JP2008282232A - System for reading optical reader - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that it is difficult to prevent any waste or dust from completely entering even by phisically providing a brush or the like to prevent waste from entering from an insertion port of a medium in a conveyance path in an optical reading device, and to solve the problem that any waste or dust may enter between a write media in the device which receives two or more media at once so that the waste can be separated and removed as waste so as not to be recognized as a write mark when the waste is included in image data. <P>SOLUTION: This optical reading device such as an OCR and OMR is configured to detect waste or dust entering from the insertion port of a medium in a conveyance path, and to separate and eliminate it from a write mark. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a reading method of an optical reading device that reads marks on an entry medium such as a data sheet.

  In a conventional optical reader, dust / dirt is determined by a method in which a thin entry mark (entry mark having a low density) is not read as a stain, or a entry mark having a small size is not read as dust.

  In the optical reader, since dust, dirt, and the like mixed from the entry port of the entry medium of the optical reader may be read as an entry mark, the mixed dust, dirt, etc. are eliminated.

  This dust, dirt, etc. shape features, color features, out-of-focus state features, means to extract the presence or absence of lifting due to the three-dimensional shape and not read as a mark, and by extracting the features of the personal mark, Reading reliability is improved by providing means for eliminating noise and other noise information.

  According to the present invention, in the optical reader, dust, dust, and the like that have entered the conveyance path from the reading medium insertion port into the conveyance path are not read as a mark on the reading medium, and reading reliability can be prevented. Improve sexiness.

  Embodiments of the present invention will be described below with reference to FIGS. FIG. 1 shows a side view of the basic structure of the mechanism of the reading apparatus and functional blocks.

  Physically, it is a structure in which an entry medium is transported along a transport path and read by a scanner.

  FIG. 2 shows a schematic processing flow of the reading apparatus according to the present invention.

  In FIG. 2, processes 20 to 22 are processes necessary for processing normal image data.

  The process No. 23 is the basic process of the present invention.

  Although the feature of the first aspect is extracted, for example, the method is an embodiment for recognizing the dust shape by obtaining the vector distribution of the mark contour.

  That is, as shown in FIG. 3, lines 30 to 32 have few vector types, and if the distribution is biased, it can be determined that the figure has been artificially entered, but there are many vector types and are discrete. If there is, determine that it is garbage and eliminate it. In this case, before removing the mark as an entry mark, display the mark of the part on the display screen of the terminal together with the format display of the entry medium, and make a judgment to the operator by the inquiry "Do you want to delete this mark?" Operation is possible.

  With regard to claim 7, what is erased in the inquiry “Would you like to delete this mark?” Is the shape information extracted, stored in the terminal, and then noise information such as dust, debris and dirt. The purpose is to increase the accuracy of dust discrimination. In this case, in the optical reading device, a structural device for physically mounting something like a brush is provided so that dust does not enter the conveyance path from the medium insertion port. Therefore, since it is rare that dust enters from the insertion port, there is a small amount of dust accumulated and stored in one terminal, and accuracy may not be improved.

  Therefore, in claim 8, the numerical value indicating the characteristics of garbage is transmitted from the terminal to the host computer, and the host computer receives the information and manages it as integrated information shared by all terminals. When there is an inquiry transmission request from the network, it has a function of transmitting the accumulation information of dust characteristics to a terminal connected to the network, thereby increasing the accumulation information of dust characteristics and improving the accuracy of determining the garbage. This makes it possible to collect the industry-specific garbage characteristics in which the system is used and increase the accuracy. For example, in a terminal system used in a sewing-related store, the characteristics of lint waste from clothes are collected, and the accuracy is improved.

  According to a second aspect of the present invention, color characteristics such as dust, dust, and dirt are determined from image data obtained by a color scanner. As shown in FIG. 4, the hardware configuration is such that the timing circuit 41 inputs a synchronization signal generated in synchronization with the conveyance speed to the color CCD, and outputs the R, G, B signals output from the color CCD, respectively. A / D conversion is performed, and the digital values are recorded as image data in R, G, and B memories, respectively. Thereafter, the processes 21 and 22 in FIG. 2 are performed. In the process of No. 23, as shown in FIG. 5, the R, G, and B component values included in one mark are normalized to 255 values within one mark as shown in the color arrangement column. This value is determined as dust at the threshold value shown in the threshold value column. In this example, the writing instrument is limited to black, and the colored image data is determined to be dust, but by determining the threshold value from the assumed color of the writing instrument and the assumed color of the dust, the writing instrument type is determined. Can be increased.

  In this example, the R, G, and B component values in one mark are normalized, but the accuracy can be further improved by setting a color unevenness threshold value based on the color distribution information.

  According to the third aspect of the present invention, dust and dirt temporarily attached to the scanner glass surface become an out-of-focus image when the depth of field is small with respect to the focal point 67 of the scanner as shown in FIG. As shown in the figure, an area with a low density value is generated around the image. In the example of FIG. 7, the area having a density value of 5 or less spreads in a donut shape. It is determined whether the area below a certain concentration value is donut-shaped.

  A certain density value threshold is determined by a ratio from the density value at the center of the image.

  Based on the ratio of the donut-shaped area in the entire area, it is determined that the dust is temporarily attached to the scanner glass surface.

  According to claim 4, in order to process the dust on the surface of the entry medium as a stereoscopic image, images are taken by switching the time difference between two scanners or a liquid crystal half mirror, and the amount of floating from the entry medium is recognized as a stereoscopic image. To do. As shown in FIG. 1, the hardware configuration includes two left and right scanners.

