JPH11112796A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To precisely identify the area for each pixel read from an original. SOLUTION: Image reading from the original is executed by dividing into pre-scanning and main scanning. The read result of pre-scanning is made into line partial data of each main scanning direction and its featured value or an identifying result sorted based on the featured value is stored in a page memory 5a. In main scanning, image data on the pixel of a local block provided in the periphery of a pixel under consideration is stored in a block memory 5b and a local calculating means 12 calculates local featured values. While referring to the partial featured value or the identifying result of sorting information stored in the page memory 5a to take into account as wide information, an area identification processing means 13 identifies an area the pixel under consideration belongs to based on the local feature value from the means 12.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital copying machine, a facsimile, and the like for improving the quality of a recorded image.
The present invention relates to an image processing apparatus that performs processing according to characteristics of an image obtained by scanning a document.

[0002]

2. Description of the Related Art Conventionally, in an image processing apparatus such as a digital copying machine or a facsimile apparatus, various image identification methods have been used to improve image quality. As a typical image identification processing method, there are known a method of dividing an image into blocks and performing pattern matching in each block, and a method of identifying each region using a characteristic parameter representing a property of a character or halftone image. Have been. In the method using pattern matching, it is necessary to prepare a large number of patterns, and there is a problem that a required storage capacity of a memory is enormous and versatility is poor. For this reason, a method using a feature parameter is often adopted.

[0003] Prior art using feature parameters is disclosed in, for example, JP-A-61-194968 and JP-A-62-1478.
60 and the like. JP-A-61-194968
In the "halftone dot discrimination method", the change in the signal level of two spatially continuous pixels is separately measured when the two pixels are continuous in the main scanning direction and when the two pixels are continuous in the sub-scanning direction. . The sum of the measured amounts in the respective scanning directions is compared with a predetermined value for each block of a plurality of pixels, and a halftone dot region is identified based on the comparison result. JP-A-62-1478
In 60 "halftone facsimile signal processing method", a difference between a maximum signal level and a minimum signal level in a block of a plurality of pixels is obtained, and the value of the difference is compared with a predetermined setting value. If the level difference is smaller than the set value, it is determined that the signal level change including the photograph portion is gentle. If the number of changes between the signal levels of two spatially consecutive pixels in the block is larger than a predetermined value in accordance with a predetermined access order, the area is determined to be a halftone dot portion. According to the determination result, signal processing corresponding to each area is performed.

[0004]

SUMMARY OF THE INVENTION In image processing, 1
Focusing on one pixel, the image identification method that uses the feature parameters of the local block when identifying the type of the region to which the pixel of interest belongs There is a problem that is likely to occur. In particular, when the classification process is completely performed for each pixel and the filtering process or the like is performed in accordance with the region identification result using a prepared filter or the like, the adverse effect on the image quality caused by the filtering process performed on the isolated erroneously identified region increases. I will.

In order to solve this problem, a method has been considered in which the identification results of adjacent pixels are stored and used for identifying the region of the pixel of interest. However, this method has a problem that incorrect identification information is sequentially propagated when the adjacent pixel to be referred to is also incorrectly identified. Further, it is not suitable for identifying an area where a large regular change in the period that cannot be accommodated in the local block occurs. The reason is that, in the case of an image in which the area changes periodically, the periodicity cannot be determined unless the identification processing is performed within a certain size range. In particular, this problem becomes remarkable when the period is larger than the local block used for identification.

An object of the present invention is to provide an image processing apparatus capable of identifying a regular region having a large cycle that does not fit in a local block.

[0007]

According to the present invention, an image signal obtained by scanning a document two-dimensionally in a main scanning direction and a sub-scanning direction is such that each pixel is a character area, a photograph area or a halftone area. In an image processing apparatus that performs processing by identifying which area the image belongs to, a control unit that controls scanning of a document so as to be divided into a pre-scan and a main scan. Using the line image data, a partial calculation unit for obtaining a partial feature amount representing the characteristic of the partial region of each line, and classifying the type of the region to which the partial region of each line belongs from the partial feature amount obtained by the partial calculation unit Recording means for recording the calculation result of the partial feature amount obtained by the partial calculation means or the identification result obtained by the classification means,
A storage unit for storing image data of a local block including a pixel of interest and a plurality of pixels in the vicinity thereof, the pixel being a pixel selected one by one from the image signal at the time of the main scan, and image data stored in the storage unit From the character area,
A local operation unit for obtaining a plurality of local feature amounts representing the characteristics of each region of the photograph region or the halftone dot region; a partial feature amount or identification result recorded in the recording unit; and a target pixel obtained by the local operation unit. Using the local feature of
Identify the region to which the pixel of interest belongs in the local block,
An image processing apparatus comprising: an area identification processing unit that performs a predetermined process according to a type of an area to be identified.

