JP4423239B2 - Image reading device - Google Patents

Description

  The present invention relates to an image reading apparatus, and more particularly to an image reading apparatus that can perform high-resolution image reading using a color camera.

  In recent years, the number of pixels in image input devices has been increasing. In particular, the number of pixels of a digital camera is remarkably increased, and a CCD sensor having a number of pixels equivalent to 10 million pixels has been commercialized. At present, a digital camera of 10 million pixel class is as expensive as several hundred thousand, but it is considered that it becomes cheaper as the number of pixels increases and it is generally used. When a digital camera of 10 million pixel class becomes widespread, the object to be photographed expands, and it is conceivable to use a document not only as a landscape and a person but also as a photograph object.

  In a mobile phone with a camera, there is already a device that recognizes characters by reading an image about the size of a word. Some commercially available software performs image correction for a digital camera semi-automatically on image data obtained by reading a document with a digital camera and generates text and general-purpose files by character recognition.

  For example, when an A4 size original of 298 mm long and 210 mm wide is scanned at 300 dpi (12 pixels / mm), the number of pixels is 298 * 12 (3576 pixels) and the width is 210 * 12 (2520 pixels). Requires about 9 million pixels with 3576 pixels * 2520 pixels.

  That is, with a digital camera equivalent to 10 million pixels, an A4 size document can be read at 300 dpi. In this way, digital cameras and camera-equipped mobile phones that can be replaced with flatbed scanners will emerge if all A4-size originals can be read at a high resolution of 300 dpi, as well as lexical characters and parts of documents. become. In addition, if characters and the like can be read at a resolution of 300 dpi, it can be sufficiently used as data used in an OCR application.

  On the other hand, the use of barcodes has been progressing in recent years, and it has rapidly spread to various fields since it has been used in POS systems such as convenience stores. In particular, recently, two-dimensional codes have been developed in order to accommodate needs such as storing more information and printing in a smaller space, and the spread thereof is expanding.

  As an input device for reading a barcode, a monochrome sensor dedicated to barcode reading is usually used. The reason is that inexpensive and high resolution can be obtained, and high recognition can be obtained by barcode recognition. This sensor cannot acquire a color image, but there is no problem if a color image is not required. However, in the future, it is desirable to use a color sensor for barcode reading so that a color image can be obtained.

  In recent years, there has been a need to use an ordinary color camera that normally uses a color sensor and additionally read a barcode. For example, a usage method for reading a barcode with a color camera mounted on a mobile phone has appeared.

Patent Document 1 is known as a barcode reader using a color sensor. In this apparatus, when the barcode is read, the barcode is read many times and a weighted average is performed to reduce noise and improve the recognition rate. Patent Document 2 discloses an information providing service device that selects and sets a photo mode and a QR code (two-dimensional code) reading mode and performs processing according to the set mode. Patent Document 3 discloses that the reading accuracy of a code reader is improved by processing a one-dimensional code with a line CCD and processing a two-dimensional code with an area CCD.
JP 2003-203198 A JP 2002-111909 A Japanese Patent Laid-Open No. 9-259215

 A reading apparatus using a color sensor as described above has a small number of pixels in the color sensor to be used, and has a low resolution because it is a single-plate color sensor, and thus has a low recognition rate. However, as described above, it is possible to cover a decrease in the recognition rate by continuously reading one code many times. As described above, at present, the level has not reached a level where one two-dimensional code can be satisfactorily recognized by only one reading.

  However, in the future, with the widespread use of two-dimensional codes, it may be necessary to read a plurality of two-dimensional codes at a time. For example, a plurality of two-dimensional codes displayed on A4 size paper can be read at one time, or a reading on which a two-dimensional code is written together with color images and characters can be read simultaneously to obtain a color image and a two-dimensional code. There may be a need to recognize both codes well. In addition, there is a possibility that color barcodes will become widespread. In any case, it seems necessary to increase the resolution of image reading using a color sensor.

