JP2001222681A - Device and method for detecting picture and device and method for recognizing picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a picture detecting device and method, a character recognizing device and method, by which a code part, etc., is correctly detected even when the size of the code part such as a character or a graphic comprised in an input picture or the thickness of a stroke, etc., are not previously distinct, or even when the input picture comprises the code part where a width or the thickness of the stroke is largely different. SOLUTION: The device for detecting a picture is provided with a picture input part 1 for inputting an original picture to be the object of picture detection, a reference area specifying part 2 which specifies a plurality of partial areas comprising a pixel to be considered and being different in largeness as a reference area, a reference value calculating part 3 for calculating a reference value concerning the plurality of specified reference areas, a pixel judging part 4 for successively comparing the original pixel value of the pixel to be considered with the plurality of reference values corresponding to the pixel in order from the smallest reference area and judging whether the pixel is the code part or not, a detection picture storage part 5 for storing a pixel judgement result and generating a binary picture and a character recognizing part 6 for recognizing the character from the binary picture.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image detection for detecting a symbol portion included in a digital image, and more particularly to an image detection device and a symbol detection device for detecting a symbol portion such as a character or a figure included in a scene image. The present invention relates to an image detection method, a character recognition device and a character recognition method for recognizing characters from a detected image.

[0002]

2. Description of the Related Art FIG.
FIG. 1 is a configuration diagram showing a conventional character string detection device disclosed in Japanese Patent Publication No. In FIG. 13, reference numeral 100 denotes an image input unit which inputs a digital signal of a predetermined area to be detected in the original image and serves as an input image, and 101 denotes a moving average value for each pixel of the input image input from the image input unit 100 A moving average unit for obtaining a high-frequency image by performing subtraction between the input image and the moving average value calculated by the moving average unit 101; An upper / lower limiter 104 for performing an upper / lower limiter process, a histogram creating / binarizing unit 104 for creating a histogram for each partial region with respect to the high-frequency image to which the upper / lower limiter process has been added, and performing a binarization process; A logical operation unit that detects only the vicinity of the character string and generates a binary image.

Next, the operation of the above-described conventional character string detecting device will be described. First, the image input unit 100
A digital signal of a predetermined area in an original image to be detected is input. The input image obtained by the image input unit 100 is
It is assumed that each pixel is a grayscale image in which pixel values take pixel values in the number of gradations from 0 to 255 according to luminance. Hereinafter, the quantized pixel value of each pixel included in the input image is referred to as “original pixel value”. The input image includes a target portion to be detected as a symbol portion such as a character or a figure. The image to be detected is composed of line elements having a constant thickness. Hereinafter, a target portion to be detected as a symbol portion such as a character or a figure in the input image is referred to as a “character portion”, and the other portions are referred to as a “background portion”.

[0004] Next, the moving averaging unit 101 calculates, for each pixel constituting the input image, one vertical pixel centered on the pixel,
A partial region of the horizontal p pixels is specified, and an average pixel value for the partial region is calculated from original pixel values of all pixels included in the partial region. The size of the partial area for which the average pixel value is calculated is defined by the number p of pixels of the width of the partial area. The value of p is set in advance so as to be several times the width of the character to be detected. The moving average unit calculates the corresponding average pixel value for all the pixels of the input image while sequentially changing the target pixel of the average pixel value calculation and the partial area corresponding to the pixel.

Next, the image difference section 102 subtracts the original pixel value of the pixel of the input image from the average pixel value obtained by the moving average section 101. The result of subtraction of the original pixel value and the average pixel value is hereinafter referred to as a “high-frequency pixel value”. The image difference unit calculates high-frequency pixel values for all pixels to be subjected to image detection, and generates a processed image from the result. The image subjected to the arithmetic processing is hereinafter referred to as a “high-frequency image”. In this high-frequency image, the high-frequency pixel value of each pixel has the following characteristics. First, in a pixel forming a character portion and having a large original pixel value, the high-frequency pixel value shifts to a large positive value. Next, in the pixels in the background and near the character, the high-frequency pixel value shifts greatly to minus. Finally, where the original pixel value changes gradually, the high-frequency pixel value becomes almost zero.

Next, the upper / lower limiter 103 performs an upper / lower limiter process on the high-frequency image. That is, when the high-frequency pixel value is larger than a predetermined upper limit, the high-frequency pixel value is limited to the upper limit. Conversely, when the high-frequency pixel value is smaller than the predetermined lower limit, the high-frequency pixel value is limited. To the lower limit.

Next, a histogram creation / binarization unit 104
Is a high-frequency image subjected to the upper and lower limiter processing,
The image data is divided into a plurality of partial regions of an appropriate size, and a histogram of high-frequency pixel values subjected to upper and lower limiters for each partial region is created. Further, the histogram generation / binarization unit calculates the average and the variance of the high-frequency pixel values in each partial region from the histogram of each partial region high-frequency pixel value, and obtains a high-luminance binary image H _{(x , y)} and a low-luminance binary image L _{(x, y)} . In these binary images, due to the characteristics of the high-frequency image, H _{(x, y)} is 1 in a character portion having a large pixel value, and L _{(x, y)} is in a background portion of a pixel in the vicinity of the character portion _{. y)} becomes 1.

[0008]

(Equation 1)

Finally, the logical operation unit 105 performs a logical operation of the following expression 2 on the high luminance binary image H _{(x, y)} and the low luminance binary image L _{(x, y)} . In Equation 2, l is the number of pixels corresponding to the thickness of the line element of the character portion to be detected,
Set in advance. By this logical operation, L on both sides is
_{(x, y)} is 1 and the center H _{(x, y)} is 1 and the H on both sides
_{(x, y)} is 1 and only the part where the center L _{(x, y)} is 1 becomes 1,
A detection image in which only the outline of the character portion is detected is generated.

[0010] ｛(H _{(x, y)} ∩L _{(xl, y)} ) ((H _{(x, y)} ∩L _{(x + l, y)} )｝ ∪ ｛(H _{(xl, y)} ∩L _{( x, y)} ) ∩ (H _{(x + l, y)} ∩L _{(x, y)} )｝ ... (Equation 2)

[0011]

In the above-described conventional character string detection apparatus, a constant (depending on the size of a symbol portion, the thickness of a line element, etc.) is used to detect a symbol portion such as a character or a figure to be detected. The above p and l) are used. Therefore, if the size and the thickness of a line element of a character or a figure included in an input image are not clear beforehand, it is not possible to accurately detect the symbol part of a character or a figure. was there.

Further, in the conventional character string detecting device, the constants (p and l) used for the character string detection are fixed through one detection process. Therefore, when an input image includes a symbol portion such as a character or a graphic having a different width or a different line element thickness, it cannot be detected at the same time.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and detects a symbol or a symbol such as a figure or a character in an input image to be detected irrespective of the size or the thickness of a line element. Even when the size of the symbol portion or the thickness of the line element is not clear in advance, or when the input image includes a symbol portion having a significantly different width or line element thickness, The purpose of the present invention is to obtain an image detection device and an image detection method, and a character recognition device and a character recognition method capable of accurately detecting an image, and further reduce the storage capacity and processing time required for the image detection process and the character recognition process. Provided are an image detection device and an image detection method, and a character recognition device and a character recognition method that can simultaneously detect a high-luminance symbol portion and a low-luminance symbol portion without increasing them. .