According to the present invention, a document having a character area, a photograph area, or a halftone area is divided into two parts, a pre-scan and a main scan, by a control means in a main scanning direction and a sub-scanning direction. Perform two-dimensional scanning. At the time of pre-scanning, a line area composed of a plurality of images continuous in the main scanning direction is set using line image data in the main scanning direction, and a character area, a photograph area, or a halftone area to which the line area belongs. Is obtained by the partial calculation means. Also, the type of each line partial area is identified by the classifying means from the partial feature amount. At this time, the length of the line partial area used in the calculation of the partial feature amount in the main scanning direction is set to be sufficiently longer than the width of the local block handled in the area separation for each pixel during the main scan. For example, it is possible to identify a feature of a region having a long cycle that cannot be identified only by the local block. The partial feature amount or the identification result calculated at the time of the pre-scan is recorded in the recording means. At the time of separation of each pixel at the time of the main scan, by using the partial feature amount or the identification result recorded in the recording means, it is possible to cope with an area having a long cycle in the main scanning direction, and Since the local feature amount from the block is also calculated, the local feature can be easily identified.

In order to enhance the identification accuracy of image identification, it is necessary that a local feature calculated from a local block including a pixel of interest and a plurality of pixels in the vicinity thereof sufficiently indicates the properties of the region. . Actually, it is difficult to know the global properties of an image only with identification information of a local block. Therefore, at the time of pre-scanning, images are broadly classified in advance by using line data that is wider in the main scanning direction than local blocks. At the time of identification for each pixel, a partial feature amount or a classification information, which is a feature amount for each line partial region at the time of pre-scan, is added to a local feature amount which is a feature amount of a local block at the time of main scan. , Identification accuracy can be improved.

Further, in the present invention, the partial calculation means calculates a partial feature quantity representing the characteristics of the character area, photograph area or halftone area using line partial read data having a plurality of pixel widths in the main scanning direction. And correcting means for correcting the partial feature amount or the identification result recorded in the recording means in units of blocks using a plurality of lines in the sub-scanning direction before performing the main scan, and re-recording the correction result in the recording means. The area identification processing means outputs the area separation information of the pixel of interest from the partial feature amount or the identification result re-recorded in the recording means and the local feature amount.

According to the present invention, the partial calculation means calculates a partial feature amount representing the characteristics of a character area, a photograph area, or a halftone dot area using line partial read data having a plurality of pixel widths in the main scanning direction. . The correction unit corrects the partial feature amount or the identification result recorded in the recording unit in units of blocks using a plurality of lines in the sub-scanning direction before execution at the time of the main scan, and re-records the data in the recording unit. Since the area identification processing unit adds the correction result by the correction unit to the area identification processing at the time of the main scan, the area identification processing unit performs more global identification as compared with the case where the one-line partial feature amount or the classification information is used. Accuracy can be improved.

Further, in the present invention, the area identification processing means uses the plurality of lines before and after the line to which the pixel of interest belongs to the partial feature amount or the identification result recorded in the recording means at the time of the main scan. Each pixel of interest is a character area,
It is characterized in that the region belongs to a photograph region or a halftone dot region, and region identification information is output.

According to the present invention, the area identification processing means uses the partial feature amount or the identification result recorded in the recording means for a plurality of lines before and after the line to which the pixel of interest belongs, so that the pixel of interest during the main scan is used. Since the region belongs to the character region, the photograph region, or the halftone region, the region can be globally identified, and highly accurate region identification information can be obtained.

Further, in the present invention, the partial calculation means includes a binarization means for binarizing the line partial image data in the main scanning direction and a binarized pixel sequence obtained from the binarization means, Adding means for adding the difference in the number of pixels between pixel rows having the same value adjacent to each other across different pixel rows, and adding the addition result of the adding means to the character area, photograph area, or net to which each line partial area belongs. It is characterized in that it is a partial feature amount for identifying which region of the point region belongs to.