  By the way, as described above, the method of covering a decrease in the recognition rate by reading one code several times continuously and taking a weighted average inevitably requires a reading time corresponding to the number of times of reading. If only users who rarely read barcodes are targeted, there is no problem. However, for users who frequently use barcodes, reading time becomes a big problem. In addition, it is expensive to provide two sensors, a line CCD and an area CCD, for reading the one-dimensional code and the two-dimensional code. For example, if a one-dimensional code can be read by an area CCD, a two-dimensional code can be read. In this case, the improvement of the resolution when using the area sensor becomes a problem.

  In a single-plate color camera, each pixel corresponds to one color of RGB color filters. For example, the luminance Y of one pixel can be obtained by an equation of Y = 0.299 * R + 0.587 * G + 0.114 * B, assuming that the luminance information R, G, B of each pixel of RGB.

  However, the single plate sensor has only luminance information of one color for one pixel. For this reason, luminance information of surrounding pixels (RGB) is used. For this reason, the smoothing process is performed when the luminance Y is converted. As a result, the resolution of the image or the difference in image information between pixels is reduced, and the recognition rate is lowered.

  Further, when acquiring a color image, each luminance information of RGB is necessary for each pixel, but since each pixel has only one color of RGB color filters, information on other colors is necessary for the surroundings. There is no choice but to substitute using information of pixels having color filters of various colors. That is, it is obtained by referring to information on pixels of different surrounding colors that are shifted from the pixel position, not the RGB color signal corresponding to the original pixel position. This degrades the resolution.

  By the way, the color of the currently popular barcode is black. I don't see colored barcodes. Therefore, assuming that the barcode color can be limited to black, it can be converted into a luminance value without reducing the resolution of the image or the difference in image information between pixels.

  For example, in general, a camera performs white balance and gray balance processing, and in the case of each color pixel, that is, the sensor color array shown in FIG. The outputs corresponding to the pixels of the filter, the Gb filter, and the B filter are adjusted.

  Since the gray area between black and white is an achromatic color, the output is the same regardless of the color of each color filter. That is, if the luminance value of each pixel is the same (R = G = B), the luminance Y = 0.299 * R + 0.587 * G + 0.114 * B is Y = R, Y = G, Y = B. . That is, the luminance value of each pixel can be considered as the luminance value of that pixel.

  That is, conventionally, the luminance value of the one pixel is generated by using the R filter pixel, the Gr filter pixel, the Gb filter pixel, and the B filter pixel around one pixel to be generated. Was. However, as described above, in the case of a monochrome image only, when converting to a luminance value, the pixel information (luminance information) of only the pixel position is used without referring to the surrounding pixels of the pixel position where the luminance value is generated. A luminance value corresponding to the pixel position can be obtained.

  That is, it is possible to convert to a luminance value without reducing the image resolution or the difference in image information between pixels, and it is possible to suppress a decrease in recognition rate. That is, it can be seen that in the gray region, even if a color sensor is used, the resolution of the image is high, and an image quality equivalent to that of a monochrome sensor can be realized without impairing the difference in pixel information between pixels.

  At this time, if the lens has chromatic aberration, red or blue light enters the edge of the gray region when reading the gray region. In such a case, simply generating a luminance value of one pixel from the R filter pixel, the Gr filter pixel, the Gb filter pixel, and the B filter pixel, the value fluctuates depending on the color, resulting in noise. turn into. Since a lens with little chromatic aberration is expensive, it is a great merit to make an inexpensive lens with chromatic aberration available. For this purpose, a noise removal process for reducing noise is required for an image read by an optical system having chromatic aberration.

  The present invention has been made based on these findings, and provides an image reading apparatus capable of obtaining a high-resolution monochrome image using a camera equipped with a color sensor.

  In order to solve the above problems, the present invention employs the following means.

  A color camera having a color sensor and capable of photographing a document including an area on which a two-dimensional barcode is printed and outputting pixel signals of at least three colors, and only pixels to which each of the pixel signals belongs A barcode area image processing unit that generates a one-pixel color signal using the signal, a pixel signal to which any one of the pixel signals belongs, and a pixel signal around the pixel A non-barcode area image processing unit for generating a barcode, and a selector for switching and outputting the output of the barcode area image processing unit and the non-barcode area image processing unit.

  Since the present invention has the above-described configuration, it is possible to provide an image reading apparatus that can obtain a high-resolution monochrome image using a camera including a color sensor.