[0014]

[Means for Solving the Problems]

In order to solve the above problems and achieve the object, an image detecting apparatus according to the present invention comprises an image input means for inputting a digital signal of a predetermined area in an original image for detecting a symbol portion. Reference area specifying means for specifying, as a reference area, a plurality of partial areas having different sizes including a pixel to be determined as a symbol part, and a reference value for calculating a reference value for each of the plurality of reference areas Calculating means for sequentially comparing a pixel value obtained as a digital signal of a pixel to be determined with the plurality of reference values corresponding to the pixel in ascending order of the size of the reference area to determine whether the pixel is a symbol part And a pixel determination unit for determining whether or not the determination is made.

In the image detecting apparatus according to the next invention, the reference area specifying means is configured to specify a plurality of partial areas having different sizes around a pixel to be determined as a reference area. It is characterized by the following.

In the image detecting apparatus according to the next invention, the reference area specifying means includes a plurality of partial areas including a pixel to be determined and having different positions and sizes relative to the pixel. It is characterized in that it is configured to be specified as a reference area.

In the image detecting apparatus according to the next invention, the reference area specifying means is configured to specify a plurality of partial areas having different shapes including pixels to be determined as reference areas. And

In the character recognition apparatus according to the next invention, furthermore, a detection image storage means for storing the determination result of the pixel determination means as a binary image, and a symbol part included in the binary image is a character. And character recognition means for recognizing the character.

In the character recognition device according to the next invention, the pixel determining means sequentially compares the pixel to be determined with a plurality of reference values, and determines whether the pixel is a part other than the symbol part around the symbol part. A determination is made as to whether or not the part is a symbol part having a higher luminance than the part, and a result of the determination is output. The detected image storage unit stores the result of the determination as to whether the pixel portion is a high-luminance symbol portion by the pixel determination unit as a binary image of the high-luminance symbol portion. The result of the determination as to whether or not there is a binary image of a low-luminance symbol portion is stored, and the character recognition means recognizes high-luminance characters from the high-luminance symbol portion binary image, and Low brightness Characterized in that the binary image of the symbol part is recognizing constituting the low brightness of the character.

In the character recognition device according to the next invention, the pixel determining means sequentially compares the pixel to be determined with a plurality of reference values, and determines whether the pixel is a part other than the symbol part around the symbol part. It is determined whether the portion is a symbol portion having a higher luminance than the portion, a symbol portion having a lower luminance than a portion other than the symbol portion, or a portion other than the symbol portion, and a result of the determination is output as a ternary value. The detected image storage means is configured to store the determination result by the pixel determination means as a ternary image, and the character recognizing means is configured to recognize both high-brightness characters and low-brightness characters from the ternary image. It is characterized in that it is configured to recognize.

In the image detecting method according to the next invention, an image inputting step of inputting a digital signal of a predetermined area in an original image for detecting a symbol portion, and an object for determining whether the symbol portion is present. A reference region specifying step of specifying a plurality of partial regions having different sizes including pixels as a reference region,
A reference value calculating step of calculating a reference value for each of the plurality of reference regions, and determining a pixel value obtained as a digital signal of a pixel to be determined by comparing the plurality of reference values and the size of the reference region corresponding to the pixel. A pixel determination step of sequentially comparing the pixels in ascending order to determine whether or not the pixel is a symbol portion.

In the image detecting method according to the next invention, the reference area specifying step is characterized in that a plurality of partial areas having different sizes around a pixel to be determined are specified as reference areas. I do.

In the image detection method according to the next invention, the reference area specifying step includes a plurality of partial areas including a pixel to be determined and having different positions and sizes relative to the pixel. It is characterized in that it is specified as a reference area.

In the image detecting method according to the next invention, the reference area specifying step is characterized in that a plurality of partial areas having different shapes including pixels to be determined are specified as reference areas.

In the character recognition method according to the next invention, there is further provided a detection image storing step of storing the determination result of the pixel determination means as a binary image, and a case where a symbol portion included in the binary image is a character. And a character recognition step of recognizing the character.

In the character recognition method according to the next invention, in the pixel determining step, the pixel to be determined is sequentially compared with a plurality of reference values to determine whether the pixel is a part other than the symbol part around the symbol part. And whether the pixel is a symbol part having a lower luminance than the part other than the symbol part, and outputs the judgment result. The image storing step stores the result of the determination as to whether the pixel portion is a high-luminance symbol portion by the pixel determination section as a binary image of the high-luminance symbol portion,
The result of the determination as to whether or not the pixel portion is a low-luminance symbol portion by the pixel determination means is stored as a low-luminance symbol portion binary image. And recognizing the low-luminance character from the binary image of the low-luminance symbol portion.

In the character recognition method according to the next invention, in the pixel determining step, the pixel to be determined is sequentially compared with a plurality of reference values to determine whether the pixel is a part other than the symbol part around the symbol part. Is determined to be a symbol part having a higher luminance than the part, a symbol part having a lower luminance than a part other than the symbol part, or a part other than the symbol part, and a result of the determination is output as a ternary value. Then, the detection image storage step includes:
The result of the determination by the pixel determining means is stored as a ternary image, and the character recognition step recognizes both high-luminance characters and low-luminance characters from the ternary image.

[0029]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 FIG. 1 is a configuration diagram of the character recognition device according to the first embodiment. In FIG. 1, reference numeral 1 denotes an image input unit for inputting an original image to be detected, and reference numeral 2 denotes a reference region specifying unit for specifying a plurality of partial regions included in the input image input by the image input unit 1 as reference regions. Reference numeral 3 denotes a reference value calculation unit that calculates reference values for a plurality of reference regions specified by the reference region specification unit 2. Reference numeral 4 denotes a pixel determination unit that performs pixel determination for all pixels included in a predetermined region where image detection is performed. Is a detected image storage unit that stores a pixel determination result output from the pixel determination unit 4 to generate a binary image, and 6 is a character recognition unit that recognizes characters from the binary image generated by the detected image storage unit 5. is there.

The operation of the above-described character recognition apparatus will be described below with reference to the drawings. FIG. 2 is a processing flowchart of the character recognition device. First, the image input unit 1 inputs a digital signal of a predetermined area in an original image to be subjected to image detection (step S1). The input image obtained by the image input unit 1 is a grayscale image in which each pixel takes a pixel value quantized to a gradation number from 0 to 255 according to luminance, and a brighter pixel has a larger pixel value. And

Next, the reference area specifying unit 2 specifies a plurality of partial areas (# 1 to #N, where N is the number of specified reference areas) included in the input image as reference areas (Step S).
2). FIG. 3 is a schematic diagram showing the reference area specified by the reference area specifying unit 2. In FIG. 3, reference numeral 7 denotes a pixel for which a pixel is determined to be a character portion or a background portion;
Reference numeral 10 denotes a partial area specified as a reference area. Hereinafter, a pixel included in a predetermined area where image detection is performed and for which a pixel is determined to be a character portion or a background portion is referred to as a “target pixel”. In FIG. 3, in particular, the reference area specifying unit 2
Shows a case where three partial regions having the same shape and different sizes are specified as reference regions around the target pixel. In this case, the number N of the specified reference areas is 3
It is. Here, the reference area specifying unit 2 sets the minimum partial area 8 as the first (# 1) reference area, sets the intermediate partial area 9 as the second (# 2) reference area, Is specified as the third (# 3) reference area. Here, the size of the minimum partial area 8 is determined in advance according to the size of the input image.