According to the present invention, the partial calculation means converts the line partial image data in the main scanning direction into binary data by the binarization means.
In the binarized pixel row, a difference in the number of pixels between pixel rows having the same value adjacent to each other across pixel rows having different values is added by an adding unit, and the addition result belongs to a line partial area. The partial feature amount corresponds to a character area, a photograph area, or a halftone dot area. If the length of the line partial area is sufficiently larger than the size of the local block composed of the target pixel and a plurality of neighboring pixels during the main scan, it is easy to separate and identify a region having a long cycle in the main scanning direction. become. In a regular halftone dot region or the like, the difference in the number of pixels between adjacent pixels having the same value is small, and the addition result of the adding means does not become too large, so that global identification can be easily performed.

Further, in the present invention, the partial calculation means includes a difference calculation means for calculating a difference between adjacent pixel values in the line partial image data in the main scanning direction, and an adjacent pixel calculated by the difference calculation means. Difference addition means for adding only when the difference between the values satisfies a condition relating to a preset threshold value, and adding the addition result of the difference addition means to a character area, a photograph area, or a halftone area to which each line partial area belongs. It is characterized by corresponding partial feature amounts.

According to the invention, the partial calculation means calculates the difference between adjacent pixel values in the line partial pixel data in the main scanning direction by the difference calculation means, and the difference between adjacent pixel values is set to a predetermined threshold value. Since the addition is performed by the difference addition means only when the difference is larger, the feature amount of the addition result becomes small in an area where the difference between adjacent pixel values is small such as a photographic area.

[0018]

FIG. 1 shows a schematic electrical configuration of an image processing apparatus 1 according to an embodiment of the present invention. Image data to be processed is generated by the line sensor 2 and subjected to electrical processing including analog / digital conversion by the image data processing means 3. Calculation means 4
Operates according to a preset program, performs arithmetic processing on digitized image data from the image data processing means 3, and performs area recognition processing. During the image processing by the arithmetic means 4, the memory 5 is used for temporary storage of image data and the like. In the line sensor 2, a plurality of pixels are arranged one-dimensionally, and this arrangement direction is the main scanning direction.
The sub-scanning means 6 moves the line sensor 2 or the document 7 to be read relatively in the sub-scanning direction perpendicular to the main scanning direction.

In the present embodiment, the control means 8 realized by the program operation of the arithmetic means 6 performs control such that the original 7 is read in two times, prescan and main scan. The image data from the image data processing means 3 is given to the partial calculation means 10 in the pre-scan by the scan switching means 9 switched by the control means 8, and the classification of the area is performed by the classification means 11 using the calculation result of the partial calculation means. Is also identified. At the time of the main scan, the control means 8 switches the scan switching means 9 to supply the image data from the image data processing means to the local operation means 12. The local operation unit 12 is configured to locally select a target pixel sequentially selected from one line of image data read from the original 7 by the line sensor 2 at the time of the main scan and a local block in a predetermined range around the target pixel. Region characteristic processing means 1
Reference numeral 3 denotes an area identification processing unit 1 based on a feature amount or identification result at the time of pre-scan recorded in the memory 5 and an operation result of the local operation unit 12 for a local block.
3 is performed for the local block around the target pixel.

FIG. 2 shows a schematic electrical configuration related to the overall processing of this embodiment. During pre-scan,
The identification result as the partial feature amount and the classification information obtained by the partial calculation unit 10 or the classification unit 11 in FIG.
The data is recorded in a page memory 5a which is a recording means provided in the memory 5 of FIG. At the time of the main scan, the image data is temporarily stored in a block memory 5b, which is a storage unit provided in the memory 5 of FIG. The area identification processing means 13
The type of the image area is determined by using the partial feature amount and the identification result by the pre-scan stored in the page memory 5a and the local feature amount calculated by the local operation unit 12 at the time of the main scan. That is, the page memory 5
Using the result of the pre-scan stored in a
Area separation processing is performed during the main scan. The image data input at the time of the main scan is stored in the block memory 5b for a plurality of lines, for example, as digital data of 8 bits for each pixel at any one of 256 levels. The local operation means 12 calculates a plurality of local feature values from the image data stored in the block memory 5b. The area identification processing means 13 obtains, by the local operation means 12, a partial feature amount or an identification result for identifying a character area, a photograph area, or a halftone dot area, which is stored in the page memory 5a and is obtained at the time of pre-scanning. The region to which each pixel belongs is identified in addition to the local feature quantity to be obtained.