  Hereinafter, the best embodiment will be described with reference to the accompanying drawings. FIG. 1 is a diagram illustrating an image reading apparatus according to an embodiment of the present invention. In FIG. 1, an image reading apparatus includes a camera head 11, a column 12 that fixes the camera head, a camera unit 1 that includes a document table 13 on which an imaging target is placed, and an image processing unit 2 that processes an image captured by the camera unit 1. It consists of and. The camera head 11 includes a single plate type color sensor. Further, the imaging direction of the camera head 11 is downward, and the document 14 placed on the document table 13 can be read. The document 14 is provided with a color image area 15 or a barcode area 16 on which a barcode is printed.

  The image processing unit 2 stores an image signal obtained by reading the original with the color sensor of the camera unit 1 in the storage unit 3. Next, the stored image signal is read, and the barcode area determination unit 6 determines whether or not the read portion is a barcode area in an image obtained by reading the document.

  Here, the determination result of the barcode area of the area in the image is stored in advance as determination data. Then, the image signal stored in the storage unit 3 is read again by the barcode area image processing unit 4 and the non-barcode area image processing unit 5 and processed. The selector 7 selects two series of output signals subjected to the processing based on the determination data and outputs them as one image signal.

  Here, the determination data determined in advance as the barcode area among the areas in the image is used to read again. However, if the barcode area is known, the image signal acquired from the camera unit 1 is not read again from the image signal stored in the storage unit 3, and the barcode area image processing unit 4 and the non-barcode area The image processing unit 5 may read and perform the processing, and the selector 7 may select the two series of outputs subjected to the processing based on the determination data and output them as one image signal.

  Further, when the image is read in advance for determining the barcode area, the barcode area can be determined by thinning out and reading the image. By doing so, the processing time can be shortened. For example, in a reading apparatus that performs “pixel shift” reading for high resolution, an image is read a plurality of times at a low resolution to obtain one high-resolution image. The barcode area can be determined from the low resolution image. In this case, it is possible to finish the determination as to whether or not it is a barcode area until a plurality of images necessary for generating a color signal or luminance (grayscale) signal are prepared.

  In FIG. 1, image data read by a color sensor constituting the camera head 11 is input to a barcode area image processing unit 4, a non-barcode area image processing unit 5, and a barcode area determination unit 6. The barcode area determination unit 6 detects a barcode area in the image data, or sets and stores data corresponding to the barcode area in advance.

  When obtaining a color image, an R color filter pixel signal, a G color filter pixel signal, and a B color filter pixel signal for one pixel obtained from the color sensor are converted into one pixel R, G, and B color signals. Convert. At this time, as will be described later, the barcode area image processing unit 4 generates a color signal of one pixel R, G, B from one color image signal of the R, G, B color filters to be processed. In this case, since RGB is generated from the image signal at the same position, the signal difference between the pixels is large, and the resolution of the image is improved. However, color reproduction is not normal in a color image or the like.

  The non-barcode area image processing unit 5 generates one pixel R, G, and B by referring to the information of the surrounding color filter pixel signals as well as the color filter pixel signals corresponding to the processing pixels, as will be described later. To do. For this reason, although the resolution of the image is deteriorated, a normal color can be reproduced.

  These two outputs are switched by a selector 7 operated by a determination signal from the barcode area determination unit 6 and output as a single color image. When obtaining a gray image, an R color filter pixel signal, a G color filter pixel signal, or a B color filter pixel signal for one pixel obtained from a color sensor is converted into a luminance signal. The area image processing means generates a luminance signal from an image signal of any one of the R, G, B color filters to be processed. That is, since it is generated only with the image signal at the same position, the signal difference between the pixels is large, and the resolution of the image is improved. However, when the read image is a color image, the apparent light and dark reproduction is not normal.

  In addition, the non-barcode area image processing unit 5 generates a luminance signal by referring to information of surrounding color filter pixel signals as well as color filter pixel signals corresponding to the processing pixels. Therefore, although the resolution of the image is deteriorated, normal light and dark can be reproduced. These two outputs are switched by the bar code area determination unit and output to synthesize one color image.