  Alternatively, in terms of cost reduction, the hardware configuration can be detected in the same way as means for switching the liquid crystal half mirror at a sufficiently high speed with respect to the conveyance speed and obtaining image data of two kinds of angles. The means for obtaining image data of two types of angles with the liquid crystal half mirror is configured as shown in FIG. 8, and the liquid crystal half mirror switching timing circuit 84 switches the liquid crystal half mirror 82 to obtain two types of image data with different angles. obtain. By taking an image with a time difference with a two-dimensional scanner as shown in FIG. 9, the same paper surface of the entry medium can be obtained as image data at a different position after being conveyed, and is handled as two image data.

  FIG. 10 shows an example of detection of the amount of floating or the presence or absence of movement of dust. The position of a mark is based on the mark frame printed on the entry medium for two pieces of image data on the surface of the entry medium. This is an example of detecting whether or not there is a difference.

  In the image processing, a method of superimposing two pieces of image data on the basis of a mark frame and digitizing as a matching amount may be used. In this case, the amount of rotation of the dust is included in the numerical value, and it can be detected as dust even if the dust rotates.

  Regarding claim 5, on the premise that one person fills in one entry medium, all of the entry forms of each person are classified as sticks or rounds. Check if it is common to the marks. Anything that deviates from this common shape is likely to be dust. If the position and number of marks to be applied are unmatched according to operational rules, those that are out of this common shape are excluded. The color feature that one person has entered is based on the premise that it is rare when the writing instrument is changed in order to enter one entry medium. The color characteristics of marks entered in one entry medium are also classified, and those that have color characteristics that are clearly different from other entry marks are highly likely to be dust. If the position and number of marks to be applied are unmatched according to operational rules, those that are out of this color feature are excluded. By determining the color feature together with the shape feature, the accuracy of determination as dust can be improved.

  Specific shape feature extraction is as in the embodiment of claim 1. Further, specific extraction of color features is as in the embodiment of claim 2.

It is the figure which showed the structure and functional block of hardware. It is the figure which showed the processing flow. It is the table | surface which showed the shape feature example and the process of determination. It is the figure which showed the hardware constitutions in the detection with a color CCD. It is the table | surface which showed the color feature example and the threshold value of determination. It is the table | surface which showed the color feature example and the threshold value of determination. It is the graph which showed the density value tendency at the time of focus shift. It is the figure which showed the optical system which acquires two image data with one CCD. It is the figure which showed the means to obtain two image data with a two-dimensional CCD. It is the figure which showed the detection method of the presence or absence of dust floating and the presence or absence of dust movement.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Left scanner, 2 ... Right scanner, 3 ... Reading position, 4 ... Optical axis, 5 ... Fill medium, 6 ... Conveyance guide, 7 ... Conveyance roller, 8 ... A / D conversion part, 9 ... Image storage memory, 10 ... CPU section, 40 ... color CCD, 41 ... timing generation circuit, 42 ... A / D converter × 3, 43 ... RB memory, 60 ... scanner, 61-point CCD, 62 ... lens, 63 ... glass plate, 64 DESCRIPTION OF SYMBOLS ... Conveyance path, 65 ... Focus position, 66 ... Adhering dust, 67 ... Focus, 80 ... CCD, 81 ... Lens, 82 ... Liquid crystal half mirror, 83 ... Mirror, 84 ... Liquid crystal half mirror switching timing circuit, 85 ... Fill medium.

Claims (8)

  An optical reader having a function of discriminating noise information such as dust, dust, and dirt by image recognition processing and discriminating it from noise information when the shape is unique to noise information (dust, dust, dirt, etc.).   An optical reader having a color scanner for determining color characteristics of noise information such as dust, dust, and dirt, and having a function of discriminating noise information from color information of the color scanner.   An optical reading apparatus having a function of discriminating dust / dirt from the dust / dirt attached to the scanner glass surface by analyzing a gray gradation state around the dirty image by recognition processing.   The dust on the surface of the entry medium is imaged by switching the time difference between the two scanners or the liquid crystal half mirror, copying the time difference twice with a two-dimensional scanner, and the amount of lift from the entry medium from the difference between the two images. Alternatively, an optical reading device that can recognize the presence or absence of dust movement and has a function of identifying as dust adhering when there is floating or dust movement.   An optical reader having a function of extracting a personal entry shape / color feature from an entry mark on an entry medium, discriminating information deviating from the entry entry shape / color feature as noise information, and removing it from the entry mark.   An optical reading device having a function of extracting a personal writing density feature from a writing mark of a writing medium, discriminating information deviating from the writing density feature as noise information and removing it from the writing mark.   Noise information such as dust, debris, and dirt is displayed in the entry field, and the reading operator has a means to check whether it is a mark. The entry mark selected by the operator is noise information such as dust, dirt, and dirt. The optical information is characterized in that it is determined that the noise information is extracted by the image recognition process, the shape features are extracted and stored, and then used when it is determined that the noise information is dust, trash, dirt, etc. Type reader.   The noise information such as dust, debris, and dirt is accumulated and stored, and the feature information of the shape is extracted by the image recognition process. This optical reader, which is a terminal, sends the dust feature information to the host computer on the network. A function for transmitting and accumulating to the host computer, and an optical reading device that can receive the feature information from the host computer and make the feature information available for discrimination as noise information.

Images