FIG. 3 shows a procedure up to recording of the partial feature amount at the time of prescan in this embodiment. The procedure of the pre-scan is started from step a0.
The image data for one line is input from the image data processing means 3 of FIG. In step a2, line portion data having a sufficient width in the main scanning direction is selected from the image data for one line input in step a1. In step a3, a partial feature amount is calculated from the selected line partial data. In step a4, the calculated partial feature amount is written and stored in the page memory 5a. In step a5, it is determined whether or not the storage of the partial feature amount has been completed for the image data for one line. If not completed, the process returns to step a2. When it is determined in step a5 that the partial feature amount for one line is stored in the page memory 5a, it is determined in step a6 whether the storage of the partial feature amount for the entire image from the document is completed. . If not finished, return to step a1,
The next line image data is input. In step a6,
When it is determined that the storage of the feature amount of the entire image has been completed, the process proceeds to the main scan in step a7.

FIG. 4 shows a procedure at the time of pre-scan according to another embodiment of the present invention. Step b0 to Step b
The procedure from step a0 to step a3 in FIG.
Correspond to each step up to. In step b4, based on the partial feature amount calculated in step b3, the type of the area to which the line partial region belongs is identified as a character area, a photograph area, or a halftone dot area based on the partial feature amount. In step b5, the identification result is written and stored in the page memory 5a as classification information. In step b6, it is determined whether or not the processing on the line portion data for one line has been completed. If not, the process returns to step b2. If finished, step b
At 7, it is determined whether the reading processing of the line image data has been completed for the entire image. If not, the process returns to step b1 to input the next line image data. If it has been completed, the process proceeds to the main scan in step b8.

FIG. 5 shows the relationship between the line image data and the line partial area at the time of prescan. Original image data 2
0 is the line image data 21 corresponding to the main scanning direction according to the image data from the pixels arranged in the main scanning direction.
Are arranged in a sub-scanning direction perpendicular to the main scanning direction. In each line image data 21, line partial image data 22 having a sufficient width in the main scanning direction is set.

FIG. 6 shows binarized line part data 2
2 shows an example. The line portion data 22 includes 24 pixels, and pixels having a pixel value of 1 are continuous like 2, 3, 2, 4, and 3. Pixels having a pixel value of 0 are continuous between pixels having a pixel value of 1, such as 2, 2, 3, and 4, respectively.
The difference between the number of consecutive pixels having a pixel value of 1 and the number of consecutive pixels adjacent to the pixel having a pixel value of 0 is 1, 1, 2, 2.
Similarly, the difference between the consecutive numbers of pixels having a pixel value of 0 is 0, 1, 1
Becomes Dividing the difference of the number of continuations by the number of continuations gives 5/5 = 1 when the pixel value is 1, and 2/5 when the pixel value is 0.
/4=0.5. If the sum of the two is taken as the feature quantity, it becomes 1.5.

FIG. 7 shows a configuration for calculating the characteristic values as shown in FIG. 6 and performing region separation at the time of pre-scanning. The image data provided from the image data processing means 3 of FIG.
0 is given. The partial operation means 10 includes a binarizing means 30, a same pixel column counting means 31, a difference calculating means 32 of the same adjacent pixel row, an adding means 33, and a dividing means. Further, in the partial operation means 10, the binarization means 30
As a different system from the division unit to the division unit 34, a calculation unit 35 for adjacent pixels, a determination unit 36, a difference addition unit 37, and a region determination unit 38 are included. The judgment result of the area judgment means 38 is
The information is stored in the page memory 5a and is referred to by the area identification processing means at the time of the main scan.

The characteristic quantities in the examples shown in FIGS. 5 and 6 are determined by the area determining means 38 based on the arithmetic means from the binarizing means 30 to the dividing means 34 in FIG. For example, assuming a halftone dot area, the i-th line image data is 230-, 240-, 230-, 50-49,
.. Are obtained as 48, 50, 229, 230, 200,...
Can be obtained. From this binary sequence,
Since the difference between the lengths of the pixel rows where adjacent values are the same is added, when a pixel row having a pixel value of 1 and a pixel row having a pixel value of 0 are regularly repeated such as a halftone dot area, , The added value becomes smaller. That is, in the character area and the photograph area, the addition value is larger in the halftone dot area. Therefore, by using this feature amount, the original can be separated into a halftone dot region and other regions.