  For this reason, it is possible to acquire a color image, and it is possible to acquire a gray image having a resolution comparable to that of a monochrome sensor reading in the gray region. Therefore, the barcode recognition rate is improved to the same level as a monochrome sensor, and good barcode recognition is possible.

  FIG. 2 is a diagram illustrating details of the barcode area image processing unit 4. The barcode area image processing unit 4 generates a luminance (grayscale) signal or an RGB color signal using the image signal read from the storage unit 3 shown in FIG. Which signal is to be generated is switched by the selector 41 in accordance with a setting signal a set by the user.

  When obtaining a gray scale signal, the image signal read out from the storage unit 3 is selected, and when obtaining a color image, the output of the RGB color generation unit 42 is selected.

  The image signal of the color sensor stored in the storage unit 3 corresponds to one of RGB color filters. Here, the pixel signal corresponding to the red color filter is R, the pixel signal corresponding to the green color filter is G, and the pixel signal corresponding to the blue color filter is B. When the image to be read has no saturation and is gray, R, G, and B have the same output as shown in FIG.

  When R, G, and B do not have the same output, even if a gray image is read, the image does not become gray. For example, as shown in FIG. When B is high, the image becomes bluish. Here, it is assumed that RGB when the camera adjustment is completed and gray is read matches and is balanced. In the case of considering the adjustment process as well, a gray balance processing unit for adjusting the balance of gray RGB may be inserted before or after the storage unit 3. This adjustment process is performed, for example, by reading a gray image in advance before reading the original document, obtaining RGB output characteristics, and obtaining a conversion formula or a converted value for each RGB from the obtained output characteristics or converting the converted values. It is good to store it in a table and convert the table when reading.

  FIG. 3 is a diagram for explaining the processing of the barcode area image processing unit 4. As described above, if the gray and RGB are balanced, the barcode area image processing unit 4 uses Y = R, Y = G when obtaining luminance information for each pixel as shown in FIG. , Y = B, one pixel is converted into a luminance value of one pixel.

  In addition, the RGB color generation unit 42 converts the pixel information of one pixel at that position as it is into RGB information as shown in FIG. 3B regardless of the color filter color.

  For example, if the pixel information of the R filter pixel is R, the generated RGB is R = R, G = R, B = R, and if the pixel information of the G filter pixel is G, R = G, G = G , B = G, and the pixel information of the B filter pixel is B, R = B, G = B, and B = B are output.

  As described above, when generating a grayscale signal or an RGB color signal, conversion can be performed without referring to surrounding pixels, and thus without reducing the resolution of the image or the difference in image information between pixels. .

  FIG. 4 is a diagram for explaining the processing of the non-barcode area image processing unit 5. The non-barcode area image processing unit 5 refers to the image signals of the surrounding color filters when generating a grayscale signal and when generating an RGB color signal. Although this is a general method, it will be briefly described with reference to FIG.

  The non-barcode area image processing unit 5 obtains a luminance signal from Y = 0.299 * R + 0.587 * G + 0.114 * B when converting to a luminance signal. At this time, since one pixel has only one color information, for example, when obtaining the upper left luminance signal Y as shown in FIG. 4A, the image signal R of the upper left R filter and the image signal G1 of the horizontal G filter are obtained. A luminance signal of one pixel is obtained by using four pixels of ', an image signal G2 of the G filter below the R filter, and an image signal B of the B filter on the side. This vertical and horizontal pixel size includes R, G, and B information. Since G has two pixels of G1 ′ and G2 ′, the average value of the two pixels is G. From this R, G, and B, Y = 0.299 * R + 0.587 * G + 0.114 * B A luminance value Y is obtained.

  Also, when acquiring an RGB color signal, it is converted into one pixel RGB using four pixels including information on surrounding RGB as shown in FIG. This conversion uses R, G, and B within the vertical and horizontal pixel size of 2 * 2 as in the case of obtaining the luminance value. For example, R is set to R as it is within the vertical and horizontal pixel size of 2 * 2. , G has two pixels of G1′G2 ′ within the vertical and horizontal 2 * 2 pixel size, and therefore the average value thereof is set to G, and B is generated as B within the vertical and horizontal pixel size of 2 * 2 as it is.