Next, the reference value calculation unit 3 calculates the average value of the original pixel values of all the pixels included in the reference area, and uses the average value as the reference value for the pixel of interest (step S3).
Here, since there are a plurality of specified reference regions for one target pixel, a plurality of calculated reference values are calculated corresponding to each of the reference regions.

The reference value calculation unit 3 repeats the reference value calculation for all the pixels included in the predetermined area where the image is to be detected in the input image while sequentially changing the pixel of interest, and includes the reference value in the predetermined area where the image is to be detected. A plurality of reference values (# 1 to # 3) corresponding to all the pixels to be obtained are obtained.

Next, the pixel determination unit 4 performs image detection by comparing the original pixel value of the input image with a plurality of reference values corresponding to the pixel referenced by the reference value calculation unit 3 (step S4). ). Hereinafter, the pixel determination processing for one target pixel will be described with reference to FIG. FIG. 4 is a processing flowchart of the pixel determination processing. In FIG. 4, n is a variable indicated for specifying a reference value. n is associated in ascending order of the size of the reference region for which the reference value is calculated, and n
= 1, the reference value for the smallest reference area, n = N
Then, it indicates the reference value for the largest reference area.

First, the pixel determination unit 4 initializes n to 1 (step S11), compares the reference value for the smallest reference area with the original pixel value of the target pixel according to Equation 3, and determines that the target pixel is a character part. It is determined whether there is (Step S12). Here, Expression 3 is a comparison condition expression used to detect a pixel having higher luminance than the surroundings, that is, a character portion having higher luminance than the background portion.

Original pixel value−Reference value> C ₁ (Equation 3) where C ₁ is a constant.

Next, the pixel judging unit judges whether the above equation 3 is satisfied (step S13). If it is satisfied (step S13, Yes), the pixel of interest is judged to be a character part, and the pixel judgment result is output. Is set to 1 (step S14), and the pixel determination process ends. On the other hand, if Expression 3 is not satisfied in Step S13 (Step S13, No), 1 is added to n (Step S15), and it is determined whether or not n is larger than the maximum value N (Step S16). . If the result of this determination is that n is less than or equal to N (step S16, N
o) Based on a reference value calculated from a reference area one step larger than the reference value compared last time, the pixel determination process for the target pixel is repeated (steps S12 and S13). If n is larger than N in step S16 (step S16, No), the target pixel is determined to be the background part because the comparison with all the reference values has not determined the character part to be the background part. The output of the determination result is set to 0 (step S17), and the pixel determination process ends.

The pixel judging section 4 repeats the above-described pixel judging process for all the pixels included in the predetermined area for performing the image detection in the input image while sequentially changing the pixel of interest. The detected image storage unit 5 stores the output of the pixel determination unit and generates a binary image in which only the pixels forming the character part are detected (Step S4). Finally, the character recognition unit 6 performs character recognition from the binary image generated by the detected image storage unit (Step S5). As a method for recognizing characters from a binary image, for example,
IEICE Transactions D-II, Vol. J81-D-
II, No. 10, pp. 2280-2287 (1998
Year) Use the method described in "Extraction of license plate area in consideration of arrangement rules of partial character strings".

As described above, in the pixel determination processing by the pixel determination section 4, in step S13 of the above-described pixel determination processing flow shown in FIG. Although the method of detecting the character portion has been described, by using Expression 4 instead of Expression 3 as a condition for pixel determination, a pixel having lower luminance than the surroundings, that is, a character portion having lower luminance than the background portion can be obtained. It is possible to detect.

Reference value−Original pixel value> C ₂ (Equation 4) where C ₂ is a constant.

Next, a case where the first embodiment is applied to the character recognition processing in a case where a character portion having a different size and a different line element thickness in an input image is described with reference to the drawings. . FIG. 5 is a schematic diagram showing a specific outline of the reference area shown in FIG. In FIG. 5, reference numeral 11 denotes an image of a dial,
Reference numerals 12 and 13 denote pixels constituting a character portion, 14 denotes a pixel constituting a background portion, 15 to 17 denote partial regions specified as reference regions corresponding to the pixel 12, and 18 to 20 denote reference regions corresponding to the pixel 13. Partial area specified, 21-2
3 is a partial area specified as a reference area corresponding to the pixel 14, and 24 is a scene image part.

The input image includes an image pickup 11 of the dial to be subjected to the image detection processing, and a scene image section 24 outside the image detection processing which is not to be subjected to the image detection processing. In the image 11 of the dial, “6” and “7” which are character portions having different sizes and line element thicknesses as character portions having higher luminance than the background portion.
Characters are included. 5. In FIG.
The character "6" in 1 is a character part with a thick and large line element, and the character "7" is a character part with a thin and small line element.

First, when the pixel 12 is set as the target pixel, the original pixel value of the pixel 12 is compared with the reference value calculated for the smallest partial area 15. In the partial region 15, the ratio occupied by the background portion is large, and the reference value is smaller than the original pixel value.

Next, when the pixel 13 is set as the target pixel, the original pixel value is compared with the reference value calculated for the smallest partial area 18. However, since the partial region 18 includes only the pixels of the character portion with high luminance, the reference value is almost the same as the original pixel value of the target pixel, and does not satisfy Expression 3. In this case, a comparison is made between the reference value calculated for the partial area 19 that is one step larger and the original pixel value. However, since the partial area 19 also includes only pixels of a character portion having a high luminance, Expression 3 is not similarly satisfied. Further, a comparison between the reference value calculated for the largest partial area 20 and the original pixel value is performed. Partial area 20
Includes a background part, the reference value is smaller than the original pixel value, and pixel 13 is determined to be a character part by satisfying Expression 3.

Finally, when the pixel 14 is set as the pixel of interest, all of the partial regions 21 to 23 include only the background portion. Therefore, the reference value calculated for each reference region and the original pixel value are compared. Equation 3 is not satisfied in the comparison. Therefore, the pixel 14 is finally determined to be a background portion.

When the target pixel is located at the end of the image pickup 11 of the dial, such as the pixel 12, for example, the partial area 17
When a large reference region such as is specified from the beginning, the scene image portion 24 outside the image of the dial is placed in the reference region.
Will also be included. In this case, the reference value calculated according to the luminance of the scene image portion changes, and the pixel determination portion cannot accurately perform the pixel determination process on the target pixel 12. Therefore, the character recognition device according to the first embodiment identifies a plurality of reference areas having different sizes, and compares the reference value and the original pixel value in ascending order of the size of the reference area, thereby capturing the image of the dial. The original pixel value of the pixel of interest is compared with the reference value so as to be as little affected by the scene image portion outside the pixel as possible.

With such a configuration, the character recognition device according to the first embodiment can detect these regardless of the size of the character portion included in the input image and the thickness of the line element. Even when the size of the character part or the thickness of the line element is not clear in advance, or when the input image includes a character part having a greatly different width or line element thickness, these are accurately imaged. Detection and character recognition can be performed. Also,
Even when one input image includes a plurality of characters having greatly different sizes and line element thicknesses, it is possible to simultaneously perform image detection and character recognition. Furthermore, even when the character to be detected is located at the end of the image of the dial, the character portion can be accurately detected without being affected by a scene image outside the image of the dial. Further, by specifying a partial region centering on the target pixel as a reference region and reflecting the original pixel values of the pixels in the peripheral region in the calculation of the reference value, it is possible to perform image detection without unevenness in luminance.