The means from the difference calculating means 35 to the difference adding means 37 in FIG. 7 are provided mainly for separating a photographic area from other areas. For example, if the pixel density value which is the signal level for each pixel is (235, 235, 236, 2
Assuming a photographic area that changes little by little, such as 37, 239, 240,..., A threshold value is set when the difference between adjacent pixel values is added, and a difference greater than or equal to the threshold value is added. In photo areas where the values change continuously and little by little,
The added value becomes smaller. Accordingly, the value of this feature amount increases in the photograph region because the value is small and in the halftone dot region, the number of repetitions of a large change in pixel value is large. In the character area, the value increases at the edge of the contour.

FIG. 8 shows the feature values calculated by the binarizing means 30 to the dividing means 34 in FIG. 7 on the horizontal axis, and the feature values calculated by the difference calculating means 35 to the difference adding means 37 of the adjacent pixels on the vertical axis. The approximate distribution state of the character area, the photograph area, and the halftone dot area is shown along the axis. If two types of feature values are used, three regions can be separated.

FIG. 9 shows a procedure in which the partial feature corresponding to the horizontal axis in FIG. 8 is performed by the binarizing means 30 to the dividing means 30 in FIG. In step c0, line image data is input, and in step c1, line partial data is extracted. In step c2, an average value is calculated for each line portion data. In step c3, binarization of pixel data is performed using the average value calculated in step c2 as a threshold. It is assumed that a pixel value larger than the threshold is “1” and a pixel value smaller than the threshold is “0”. In step c4, the number of continuations of pixel rows having the same value of the binarized pixel value is counted and counted. In step c5, the difference between the lengths of the pixel rows having the same pixel value adjacent to each other with the pixel rows having different values interposed therebetween is counted, and the difference between the lengths of the pixel rows having the same value adjacent to each other is added. In step c6, it is determined whether or not the processing in the line partial area has been completed. If not, the process returns to step c3. If it has been completed, in step c7, a division process of dividing the addition value obtained in step c5 by the number of consecutive pixel columns having the same value in the line partial area is performed. Step c8
Then, the result of the pixel values 1 and 0 is added, and the added value is output as a feature value in step c9.

FIG. 10 shows a calculation procedure of the feature amount processed by the difference calculating means 35 to the difference adding means 37 of the adjacent pixels shown in FIG. In step d0, line image data is input, and in step d1, line partial data is extracted. In step d2, the node difference between adjacent pixels is calculated, and in step d3, it is determined whether the node difference is larger than a threshold. When it is determined that the node difference is larger than the threshold value, the node difference is added in step d4. After step d4, or when it is determined in step d3 that the node difference is equal to or smaller than the threshold value, it is determined in step d5 whether the processing in the line partial area has been completed.
When it is determined that the processing has not been completed, the process returns to step d2. If it is determined that the process has been completed, the process proceeds to step d6.
Then, the added value is output as a feature value.

FIG. 11 shows a procedure for identifying the type of the area to which the pixel of interest belongs during the main scan. In step e0, it is assumed that the partial features of the entire image and the identification result as the classification information are stored in the page memory 5a at the time of the pre-scan. In step e1, the main scan is started, and in step e2, pixel data for a plurality of lines is stored in the block memory 5b in a local block, and the feature amount is calculated. In step e3, the feature amount of the target pixel is calculated from the feature amount calculated in step e2 and the partial feature amount or the identification result stored in the page memory 5a in step e0. In step e4, it is determined whether the region to which the target pixel belongs is a character region, a photograph region, or a halftone dot region based on the feature amount calculated in step e3. In step e5, image processing according to the identification result in step e4 is performed.

FIG. 12 shows the basic concept of area determination in the present embodiment described above. FIG. 12 (1)
Indicates a pre-scan state. In the pre-scan, the line image data 21 in the main scanning direction is read from the document image data 20, and the line partial data 22 is set for a plurality of W pixels. Next, as shown in FIG. 12B, rough region separation is performed for each line partial data 22. Next, as shown in FIG. 12C, a main scan is performed, and W>
In FIG. 12 (4), N local blocks 40 each having a size of N are set around the pixel of interest 41, and the area of the pixel of interest 41 is determined.
Based on the result of (3), the area of each pixel is identified.