  Note that the number of pixels used when acquiring the RGB color signal can be four or more. The non-barcode area pixel generation unit 5 deteriorates because the resolution of the image is generated from four pixels when generating a grayscale signal and a color signal. There is no problem.

  FIG. 5 is a diagram for explaining an example of a barcode. For example, in the case of a QR code that is a two-dimensional barcode, there are three cut symbols (A, B, C) as shown in FIG. 5, and the area surrounded by them is the QR code area. Therefore, in the case of a QR code, if this cut symbol is detected by pattern matching, a two-dimensional code area can be determined.

  FIG. 6 is a diagram for explaining the processing of the barcode area determination unit 6. For the image signal read from the storage unit 3, the pattern matching unit 61 performs pattern matching using the normalized correlation using the cut symbol as a template, and detects the cut symbol. After detecting one image, the coordinates of x and y having a high normalized correlation value are obtained and set as the coordinate point (x, y) of the cut symbol. The area surrounded by the coordinate points is obtained by the determination unit 62 as a barcode area.

  For example, when the coordinate points A, B, and C of the cut symbol obtained as shown in FIG. 5 are A (50, 10), B (10, 10), and C (10.50), the coordinates of the barcode area Is the range from the 10th pixel to the 50th pixel in the X direction and the range from the 10th pixel to the 50th pixel in the Y direction. For example, the barcode area is set to “1” and the non-barcode area is set to “0”, and the determination result corresponding to each pixel is stored in the barcode area storage unit 63. Here, it is assumed that reading is performed at a fixed position without tilting the barcode, but there may be rotation or distortion. In this case, processing corresponding to rotation and distortion is required. When there are a plurality of barcode areas, the determination unit 62 needs to group each of the three cut out symbols. In this case, for example, distance information between cut symbols may be obtained and grouped based on the distance information.

  For example, when the cut symbols are A, B, and C as shown in FIG. 5, the cut symbols are square, so AB: BC: CA is 1: 1: √2. Therefore, when the distance AB is set to 1, the cut symbols whose BC is close to 1 and CA is close to √2 are obtained and grouped, and the range surrounded by 3 cut symbols in each group is barcoded. This can be dealt with by determining the area.

  Further, since the barcode is composed of bars and cells, the area has a high spatial frequency of the image signal. Therefore, you may determine the location where a spatial frequency is high as a barcode area | region. Further, assuming that the barcode is a gray region with black and little saturation, a region with a small difference in R, G, B image signals may be determined as the barcode region. However, it is also possible to exclude white at this time.

  If the shape of the bar area is known, the information can be used for detection. In addition, when there are a plurality of types of cut-out symbols, a plurality of templates can be provided and matching can be detected by repeating a plurality of times. It is also conceivable to first find a part that seems to be a bar code, detect the size of the cell, estimate the size of the cut symbol from the size of the cell, and narrow down the template.

  As described above, if the format is determined and the barcode area is known information in advance, the area can be determined based on the known information. For example, if the barcode area is known in advance from the 100th line to the 500th line, the image signal is not required and only the timing for reading the image is used to determine the line number of the current image signal. . Therefore, the barcode area can be determined using this information.

  For the area determined as the barcode area in this way, the output of the barcode area image processing unit 4 is selected and the surrounding pixels are not referred to, so that the resolution of the image or the difference in image information between the pixels can be reduced. First, image conversion is performed as described above. As a result, a gray image having a resolution comparable to that read by a monochrome sensor can be acquired. Therefore, the barcode recognition rate is improved to the same level as a monochrome sensor, and good barcode recognition is possible.

  For an area determined to be a non-barcode area, the output of the non-barcode area image processing unit 5 is selected, and a color image or the like can obtain a good color image without impairing light and darkness or color reproducibility. That is, it is possible to achieve both good color image acquisition and high-definition barcode image acquisition.

  Here, if the lens has chromatic aberration as described above, a false color is generated at the edge of the gray region. If a color is generated when converting to a luminance value, that portion may become noise. When obtaining from the surrounding R signal, G signal, and B signal as in the prior art, the noise is reduced by averaging, but the R signal, G signal, and B signal are directly used as the luminance value Y as in this embodiment. Has a great impact. For this reason, it is conceivable that the image quality deterioration and the recognition rate of barcode recognition deteriorate.