In the first embodiment, the number of reference areas specified by the reference area specifying unit is three. However, the number may be two or four or more. In the first embodiment, “the average value of the original pixel values of all the pixels included in the reference area” is used as the reference value calculated by the reference value calculation unit. , '80 /
4, Vol. J63-D, No. 4, p. 349-35
6 "Automatic threshold value selection method based on discrimination and least square criterion" may be used, and other "intermediate values among original pixel values of all pixels included in reference area" and "reference area" For example, a value calculated from the original pixel values of the pixels included in the reference area may be used, such as "the pixel value having the highest number of appearances among the original pixel values of all the pixels included in". Further, the method of recognizing characters from a binary image is the method described in the above-mentioned document, but this may be another method.

Embodiment 2 In the second embodiment, when an original image includes both a character portion having a higher luminance than the background portion and a character portion having a lower luminance than the background portion, one time is applied to both of these character portions. Image detection is performed in the processing, and both high-brightness characters and low-brightness characters are recognized from the result. Hereinafter, a character portion having a higher luminance than the background portion to be detected is referred to as a “high-luminance character portion”, and a character portion having a lower luminance than the background portion is referred to as a “low-luminance character portion”.
Call.

FIG. 6 is a block diagram of the character recognition apparatus according to the second embodiment. Note that the second embodiment is different from the first embodiment described above in that the pixel determination unit, the detected image storage unit, and the character recognition unit are different, and the other configurations are the same. The description is omitted by attaching reference numerals. 4a
Is a pixel determination unit that performs a pixel determination for all pixels included in a predetermined area for performing image detection, determines whether or not a target pixel is a high-luminance character portion, and outputs the determination result. It determines whether or not it is a character part and outputs the determination result. Next, 5a is a detection image storage unit that stores a pixel determination result of the input image to generate a binary image,
A high-luminance character portion binary image is generated from the determination result as to whether the pixel is a high-luminance character portion output from the pixel determination portion 4a,
A binary image of the low-luminance character portion is generated from the determination result as to whether or not the character portion is the low-luminance character portion. Reference numeral 6a denotes a character recognition unit for recognizing a character from the generated binary image. The character recognition unit 6a recognizes a high-luminance character from the high-luminance character portion binary image generated by the detection image generation unit 5a, and outputs a low-luminance character portion. A low-luminance character is recognized from the binary image.

Hereinafter, the image detection processing and the character recognition processing in the character recognition device configured as described above will be described.
Note that the overall processing flow of the character recognition device of the second embodiment differs from the processing flow of the first embodiment shown in FIG. 2 only in image detection and character recognition (steps S4 and S5). Therefore, the description of the other processing steps is omitted.

First, the pixel determination section 4a determines whether the target pixel is a high-luminance character portion and outputs the determination result, and determines whether the target pixel is a low-luminance character portion and outputs the determination result. . Here, when the pixel of interest is a high-brightness character portion, the output of the pixel determination portion is 1 when the determination result is a high-brightness character portion and is 0 when the target pixel is another low-brightness character portion. (The output of the pixel determination unit 4a is [1, 0].) Conversely, when the pixel of interest is a low-luminance character portion, the determination result as to whether the pixel is a low-luminance character portion is 1 and the determination result as to whether the pixel is a high-luminance character portion is 0 ([0, 1]). ). When the pixel of interest is the background portion, both the determination result as to whether it is a high luminance character portion and the determination result as to whether it is a low luminance character portion are 0 ([0,
0]).

Hereinafter, the processing of the pixel determining unit 4a according to the second embodiment will be described with reference to the drawings. FIG. 7 is a processing flowchart of the pixel determination processing 4a. First, a variable n indicating the reference area is initialized with 1 (step S21), and the reference value calculated for the smallest reference area is compared with the original pixel value of the target pixel (step S22). Next, Embodiment 1
Is used as a conditional expression for determining a high-luminance character portion, it is determined whether or not the condition of the high-luminance character portion is satisfied (step S2).
3) If the condition is satisfied (step S23, Ye
s), the target pixel is determined to be a high-luminance character part (step S24), and the determination result as to whether the pixel of interest is a high-luminance character part is 1
Then, the pixel determination process for the pixel of interest ends.
In this case, the output of the pixel determination unit 4a is [1, 0].

If the condition for the high-brightness character portion is not satisfied (step S23, No), the expression 4 shown in the first embodiment is used as a conditional expression for determining the low-brightness character portion. It is determined whether or not the condition is satisfied (step S25). If the condition is determined to be satisfied (step S25, Ye
s), the target pixel is determined to be a low-luminance character portion (step S26), and the determination result as to whether the target pixel is a low-luminance character portion is set to 1
Then, the pixel determination process for the pixel of interest ends.
In this case, the output of the pixel determination unit 4a is [0, 1].

If the condition for the low-luminance character portion is not satisfied (step S25, No), 1 is added to n (step S27), and whether n is larger than the number N of the specified reference areas or not is determined. That is, it is determined whether the comparison with all reference values corresponding to the target pixel has been completed (step S2).
8). As a result, if n is equal to or less than N (step S2
8, No), returning to step S22, comparing the reference value calculated for the reference region one step larger than the reference value compared last time with the original pixel value of the target pixel, and repeating the pixel determination for the target pixel (step S22). S23, S25). On the other hand, if the result of the determination in step S28 is that n is greater than N (step S28, Yes), the pixel of interest is determined to be a background portion (step S29), and whether or not it is a low-luminance character portion The determination result and the determination result as to whether or not the pixel is a high-luminance character portion are both output as 0 ([0, 0]), and the pixel determination process for the pixel of interest ends.

The pixel determination section 4a repeatedly performs the above-described pixel determination process for all pixels included in a predetermined area for performing image detection in the input image while sequentially changing the pixel of interest. As a result, for each pixel, two determination results, that is, a determination result as to whether it is a high-luminance character portion and a determination result as to whether it is a low-luminance character portion, are output. Next, the detected image storage unit 5a stores these outputs, generates a binary image of a high-luminance character from the determination result of the high-luminance character portion, and generates a binary image of the low-luminance character portion from the determination result. , A binary image of a low-luminance character portion is generated.

Finally, the character recognition section 6a performs character recognition from the two binary images generated in the detected image storage section. In the second embodiment, first, for a binary image of a high-luminance character portion, character recognition is performed by, for example, the method described in the first embodiment, and high-luminance characters such as white characters included in the input image are recognized. Thereafter, character recognition is similarly performed on the binary image of the low-luminance character portion to recognize high-luminance characters such as black characters.

As described above, in the second embodiment,
Even if the input image contains one or both of the high-brightness character part and the low-brightness character part, and the size and line element thickness of each are not clear in advance, the image Detection and character recognition can be performed. Therefore, even if it is not possible to predict whether the character portion included in the input image has lower luminance or higher luminance than the background portion, character recognition can be reliably performed. Furthermore, since one high-luminance character portion and low-luminance character portion can be detected by one image detection process for one input image, the processing time required from image input to character recognition can be reduced. Can be expected.

Embodiment 3 The character recognition device according to the third embodiment differs from the character recognition device according to the second embodiment shown in FIG. 6 only in the configuration of the reference area specifying unit 2. The reference area specifying unit 2 specifies, as a reference area, a plurality of partial areas including a target pixel and having different positions and sizes relative to the target pixel, and sets a target value based on a reference value calculated from the plurality of reference areas. The pixel determination of the pixel is performed. The overall processing flow of the third embodiment is the same as the processing flow shown in FIG.