FIG. 13 shows another embodiment of the present invention, in which the partial feature amount and the identification result at the time of pre-scanning are temporarily stored in the page memory 5a in FIG. 5 shows a process for correcting stored contents before scanning. The processing procedure is shown in FIG. In step f0 of FIG. 14, the identification result of the partial feature amount and the classification information is stored in the page memory 5a as a pre-scan. In step f1, correction is performed based on the identification results of a plurality of rows. At step f2, it is determined whether or not the correction has been completed for all the lines. If it is determined that the processing has not been completed, the process returns to step f1. When the correction for all the lines is completed, the process proceeds to the main scan in step f3. By performing such correction, the identification result of the partial feature amount or the classification information obtained for each line partial area at the time of the pre-scan is corrected using a plurality of lines before the main scan is performed, and the main scan direction and the main scan direction are corrected. Global handling can be performed in the sub-scanning direction.

FIG. 15 shows, as still another embodiment of the present invention, a method in which the area identification processing means 13 checks the identification result of the area indicating the partial feature amount or the classification information stored in the page memory 5a at the time of the prescan. By referring to a plurality of lines before and after including the pixel, the area to which the pixel of interest belongs is identified together with the local feature obtained from the block memory 5b at the time of the main scan. Even in such a method, a global area can be identified.

[0035]

As described above, according to the present invention, an original is read twice by dividing the original into a pre-scan and a main scan, and the area is roughly separated in advance using line image data at the time of the pre-scan. The amount is calculated or the type of region is classified based on the partial feature amount. Using this result as a feature parameter for region identification at the time of main scan, it is possible to accurately perform region separation for each pixel, which is difficult to identify only with local information. In other words, it is difficult to identify a region having a long cycle that does not fit in the local block by the identification using only the local region. However, in pre-scanning, it is not sufficient in the main scanning direction compared to the width of the image data used for the local block. Since a partial feature amount from large line partial data is taken into account, a halftone dot region having a relatively long cycle in the main scanning direction can be easily identified.

Further, according to the present invention, the partial feature amount and the identification result obtained from the line partial data at the time of the pre-scan are determined by:
The correction is performed using a plurality of lines, the correction result is re-recorded, and added to the local feature amount at the time of region identification at the time of the main scan. Thus, a more comprehensive identification result can be obtained and the identification accuracy can be improved.

Further, according to the present invention, a partial feature amount or classification information obtained based on line partial data at the time of pre-scan is used as a local feature amount for region identification at the time of the main scan for a plurality of lines before and after the target pixel. Therefore, the parameters can be used as a plurality of parameters at the time of the main scan, and global identification errors at the time of the pre-scan can be reduced.

Further, according to the present invention, the partial feature amount obtained by binarizing the line data at the time of pre-scanning and adding the difference in the number of pixels between adjacent pixel rows having the same value across pixel rows having different values. Has a small value in a periodic region, and has a large value in a non-periodic region. By utilizing this, it is possible to separate the halftone dot area from other character areas and photograph areas.

Further, according to the present invention, partial feature values are obtained by adding only when the difference between adjacent pixel values is larger than a preset threshold value using line data at the time of pre-scanning. Is small, the partial feature amount is small. By utilizing this, it is possible to easily separate a photograph region from other character regions and halftone dot regions.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic electrical configuration of an embodiment of the present invention.

FIG. 2 is a block diagram showing a schematic logical configuration for area identification by the image processing apparatus of FIG. 1;

FIG. 3 is a flowchart showing a processing procedure at the time of pre-scan in the embodiment of FIG. 1;

FIG. 4 is a flowchart showing a processing procedure at the time of pre-scan in the embodiment of FIG. 1;

FIG. 5 is a diagram showing line image data and a line partial area at the time of pre-scan in the embodiment of FIG. 1;

6 is a diagram illustrating an example of calculating a feature amount based on line image data and line partial data read in FIG. 5;

FIG. 7 is a block diagram showing a logical configuration for calculating a feature amount at the time of a pre-scan;

8 is a diagram showing a relationship between a feature amount calculated by the configuration of FIG. 7 and a type of a region.