  7 and 8 are diagrams illustrating the shape correction filter. This shape correction filter is a cell based on the assumption that, for example, chromatic aberration does not extend over several pixels, and conversely, the points or cells constituting a character or a two-dimensional barcode are over several pixels. It is a filter which correct | amends with respect to this shape.

  As shown in FIG. 7, the shape correction filter 43 is connected to the subsequent stage of the selector 41. As described above, the shape correction filter 43 is a filter that corrects the size or shape of the points or cells constituting the two-dimensional barcode.

  For example, as shown in FIG. 8, the shape correction filter 43 is assumed to be a quadrangle having a size of one cell and three pixels in length and width. In this case, the corners are likely to be blurred due to the noise, and the quadrilateral looks like an irregular shape such as a circle or a trapezoid. However, if it is known that the original shape is not less than 3 pixels in length and width, it can be corrected to approach 3 pixels in length and width.

  For this reason, the shape correction filter has a filter size larger than the number of vertical and horizontal pixels of one cell when a two-dimensional code cell is read as shown in FIG. 8, and the filter center corresponds to the vertical and horizontal centers of one cell. When there is a positive weight at a position corresponding to the inside of one cell, a negative weighting coefficient is provided at a position corresponding to the outside. Here, since one cell has three pixels in the vertical and horizontal directions, the filter size needs to be larger than three pixels in the vertical and horizontal directions. However, even if it is too large, the amount of hardware or software processing increases, so the filter size is 5 pixels vertically and horizontally.

  In the case of this example, the weight of 3 pixels in the center corresponding to the size of one cell is set as a positive weight. The square corners are not blurred because they are easily blurred and increase in luminance value (weight is set to 0), but weights may be given. In order to correct the shape of the corners of the four shapes, it is necessary to correct the blurring of the corners, and here, negative weights are applied to pixels adjacent to the corners in the vertical and horizontal directions. Since how much weight is emphasized and varies depending on the optical system, an example is shown here, but it is not necessary to stick to this. Similarly, the position of the weight need not be particular.

  By providing the shape correction filter 43 that removes noise and corrects the shape by using the size and shape of one cell in this way, it is possible to obtain good image quality even when noise occurs due to chromatic aberration of the lens. Thus, the barcode recognition rate can be improved.

FIG. 9 is a diagram illustrating an example in which the output from the barcode area determination unit 6 is invalidated by the changeover switch set by the user, and the output of the image processing unit is fixed to the output of the image processing unit for the area designated by the user. is there.
In the above example, the bar code area determination unit 6 switches the output of the bar code area image processing unit 4 and the output of the non-bar code area image processing unit 5, but the changeover switch set by the user as shown in FIG. Thus, the output from the barcode area determination unit 6 can be disabled by the changeover switch 9 and fixed to the output of the area image processing unit designated by the user.

  FIG. 11 is a diagram illustrating the configuration of the changeover switch 9. In the figure, the output signal b of the barcode area determination unit 6 and the setting signals from the user setting switch 8 set by the user are signal c and signal d. When the signal c is “1”, the output signal b of the barcode area determination unit 6 becomes the output signal e of the changeover switch 9 as it is. When the signal c is “0”, the output signal b of the barcode area determination unit 6 becomes invalid, and the set value of the signal d becomes the output signal e of the changeover switch 9 as it is.

  Here, the changeover switch is configured by a hard wafer, but may be performed by software processing. In that case, for example, the output of the barcode region image processing unit 4 and the non-barcode region image processing unit 5, the output of the barcode region determination unit 6, the setting signal of the user setting switch, etc. are displayed on the GUI screen. You may want to select.

  FIG. 10 is a diagram illustrating an example in which the output of the barcode area image processing unit 4 and the non-barcode area image processing unit 5 is switched according to the setting signal of the user setting switch 8. In the case of this example, it may not be switched within one image area. This switching process may be realized by hardware processing or software processing.

  Further, in the above description, the image processing unit 2 is arranged outside the camera unit 1, but it can also be housed in the camera unit 1. In this case, for example, the image processing unit 2 can be stored in the camera head 11.