The reference area specifying section 2 specifies a plurality of partial areas included in the input image as reference areas by the method shown in FIG. FIG. 8 is a schematic diagram illustrating a reference area specified by the reference area specifying unit according to the third embodiment. In FIG. 8, reference numerals 25 to 29 denote partial regions specified as reference regions. First, the reference region specifying unit 2 specifies the portion region 25 centered on the target pixel 7 as a reference area A _1. Further, the reference area specifying unit 2 sets the partial area 26 having the target pixel at the right end center as the reference area B ₁ , the partial area 27 having the target pixel at the left end center as the reference area C ₁ , and the partial area having the target pixel at the upper end center. 28 is specified as a reference area D ₁ , and a partial area 29 with the target pixel at the lower center is specified as a reference area E ₁ . The reference areas B _{1 to} E ₁ all have the same size and the same shape as the reference area A ₁ . These reference areas A _{1 to} E ₁ which are partial areas of the same size are specified as one reference area group (# 1).

Next, the reference area specifying unit 2 repeatedly specifies the reference area group by changing the size of the partial area, and specifies a plurality of reference area groups (# 1 to #N). Note that N is the number of reference area groups specified by the reference area specifying unit 2.
A method of specifying the plurality of reference region groups will be described with reference to the drawings. FIG. 9 is a schematic diagram showing a specific outline of the reference area group. In FIG. 9, first, the reference area specifying unit 2 specifies a first reference area group (# 1) including the minimum reference areas A _{1 to} E ₁ . Next, the size of the partial region is increased by one step, and reference regions A _{2 to} E ₂ having different positions relative to the pixel of interest and the same size and shape are specified as a second reference region group (# 2). I do. In the third embodiment, it is assumed that the number of specified reference area groups is three.
Becomes 3. In FIG. 9, the reference area groups # 1 and # 2
Are shown, and the specification of the reference region group (# 3) of the third edition in which the size of the partial region is the maximum is omitted.

Next, the reference value calculation section 3 calculates reference values for the specified plurality of reference region groups. In the third embodiment, the reference value calculation unit 3 calculates reference values for a plurality of reference regions for one reference region group. For example, the reference area group (# 1) is composed of five reference areas A _{1 to} E ₁ , and the reference value is calculated for all of them. Hereinafter, the reference value a ₁ reference value calculated from the reference area A _1, the reference region B ₁ and the reference value b ₁ the reference value calculated from the reference area reference value c ₁ of the reference value calculated from C ₁
The reference value calculated from the reference area D ₁ is referred to as a reference value d _1, and the reference value calculated from the reference area E ₁ is referred to as a reference value e ₁ .
Since the reference area specifying unit 2 specifies a plurality of reference area groups (# 1 to # 3) as described above, the reference value calculation unit 3 sets a plurality of reference values (a
Repeating the calculation of _{1 ~e 1, a 2 ~e 2} , a 3 ~e 3).

Next, the pixel determination section 4a performs image detection by comparing the original pixel value of the pixel of interest with a reference value corresponding to the pixel. Hereinafter, the process of the pixel determination unit according to the third embodiment will be described with reference to the drawings. FIG. 10 is a processing flowchart of the pixel determination processing. In the following, n is a variable that specifies a reference area group. First, n is initialized (step S
31), and compares the minimum and the reference value a _1, which is calculated from the reference area A ₁ and the original pixel value of the pixel of interest (step S32).
Next, it is determined whether or not the pixel of interest is a high-luminance character portion by using Equation 3 as condition 1 for high-luminance character portion determination (step S33). As a result, if the conditions 1 are determined to a high intensity characters section (step S33, Yes), subsequently the original pixel value of the pixel of interest and the reference value b ₁ to e ₁ of the reference area group (n = 1) A comparison is made (step S34). Here, the following Expression 5 is used as the condition 2 for the high brightness character portion determination.

Original pixel value−MAX (reference value b _n , reference value c _n , reference value d _n , reference value e _n )> C ₃ (Expression 5) where MAX is an operator indicating the maximum value. C ₃ is a constant.

As a result, if it is determined in step S35 that the character portion is a high-luminance character portion for condition 2, (step S3
5, Yes), the pixel of interest is determined to be a high-luminance character portion (step S36), and the result of determination as to whether the pixel is a high-luminance character portion is output as 1 and the result of determination as to whether it is a low-luminance character portion is output as 0. ([1, 0]), the pixel determination processing for the pixel of interest ends.

If the condition 1 for determining the high-luminance character portion is not satisfied in step S33 (step S33,
No), Equation 4 is used as the condition 1 for the low-luminance character portion determination.
It is determined whether the target pixel is a low-luminance character portion (step S37). As a result, if the conditions 1 are determined to a low intensity characters section (step S37, Yes), similarly to the case of the detection of the high luminance character unit, the reference value b ₁ to e ₁ and the original pixel value of the pixel of interest Are compared (step S38). Here, the following Expression 6 is used as the condition 2 for the low luminance character portion determination.

MIN (reference value b _n , reference value c _n , reference value d _n , reference value e _n ) −original pixel value> C ₄ (Equation 6) where MIN is an operator indicating the minimum value, C ₄ are constants.

As a result, if it is determined in step S39 that the character portion is a low-luminance character portion for condition 2, (step S3
9, Yes), the pixel of interest is determined to be a low-luminance character portion (step S40), and the result of determination as to whether the pixel is a low-luminance character portion is output as 1 and the result of determination as to whether or not it is a high-luminance character portion is output as 0. ([0, 1]), the pixel determination process for the pixel of interest ends.

In this pixel determination process, the reference values b _n to
Comparison with e _n has a meaning as assays taking into account the direction of the luminance change of the original pixel values. In the actual pixel determination process, _if only the reference value an calculated from the reference region centering on the pixel of interest is used in the pixel determination process, it is located in the reference region and around the character portion to be detected. In some cases, an edge region having a large variation in shading is determined as a character portion. Therefore, partial regions including the target pixel at the upper, lower, left, and right ends are also specified as reference regions, and reference values are calculated. Then, pixels larger than all of these reference values (pixels smaller than all) are referred to as high-luminance character portions (low-luminance characters). Further, pixel determination is performed according to the condition 2 of the high-luminance character part determination (low-luminance character part determination) that the character part is a character part), so that the luminance is higher (low luminance) than the background part, such as a line element constituting a character. It is possible to reliably determine a pixel in a proper character portion.

If no low-luminance character portion or high-luminance character portion is determined in steps S33 and S37 (step S37, No), or
If the high brightness character portion determination condition 2 is not satisfied in 35 (step S35, No), or step S3
If the condition 2 for the low-luminance character portion determination is not satisfied in Step 9 (Step S39, No), after adding 1 to n (Step S41), n is larger than the number N of reference areas (= 3). It is determined whether the value has been reached (step S42). If the result of this determination is that n is N or less (step S4
2, No), the pixel determination process is repeated for the reference region group formed of the next larger partial region (step S32)
To S35, steps S37 to S39). Step S42
When n becomes larger than N (step S4).
2, Yes), the pixel of interest is determined to be a background part (step S43), and both the determination result of whether the pixel is a low-luminance character part and the determination result of whether it is a high-luminance character part are output as 0 ([ 0, 0]), and ends the pixel determination process for the target pixel.