FIG. 9 is a flowchart illustrating a calculation procedure of a feature amount related to the horizontal axis in FIG. 8;

FIG. 10 is a flowchart showing a procedure for calculating a feature amount on the vertical axis in FIG. 8;

FIG. 11 is a flowchart showing a processing procedure at the time of a main scan in the processing procedure of FIG. 2;

FIG. 12 shows original image data 20 based on the embodiment of FIG. 2;
FIG. 6 is a diagram showing a relationship between the processing of FIG.

FIG. 13 is a block diagram showing a partial logical configuration of another embodiment of the present invention.

FIG. 14 is a partial flowchart showing the operation of the embodiment of FIG.

FIG. 15 is a simplified block diagram showing a logical configuration of still another embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image processing apparatus 4 Computing means 5 Memory 5a Page memory 5b Block memory 7 Document 8 Control means 10 Partial computing means 11 Classification means 12 Local computing means 13 Area identification processing means 20 Document image data 21 Line image data 22 Line partial data 30 2 Value conversion means 31 Same pixel column counting means 32 Same adjacent pixel row difference calculation means 33 Addition means 34 Division means 35 Neighbor pixel difference calculation means 36 Judgment means 37 Difference addition means 38 Area judgment means 40 Local block 41 Target pixel 50 Correction means

Claims (5)

[Claims] 1. An image signal obtained by two-dimensionally scanning a document in a main scanning direction and a sub-scanning direction, wherein each pixel belongs to a character area, a photograph area, or a halftone area. In an image processing apparatus that identifies and processes a document, a control unit that controls scanning of a document into pre-scanning and main scanning is performed, and during pre-scanning, line image data in the main scanning direction is used. A partial operation means for obtaining a partial characteristic amount representing a characteristic of the partial area of the line; a classification means for classifying and classifying a type of an area to which the partial area of each line belongs from the partial characteristic amount obtained by the partial operation means; A calculation result of the partial feature amount obtained by the calculation means,
Or a recording unit for recording an identification result obtained by the classifying unit; and image data of a local block including a pixel of interest and a plurality of pixels near the pixel of interest, where a pixel selected one by one from an image signal is used as a pixel of interest during the main scan. Storage means for storing a plurality of local feature values representing characteristics of each of a character area, a photograph area, and a halftone dot area from image data stored in the storage means; The region to which the pixel of interest belongs in the local block is identified using the partial feature amount or the identification result obtained and the local feature amount for each pixel of interest obtained by the local operation means, and the type of the region to be identified An image processing apparatus comprising: an area identification processing unit configured to perform a predetermined process according to the image processing. 2. The method according to claim 1, wherein the partial calculation unit calculates a partial feature amount representing a characteristic of the character area, the photograph area, or the halftone dot area using line partial read data having a plurality of pixel widths in the main scanning direction. A correction unit configured to correct the partial feature amount or the identification result recorded in the recording unit in units of blocks using a plurality of lines in the sub-scanning direction before performing the main scan, and re-record the correction result in the recording unit; 2. The image according to claim 1, wherein the area identification processing unit outputs the area separation information of the pixel of interest based on the partial feature amount or the identification result re-recorded in the recording unit and the local feature amount. 3. Processing equipment. 3. The area discriminating processing unit uses a plurality of lines before and after the line to which the pixel of interest belongs to each of the partial feature amounts or the identification results recorded in the recording unit, and Pixels are text area, photo area,
2. The method according to claim 1, further comprising: identifying to which of the halftone dot regions the region belongs, and outputting the region identification information.
Or the image processing device according to 2. 4. The method according to claim 1, wherein the partial operation unit binarizes the line partial image data in the main scanning direction.
A binarizing means, and an adding means for adding the difference in the number of pixels between adjacent pixel rows having the same value across the pixel rows having different values in the binarized pixel row obtained from the binarizing means. 2. The method according to claim 1, wherein the addition result of the means is a partial feature amount for identifying which of the character area, the photograph area, and the halftone dot area each line partial area belongs to. The image processing apparatus according to any one of claims 1 to 3, 5. The partial calculating means includes: a difference calculating means for calculating a difference between adjacent pixel values in the line partial image data in the main scanning direction; and a difference between adjacent pixel values calculated by the difference calculating means. And a difference adding unit that adds only when a condition regarding a preset threshold value is satisfied, and a part corresponding to a character area, a photograph area, or a halftone area to which each line partial area belongs. The image processing apparatus according to claim 1, wherein the image processing apparatus is a feature amount.