  In the above description, an example using a single-plate color sensor as the color sensor has been described. However, a three-plate color in which R, G, and B pixel signals cannot be obtained at the original position by performing processing such as shifting the image. The present invention can also be applied to a sensor or a line sensor.

  As described above, according to the present embodiment, barcode reading using a color sensor enables barcode image acquisition with high resolution equivalent to reading with a monochrome sensor in the barcode area, and barcode recognition. It can contribute to rate improvement. In addition, a good color image can be acquired even in the color image region.

It is a figure explaining the image reading apparatus which concerns on embodiment of this invention. It is a figure explaining the detail of the image process part 4 for barcode area | regions. It is a figure explaining the process of the image process part 4 for barcode area | regions. It is a figure explaining the process of the image processing part 5 for non-barcode areas. It is a figure explaining the example of a barcode. It is a figure explaining the process of the barcode area | region determination part. It is a figure explaining a shape correction filter. It is a figure explaining a shape correction filter. It is a figure explaining switching by the changeover switch which a user sets. It is a figure explaining the switching by the setting signal of the user setting switch. FIG. 3 is a diagram illustrating a configuration of a changeover switch 9. It is a figure explaining a gray balance. It is a figure explaining the color arrangement | sequence of a sensor.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Camera part 2 Image processing part 3 Memory | storage part 4 Barcode area | region image processing part 5 Non-barcode area | region image processing part 6 Barcode area | region determination part 7 Selector 8 User setting switch 9 Changeover switch 11 Camera head 12 Support | pillar 13 Document stand 14 Document 15 Color image area 16 Barcode area 41 Selector
42 RGB color generation unit 43 Shape correction filter 61 Pattern matching unit 62 Determination unit
63 Barcode area storage unit

Claims (7)

A color camera having a color sensor and capable of shooting a document including an area on which a two-dimensional barcode is printed and outputting pixel signals of at least three colors;
An image processing unit for a barcode area that generates a one-pixel color signal using a signal of only a pixel to which any one of the pixel signals belongs;
A non-barcode area image processing unit that generates a one-pixel color signal using a signal of a pixel to which any one of the pixel signals belongs, and a signal of pixels around the pixel;
An image reading apparatus comprising a selector that switches and outputs the output of an image processing unit for a barcode area and an image processing unit for a non-barcode area. The image reading apparatus according to claim 1,
A bar code area determination unit that determines whether the area processed by the bar code area image processing unit is a bar code printed area, and switches the selector based on the determination output of the determination unit. An image reading apparatus. The image reading apparatus according to claim 1,
The bar code area image processing unit includes a shape correction filter for correcting the shape of a two-dimensional bar code read in a subsequent stage. The image reading apparatus according to claim 3.
When the two-dimensional barcode cell is read, the shape correction filter has a filter size larger than the number of vertical and horizontal pixels of one cell, and is placed inside one cell when the filter center corresponds to the vertical and horizontal center of one cell. An image reading apparatus characterized in that there is a positive weight at a corresponding position and a negative weight coefficient is provided at a position corresponding to the outside. A single-plate color camera that includes a three-color sensor and can capture a document including an area on which a two-dimensional bar code is printed and output any one of R, G, and B pixel signals;
An image processing unit for a barcode area that generates a one-pixel color signal using a signal of only a pixel to which any one of the R, G, and B pixel signals belongs;
An image processing unit for a non-barcode area that generates a one-pixel color signal using a signal of a pixel to which each of the R, G, and B pixel signals belongs, and a signal of a pixel around the pixel;
A storage unit storing an area in which a two-dimensional barcode is printed in advance;
An image reading apparatus comprising: a selector that switches and outputs the output of the barcode area image processing section and the non-barcode area image processing section based on a storage signal of the storage section. The image reading apparatus according to claim 1.
An image reading apparatus comprising a user setting switch for setting a selector switching position. The image reading apparatus according to claim 2.
A user setting switch for setting a switching position of the selector is provided, and a switch that uses either the user setting switch setting output or the determination output of the barcode area determination unit as a switching signal of the selector is provided. Image reading device.

Images