowchart for explaining the binarization execution block setting step. In the second block setting step, the image is divided into blocks of 16 × 16 pixel units as shown in FIG. In the binarization execution block setting step, a block for executing the binarization is determined from the entire second block. In the second block selection step S20, a second block for pre-binarization is selected. The selected second block samples the pixels every several pixels (for example, 4 pixels) in the sampling step S21, and the luminance of the selected pixel position is the first pixel to which the pixel position belongs in the previous binarization step S22. It is compared with the first block binarization threshold of the block, and if the luminance value of the pixel is larger than the block binarization threshold, it is determined to be white, otherwise it is determined to be black. If one or more sampled pixels in the selected second block are one or more black pixels, in the present binarization flag setting step S24, the second block and the neighboring blocks that are subject to the binarization are selected. Set the value flag. As shown in FIG. 7, the neighboring block means three blocks on the upper left, upper, and left when the lower right block is the second block. Therefore, the four blocks of the block and the neighboring three blocks are determined as the binarization execution block.

FIG. 15 is a flowchart for explaining the binarization processing step. The second block is selected in the second block selection step S30, and it is determined whether or not the block has the present binarization flag (S31). When it is determined that the <binarization flag> is present, the pixel threshold at the pixel position is interpolated from the first block binarization threshold in the pixel threshold interpolation step S32. Otherwise, select the next block. This interpolation method is equivalent to that already described. Thereafter, the pixel is binarized in this binarization step S33 using the interpolated pixel threshold. If the luminance value of the pixel is larger than the pixel threshold value, it is determined to be white, otherwise it is determined to be black. This binarization process is performed on all pixels and all blocks, and the process ends.
As in all the embodiments, the binarization threshold value is interpolated in units of pixels during the binarization. If this is calculated not in units of pixels but in units of smaller blocks such as 2 × 2 pixels, the amount of calculation is reduced, and further speeding up of binarization can be achieved.

It is a block diagram showing the image binarization apparatus of this invention. It is the figure which divided | segmented the image into the block of a 128x128 pixel unit. It is a figure which samples every 8 pixels horizontally and vertically. It is a block diagram showing the internal structure of a 1st block binarization threshold value calculator. It is the figure which divided | segmented the image into the 2nd block of a 16x16 pixel unit. It is a block diagram showing the internal structure of this binarization execution block setting device. It is a figure explaining this binarization execution block determination. It is a block diagram showing the internal structure of this binarization processor. It is a figure explaining detailed operation | movement of a pixel threshold value interpolator. It is a flowchart when implement | achieving this binarization method by software. It is a figure which shows the other Example of this binarization execution block setting device of this invention. It is a flowchart explaining this binarization execution block setting step. It is a flowchart explaining a 1st block binarization threshold value calculation step. It is a flowchart explaining this binarization execution block setting step. It is a flowchart explaining this binarization process step.

Explanation of symbols

  1 multi-valued image, 2 first block setter, 3 first block binarization threshold calculator, 4 second block setter, 5 binarization execution block setter, 6 binarization processor, 7 2 Binary image, 100 image binarization device

Claims (10)

A first block setting means for inputting a multi-value image and setting a block including a plurality of pixels for the multi-value image;
Block binarization threshold value calculating means for calculating a block binarization threshold value for each block in the first block set by the first block setting means based on the luminance average value;
A second block setting means for setting a second block containing fewer pixels than the first block with respect to the multi-valued image,
Comparing the brightness value of the pixel of the second block set by said second block setting means and the said block binarization threshold, if the luminance value is small, the second block of interest and the binarization execution block setting means in which a plurality of the second block is set to the binarization execution block and adjacent to the second block of interest,
This in binarization execution block set by the main binarization execution block setting means, and the luminance value of each pixel, based on the block binarization threshold of the plurality of first blocks surrounding the pixel An image binarization apparatus comprising: a binarization unit that performs binarization by comparing with a calculated pixel threshold value.   The block binarization threshold value calculation means includes sampling means for sampling pixels from the input first block, and low luminance exclusion means for excluding them from the target of luminance average calculation when the pixels have low luminance values. The average luminance calculation means for calculating the average luminance, and the block binarization threshold value calculation means for calculating a block binarization threshold value by a predetermined calculation from the average luminance. Image binarization device. The present binarization execution block setting means, two with a sampling means for sampling the pixels from the second block input, the block binarization threshold determined the pixel in the block binarized threshold calculating means before binarizing means for binarizing a plurality of adjacent second block being the second block and interest are the focus when there are pixels determined as black by front binarizing means image binarization apparatus of claim 1, the second the block and the binarization execution blocks and determining the binarizing performed block determination unit, characterized by comprising a. The present binarizing means, said a binarization block selecting means for selecting only the blocks determined as the binarization execution block from the second block input, all the pixels of the second block The image binarization apparatus according to claim 1, further comprising: binarization means for binarizing with a pixel threshold value. The present binarizing means receives the second block, the binarization block selecting means for selecting and outputting the binarized execution block, a plurality of pixels of the second block of the selected a pixel threshold interpolation means for calculating a binarization threshold value for the pixel block from the block binarization thresholds corresponding to the plurality of first blocks of the peripheral pixel blocks comprising, the pixel block of all the pixels of the second block The image binarization apparatus according to claim 1, further comprising: binarization means for binarizing with a binarization threshold calculated as a unit. A first block setting step for inputting a multi-value image and setting a block including a plurality of pixels for the multi-value image;
A block binarization threshold value calculating step for calculating a block binarization threshold value for each block in the first block set by the first block setting step based on the luminance average value;
A second block setting step of setting a second block containing fewer pixels than the first block with respect to the multi-valued image,
Comparing the brightness value of the pixel of the second block set by said second block setting step and the said block binarization threshold, if the luminance value is small, the second block of interest and the binarization execution block setting step in which a plurality of the second block is set to the binarization execution block and adjacent to the second block of interest,
This in binarization execution block set by the main binarization execution block setting means, and the luminance value of each pixel, based on the block binarization threshold of the plurality of first blocks surrounding the pixel An image binarization method comprising: a binarization step for performing binarization by comparing with a calculated pixel threshold value.   The block binarization threshold value calculating step includes a sampling step for sampling pixels from the input first block, and a low luminance exclusion step for excluding from the target of luminance average calculation when the pixel has a low luminance value; The image binarization method according to claim 6, further comprising: calculating an average luminance, and calculating a block binarization threshold value by a predetermined calculation from the average luminance. The present binarization execution block setting step, two with a sampling step of sampling the pixels from the second block input, the block binarization threshold determined the pixel in the block binarized threshold calculating step a binarization step prior to digitizing, a plurality adjacent to the second block being the second block and interest are the focus when there are pixels determined as black by the front binarization step image binarization method of claim 6, characterized in that it comprises second and the binarization execution block determining step determines that the binarization execution block blocks, a.   The binarization step should be binarized from a binarization block selection step for selecting only a block determined to be the binarization execution block from the input second block, and a block binarization threshold. The pixel threshold interpolation step for interpolating the pixel threshold at the pixel position, and the binarization step for binarizing all the pixels in the second block with the pixel threshold. Image binarization method. The present binarization step is to enter the second block, the binarization block selecting step of selecting and outputting the binarized execution block, a plurality of pixels of the second block of the selected a pixel threshold interpolation calculating a binarization threshold value for the pixel block from the block binarization thresholds corresponding to the plurality of first blocks of the peripheral pixel blocks comprising, the pixel block of all the pixels of the second block The image binarization method according to claim 6, further comprising: binarization means for binarizing with a binarization threshold calculated as a unit.

Images