The pixel determination section 4a repeats the above-described pixel determination process for all pixels included in a predetermined area for performing image detection in the input image while sequentially changing the pixel of interest. The detected image storage unit 5a stores these outputs, generates a binary image of a high-brightness character from the determination result as to whether it is a high-brightness character part, and generates a low-brightness character from the determination result as to whether it is a low-brightness character part. A binary image of a character portion is generated. Finally, the character recognition unit 6a
Performs character recognition from the two binary images generated by the detected image storage unit 5a.

As described above, the character recognition apparatus according to the third embodiment can perform image detection and character recognition even when the size of a character or the thickness of a line element included in an input image is not clear in advance. It is possible. In addition, a character recognition device that can accurately detect an image and further improve the accuracy of character recognition by reliably determining an edge region having a large variation in shading around a character portion as a background portion. Obtainable.

In the third embodiment, as the reference area included in the reference area group specified by one size,
Five partial regions centered on the pixel of interest and partial regions including the pixel of interest at the center of the right end, the center of the left edge, the center of the upper end, and the center of the lower end are specified. Any number of partial regions may be used, and the relative positional relationship between the target pixel and the reference region is not necessarily specified as described above. Although the number of reference area groups specified by the reference area specifying unit is three, any number may be used as long as the number is plural. Further, in the third embodiment, the character recognition device of the second embodiment includes a reference region specifying unit that includes a target pixel and specifies a plurality of reference regions having different positions and sizes relative to the target pixel. Thus, the accuracy of image detection is improved, but the same effect can be expected by applying the reference area specifying unit to the image recognition device of the first embodiment.

Embodiment 4 In the fourth embodiment, the pixel determination unit determines whether the pixel of interest is a high-luminance character portion,
The detection result storage unit outputs a ternary determination result as to whether it is a low-luminance character portion or a background portion, and stores the result as a ternary image, and performs character recognition based on the ternary image. The unit recognizes high-brightness characters and low-brightness characters.

The character recognition apparatus according to the fourth embodiment is different from the above-described character recognition apparatus shown in FIG. 1 in the pixel determining section, the detected image storage section, and the character recognition section. Since they are the same, the same reference numerals are given and the description is omitted. Reference numeral 4 denotes a pixel determination unit that determines whether the target pixel is a high-luminance character portion, a low-luminance character portion, or a background portion, and outputs a ternary determination result. A detection image storage unit that stores a ternary determination result output from the pixel determination unit 4 and generates a ternary image, and 6 detects a high-luminance character and a low-luminance character from the ternary image generated by the detection image storage unit 5 This is a character recognition unit that recognizes a luminance character. Further, since the entire processing flow of the character recognition device according to the fourth embodiment is the same as the processing flow of the first embodiment shown in FIG.
4 and S5) will be described.

The pixel determination section 4 compares the original pixel value of the target pixel with the reference value, and determines whether the target pixel is a high-luminance character portion, a low-luminance character portion, or a background portion. And outputs the determination result as a ternary determination result. Hereinafter, the processing flow of the pixel determination process in the pixel determination unit according to the fourth embodiment will be described with reference to FIG.

First, a variable n indicating the reference area is initialized with 1 (step S21), and the reference value calculated from the minimum reference and the original pixel value of the pixel of interest are compared according to the above-described equations (3) and (4). Step S22). Next, it is determined whether or not the condition of the high-brightness character portion of Expression 3 is satisfied (step S23). If the condition is satisfied (step S23, Ye)
s), the target pixel is determined to be a high-luminance character portion (step S)
24), the pixel determination result of the target pixel is set to 2, and the pixel determination process ends.

Next, if the condition for the high-luminance character portion is not satisfied in step S23 (step S23, No),
Further, it is determined whether or not the condition of the low-luminance character portion of Expression 4 is satisfied (step S25). If the condition is satisfied (step S25, Yes), the target pixel is determined to be the low-luminance character portion (step S26). ), The pixel determination result of the target pixel is set to 1, and the pixel determination process ends.

If the condition for the low-luminance character portion is not satisfied in step S25 (step S25, N
o) After adding 1 to n (step S27), it is checked whether n has become larger than N, that is, whether or not comparison with all reference values has been completed (step S28). n is N
In the following cases (Step S28, No), Step S28
22, the pixel determination for the target pixel is repeated based on the reference value calculated from the reference area one step larger than the reference value compared last time (steps S23 and S2).
5). If n becomes larger than N (step S
28, Yes), the pixel is determined to be the background part (step S29), and the pixel determination result of the target pixel is set to 0, and the pixel determination process ends.

The pixel judging section 4 repeats the above-mentioned pixel judging process for all the pixels included in the predetermined area for performing the image detection in the input image while sequentially changing the pixel of interest, and the ternary pixel judgment result (2 / 1/0). Next, the detected image storage unit 4 stores the ternary pixel determination result output from the pixel determination unit as a ternary image. In this ternary image, one detection image includes both the detection result of the high luminance character portion and the detection result of the low luminance character portion.

Finally, the character recognition section 6 performs character recognition from the ternary image. In the fourth embodiment, first, character recognition is performed by regarding a high-luminance character portion of a ternary image, that is, only a pixel whose pixel determination result is 2 as a character portion and the other as a background portion. Next, character recognition is performed by assuming that only the low-luminance character portion, that is, the pixels for which the pixel determination result is 1, is a character portion, and the others are background portions.

As described above, the character recognition apparatus according to the fourth embodiment can perform image detection and character recognition even when the size of a character or the thickness of a line element included in an input image is unknown. It is. Furthermore, since the detection results of both the high-brightness character portion and the low-brightness character portion are generated as one ternary image, and character recognition is performed based on this, the detection result of the high-brightness character portion and the detection of the low-brightness character portion are performed. The results are two independent two
Compared to the case of storing in a value image, the amount of information as the entire detected image can be reduced, and even with a small storage capacity, image detection and character recognition of a low-luminance character portion and a high-luminance character portion can be performed. It is possible to obtain a character recognition device that can perform the operation.

Embodiment 5 The character recognition device according to the fifth embodiment differs from the character recognition device according to the first embodiment shown in FIG. 1 only in the reference area specifying unit, and the other configurations are the same. The description is omitted. 2
Is a reference region specifying unit that specifies a plurality of partial regions having different sizes and shapes as reference regions.

In the fifth embodiment, the reference area specifying unit 2
Specifies a plurality of partial areas having different sizes and shapes as reference areas (# 1 to #N, where N is the number of specified reference areas) by the method shown in FIG. FIG. 11 is a schematic diagram illustrating a method of specifying the reference area by the reference area specifying unit 2.
In FIG. 11, reference numerals 30 to 32 denote partial areas specified as reference areas corresponding to the target pixel 7. In FIG.
In particular, the case where the reference region specifying unit specifies three partial regions having different sizes and shapes around the target pixel as the reference region is shown. In this case, the number N of the specified reference regions is three. The smallest partial area 30 is a square area centered on the pixel of interest, and is specified as the first reference area (# 1). Next largest partial area 31
Is a rectangular area that is long in the horizontal direction centering on the target pixel, and is specified as the second reference area (# 2). The largest partial region 34 is a rectangular region that is long in the horizontal direction around the pixel of interest and is specified as the third reference region (# 3).

FIG. 12 is a schematic diagram showing a specific outline of the reference area according to the fifth embodiment. In FIG. 12, reference numeral 33 denotes an image of a dial. The input image includes an image pickup 33 of a dial to be subjected to image detection processing, and a scene image portion 24 outside of the image detection processing to be subjected to image detection processing. Furthermore, the image 33 of the dial includes a character portion having a higher luminance than the background portion indicated by the character “74”.

As shown in FIG. 12, when the upper and lower ends of the character portion to be detected are close to the imaging edge of the dial, it is assumed that the shapes of the reference areas are all square and the same. Then, in the reference area having a large area, the scene image portion 24 outside the image pickup of the dial is included in the reference area at a considerable ratio. In this case, the reference value calculated by being influenced by the brightness of the scene image portion changes greatly, and the pixel determination portion cannot accurately perform the pixel determination process on the target pixel 7. Therefore, in the reference region specifying unit according to the fifth embodiment, in order to reduce the influence of the scene image portion included in the reference region as much as possible, the partial regions 30 to 32 are used.
, The large partial areas are specified as reference areas, as horizontally long rectangular areas.

Thereafter, similarly to the character recognition device according to the first or second embodiment, the reference value calculation unit 3 calculates a reference value for each of the reference areas specified by the reference area specification unit 2 and outputs the reference value to the pixel determination unit. Reference numeral 4 compares the reference value with the original pixel value of the pixel of interest to detect a character portion, and a detected image storage unit 5
Generates a binary image from the detection result of the pixel determination unit 4, and finally, the character recognition unit 6 performs character recognition from the binary image.

As described above, the character recognition apparatus according to the fifth embodiment can perform image detection and character recognition even when the size of a character portion included in an input image or the thickness of a line element is unknown. In addition, even in the case where a symbol portion such as a character or a figure to be detected is close to the edge of the imaging of the dial, in consideration of the arrangement of the character portion in the imaging of the dial,
It is possible to accurately detect a character portion while minimizing the influence of a scene image portion outside the image of the dial plate.

In the fifth embodiment, the shape of the reference region is specified so that the larger the reference region, the longer the rectangle in the horizontal direction. However, the larger the larger reference region, the longer the rectangle in the vertical direction. good. Also, for example, the shape of the partial area may be changed such that a small one is a square, a large one is a circle or a triangle, and conversely, a small one is a circle and a large one is a square or a rectangle. May be changed. Further, in the fifth embodiment, a plurality of reference regions centering on the target pixel are specified to perform the pixel determination. However, as in the third embodiment, only the reference region centering on the target pixel is used. Alternatively, it is naturally possible to specify these as a reference region group including the reference region including the target pixel at the right end center, the left end center, the upper end center, and the lower end center, and to perform the pixel determination process based on the reference region group. .

As described above, the character recognition device according to the first to fifth embodiments includes the detected image storage units 5 and 5a and the character recognition units 6 and 6a, and uses the pixel determination results output from the pixel determination units 4 and 4a. A binary image is generated and character recognition is performed, but the detected image storage unit and the character recognition unit are omitted, and the image input unit, the reference area specifying unit, the reference value calculation unit, and the pixel determination unit are configured. It is possible to obtain an image detecting device.

In such an image detection device, first, the image input section inputs a digital signal of a predetermined area in an original image to be subjected to image detection to obtain an input image. Next, the reference area specifying unit specifies a plurality of partial areas included in the input image as reference areas. Next, the reference value calculation unit calculates a reference value for each of the specified reference regions. Finally, the pixel determination unit compares the original pixel value of the input image with a plurality of reference values corresponding to the pixel, performs image detection, and directly outputs a pixel determination result.

In such an image detection apparatus, even if the size or the thickness of a line element of a symbol portion such as a character or a figure included in an input image is not clear beforehand, a character or a symbol to be detected is accurately detected. It is possible to detect a symbol part such as a figure. Further, even when a single input image includes a plurality of symbol parts such as a plurality of characters and figures having different sizes and different line elements, they can be simultaneously detected. Therefore, for example, by applying the present invention to an automatic detection device for a traffic sign, etc., when the input image to be recognized includes a character, a symbol, a simple diagram, etc. The effect that accurate detection is possible regardless of the thickness is expected and useful.

[0093]

As described above, according to the present invention, a plurality of reference regions having different sizes including a pixel of interest are specified, and reference values are calculated for each of these reference regions. By comparing the reference value with the pixel value of the pixel of interest and performing pixel determination of the pixel of interest, the size and line element thickness of the symbol portion such as characters and figures included in the input image Irrespective of the above, it is possible to accurately detect the symbol portion. Therefore, it is possible to perform image detection even when the size and the thickness of the line element are not clear beforehand, and a plurality of symbol parts having different sizes and the thickness of the line element are included in one input image. Even in this case, all of them can be simultaneously detected by a single image detection process. Furthermore, even when the pixel of interest is located at a position near the end of the predetermined area to be subjected to image detection, image detection can be performed accurately by reducing the influence of the scene image portion.

Further, a plurality of reference regions including the target pixel and having different positions and sizes relative to the target pixel are specified, and a reference value is calculated for each of these reference regions.
By comparing the reference value with the pixel value of the pixel of interest in ascending order of the reference area and performing the pixel determination of the pixel of interest, the size of a symbol portion such as a character or a figure included in the input image can be reduced. Regardless of the thickness of the line element, it is possible to accurately detect the symbol portion. Therefore, image detection can be performed even when the size and the thickness of a line element of a character or figure included in the input image are not clear beforehand, and the size and the line element of one input image can be detected. Even if there are multiple parts with different thicknesses,
All of these can be detected simultaneously by one image detection process. Furthermore, since the edge region having a large variation in shading around the symbol portion is reliably determined as the background portion, the accuracy of image detection can be improved.

Further, since a plurality of reference areas having different shapes including the target pixel are specified, even if the size of a character or the thickness of a line element included in an input image is unknown, image detection can be performed. In addition to being able to perform character recognition, even if a symbol portion such as a character or a figure to be detected is close to an end of the predetermined region to be image-detected, the predetermined region to be image-detected In consideration of the arrangement of the symbol portions in the image, the influence of the scene image portion outside the predetermined region to be detected can be reduced as much as possible, and the accuracy of the image detection can be improved.

Further, the image detecting apparatus having the above-described effects further includes a detected image storing means for storing the image detection result as a binary image, and a character recognizing means for recognizing characters from the binary image. Accordingly, even if the size of a character and the thickness of a line element included in the input image are not clear beforehand, it is possible to accurately perform character recognition. Also,
Even if a single input image contains a plurality of characters with different sizes and line element thicknesses, it is possible to recognize all these characters simultaneously with a single image detection process and character recognition process. it can.

Also, it is determined whether or not the pixel of interest is a high luminance symbol portion, and the result of the determination is output. At the same time, it is determined whether the pixel is a low luminance symbol portion, and the determination result is output. Is stored as a binary image of a high-luminance symbol portion, and the result of determination as to whether or not the symbol portion is a low-luminance symbol portion is stored as a binary image of the low-luminance symbol portion. A high-brightness character portion is recognized from the high-brightness symbol portion binary image, and a low-brightness character is recognized from the low-brightness symbol portion binary image. Therefore, when the input image includes one or both of the high-luminance symbol portion and the low-luminance symbol portion, the size of each symbol portion and the thickness of the line element are not clear in advance. In case , It is possible to perform accurate image detection and character recognition. Furthermore, for one input image, both a high-luminance symbol portion and a low-luminance symbol portion can be detected by one image detection process, thereby avoiding an increase in processing time required for the image detection process. be able to.

It is determined whether the pixel of interest is a high-luminance symbol part, a low-luminance symbol part, or a background part, and the result of the determination is output as a ternary value. By storing the result as a ternary image and recognizing both high-brightness characters and low-brightness characters from the ternary image, the size of the symbol portion included in the input image and the Even when the thickness is unknown, the image detection processing and the character recognition processing can be performed. Furthermore, since character recognition is performed based on one ternary image, image detection and character recognition of low-brightness characters and high-brightness characters can be performed without increasing the storage capacity required for character recognition. It is possible to obtain a simple character recognition device.

[Brief description of the drawings]

FIG. 1 is a first embodiment of a character recognition device according to the present invention.
FIG.

FIG. 2 is a processing flowchart of the character recognition device according to the first embodiment of the present invention.

FIG. 3 is a schematic diagram illustrating a reference area specified by a reference area specifying unit according to the first embodiment of the present invention.

FIG. 4 is a process flowchart of a pixel determination process according to the first embodiment of the present invention.

FIG. 5 is a schematic diagram showing a specific outline of a reference area according to the first embodiment of the present invention.

FIG. 6 is a second embodiment of the character recognition device according to the present invention.
FIG.

FIG. 7 is a process flowchart of a pixel determination process according to the second embodiment of the present invention.

FIG. 8 is a schematic diagram showing a reference area specified by a reference area specifying unit according to the third embodiment of the present invention.

FIG. 9 is a schematic diagram showing a specific outline of a reference area group according to the third embodiment of the present invention.

FIG. 10 is a process flowchart of a pixel determination process according to the third embodiment of the present invention.

FIG. 11 is a schematic diagram showing a reference area specified by a reference area specifying unit according to the fifth embodiment of the present invention.

FIG. 12 is a schematic diagram showing a specific outline of a reference area according to the fifth embodiment of the present invention.

FIG. 13 is a configuration diagram showing a conventional character string detection device.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Image input part 2 Reference area | region identification part 3 Reference value calculation part 4, 4a Pixel determination part 5, 5a Detected image storage part 6, 6a Character recognition part 7 Target pixel 8, 9, 10, 15, 16, 17, 18, 19, 20,
21, 22, 23, 25, 26, 27, 28, 29, 3
0, 31, 32 Partial area specified as reference area 11, 33 Image of dial 12, 13, 14 Pixel 24 Scene image part

Claims (14)

[Claims] 1. An image input means for inputting a digital signal of a predetermined area in a digital image for detecting a symbol part, and a plurality of partial areas having different sizes including pixels to be determined whether or not the symbol part. Reference area specifying means for specifying a reference area as a reference area; reference value calculating means for calculating a reference value for each of the plurality of reference areas; and a pixel value obtained as a digital signal of a pixel to be determined corresponding to the pixel. An image detection device comprising: a pixel determination unit that sequentially compares the plurality of reference values with the reference area in ascending order of the size of the reference area to determine whether the pixel is a symbol part. 2. The method according to claim 1, wherein the reference area specifying unit is configured to specify a plurality of partial areas having different sizes around a pixel to be determined as a reference area. Image detection device. 3. The reference region specifying means is configured to specify, as a reference region, a plurality of partial regions including a pixel to be determined and having different positions and sizes relative to the pixel. The image detection device according to claim 1, wherein: 4. The reference area specifying means according to claim 1, wherein said reference area specifying means specifies a plurality of partial areas having different shapes including pixels to be determined as reference areas. An image detection device according to any one of the above. 5. The image detection device according to claim 1, further comprising: a detection image storage unit configured to store a determination result of the pixel determination unit as a binary image; A character recognition device, comprising: character recognition means for recognizing a character when the symbol portion is a character. 6. A pixel determining unit sequentially compares a pixel to be determined with a plurality of reference values, and determines whether the pixel is a symbol portion having a higher luminance than a portion other than the symbol portion around the symbol portion. And outputting a result of the determination, determining whether the pixel is a symbol part having a lower luminance than a part other than the symbol part, and outputting the result of the determination. Is stored as a binary image of the high-luminance symbol portion, and the result of the determination as to whether or not the symbol portion is a low-luminance symbol portion by the pixel determining means is stored as the low-luminance symbol portion. The character recognizing means recognizes high-brightness characters from the high-brightness symbol part binary image, and stores a low-brightness symbol part binary image from the low-brightness symbol part. Recognize the characters Character recognition device described in claim 5, characterized in that it is configured to. 7. A pixel determination unit sequentially compares a pixel to be determined with a plurality of reference values, and determines whether the pixel is a symbol part having a higher luminance than a part other than the symbol part, or a pixel other than the symbol part. It is configured to determine whether the mark portion is a low-luminance symbol portion or a portion other than the symbol portion, and to output the determination result as a ternary value. 6 is stored as a ternary image, and the character recognizing means is configured to recognize both a high-luminance character and a low-luminance character from the ternary image. Character recognition device. 8. An image input step of inputting a digital signal of a predetermined area in a digital image for detecting a symbol part, and a plurality of partial areas having different sizes including pixels to be determined as to whether or not the symbol part. A reference region specifying step of specifying a reference value as a reference region; a reference value calculating step of calculating a reference value for each of the plurality of reference regions; and a pixel value obtained as a digital signal of a pixel to be determined corresponding to the pixel. A pixel determination step of sequentially comparing the plurality of reference values and the reference area in ascending order of size to determine whether the pixel is a symbol part. 9. The image detection method according to claim 8, wherein in the reference area specifying step, a plurality of partial areas having different sizes around a pixel to be determined are specified as reference areas. . 10. The reference area specifying step includes specifying, as a reference area, a plurality of partial areas including a pixel to be determined and having different positions and sizes relative to the pixel. Item 9. The image detection method according to Item 8. 11. The reference area specifying step includes specifying a plurality of partial areas having different shapes including pixels to be determined as reference areas.
0. The image detection method according to any one of 0. 12. The image detection method according to claim 8, further comprising: a detection image storage step of storing a determination result of the pixel determination unit as a binary image; A character recognition step of recognizing the character when the symbol portion is a character. 13. The pixel determining step includes sequentially comparing a pixel to be determined with a plurality of reference values to determine whether the pixel is a symbol part having a higher luminance than a part other than the symbol part around the symbol part. And outputting a result of the determination, determining whether the pixel is a symbol part having a lower luminance than a part other than the symbol part, and outputting the determination result. Is stored as a binary image of a high-luminance symbol portion, and the result of determination as to whether or not the symbol portion is a low-luminance symbol portion is stored as a binary image of the low-luminance symbol portion. Storing the character image as a value image, and recognizing a high-luminance character from the binary image of the high-luminance symbol portion and recognizing a low-luminance character from the binary image of the low-luminance symbol portion. Characterized by Character recognition method described in Motomeko 12. 14. The pixel determination step includes sequentially comparing a pixel to be determined with a plurality of reference values, and determining whether the pixel is a symbol part having a higher luminance than a part other than the symbol part or a part other than the symbol part. It is determined whether it is a low-luminance symbol part or a part other than the symbol part, and the result of the determination is output as a ternary value. 13. The character recognition method according to claim 12, wherein the character is stored as an image, and in the character recognition step, both high-brightness characters and low-brightness characters are recognized from the ternary image.