JP3710068B2 - Health insurance card automatic recognition method and apparatus, and geriatric medical insurance card automatic recognition method and apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
TECHNICAL FIELD The present invention relates to a health insurance card automatic recognition method and apparatus, and a geriatric medical insurance card automatic recognition method and apparatus. For example, a health insurance card suitable for use in confirming or collating a health insurance card or a geriatric medical insurance card. The present invention relates to an automatic recognition method and apparatus, and an elderly medical insurance card automatic recognition method and apparatus.
[0002]
[Prior art]
Conventionally, hospitals have been required to check a health insurance card (hereinafter referred to as an insurance card) once a month under the guidance of the Ministry of Health and Welfare. Health insurance associations are also obliged to promptly collect insurance cards when members lose their qualifications. However, when the hospital treats the patient as insurance simply by checking the examination card and requests the health insurance association for the insurance contribution, there may be cases where the patient has already lost his / her insurance qualification.
[0003]
At this time, the health insurance association is not obligated to pay, and uncollected money is generated on the hospital side because the hospital neglected the obligation to confirm the insurance card. That amount is 0.7% to 1.0% of medical expenses. Therefore, the total medical expenses of each insurance association is 25 trillion yen in 1993, so the uncollected money is 175 to 250 billion yen in total.
[0004]
At present, the hospital has to increase the number of clerks significantly from the current level in order to ensure the confirmation of all patients' insurance cards.
[0005]
The number of outpatients reaches as many as 1500 people per day in a hospital with about 1000 beds, and it is concentrated within the short reception hours determined per day, so it is necessary to check which patient's insurance card. It is necessary to increase the number of check personnel to determine whether there is any.
[0006]
In addition, among the members of the medical insurance system, elderly people over 70 years old and bedridden elderly people over 65 years old are issued with a medical insurance card for the elderly from municipalities, etc., and medical treatment benefits from medical institutions Can receive.
[0007]
The target for issuing a geriatric health insurance card is stipulated by the geriatric health insurance system as follows. In other words, you are a health insurance member and your income is lower than the prescribed amount set by each local government. However, this will be reviewed annually. Next, be the target age (but less than 70 years old) determined by each local government. A geriatric health insurance card is issued for the elderly who meet this condition. In addition, this geriatric health insurance card can be used only within the prefecture of the address.
[0008]
In the case of a geriatric health insurance card, it will continue even though the geriatric health insurance card you currently have is invalid due to changes in income, unannounced screening, or moving or moving out. May be used. However, as in the case of health insurance cards, manual confirmation work takes time, so either a significant number of people are assigned to this work or no confirmation work is performed at all. Currently.
[0009]
The flowchart of FIG. 86 shows the work procedure of the confirmation work for the geriatric health insurance card. First, in step S911, the insured submits a geriatric medical insurance card to a medical institution. Next, in step S912, it is determined whether the monthly confirmation of the geriatric medical insurance card has been completed. If it is determined that the monthly health insurance card has not been checked yet, the process proceeds to step S913, and the work for checking the health insurance card is performed manually. Alternatively, depending on the medical institution, the confirmation operation is not performed at all, and the process proceeds to step 914 as it is.
[0010]
On the other hand, when it is determined that the monthly health insurance card has been confirmed, no processing is performed and the process proceeds to step S914.
[0011]
In step S914, a medical cost claim amount is calculated by a predetermined method. Next, proceeding to step S915, the medical institution charges a medical fee by submitting a predetermined document to a predetermined payment fund. Next, in step S916, the payment fund examines the document submitted by the medical institution, and determines whether or not the document is defective in step S917.
[0012]
If it is determined by the payment fund that the document is incomplete, the process proceeds to step S918, and the payment fund sends a return notification to the medical institution. On the other hand, if it is determined that there are no deficiencies in the documents, the process proceeds to step S919, and the payment fund pays medical expenses to the medical institution.
[0013]
[Problems to be solved by the invention]
Due to the situation as described above, there was a problem that the hospital side could not collect the full amount of medical expenses.
[0014]
In addition, in order to avoid congested reception, there are cases where a hospital that issues a magnetic examination card has an automatic re-entry reception machine, which can determine the order of the examination, Since there was no confirmation function, there was a problem that was not useful for labor saving in insurance medical care reception work.
[0015]
The present invention has been made in view of such a situation, and it is possible to easily determine the validity of health insurance cards and geriatric medical insurance cards, to realize labor saving in insurance medical reception work, In order to collect as much as possible.
[0016]
[Means for Solving the Problems]
The health insurance card automatic recognition method according to claim 1 reads an image of a health insurance card,By binarizing the read image with a predetermined threshold value, the image is divided into a first image and a second image. The predetermined position of the image is used as a search start point, and the vertical direction or the horizontal direction is set with respect to the health insurance card. The search direction is set, the direction of the ruled line to be searched is the direction perpendicular to the search direction with respect to the health insurance card, and the second image of the image from the search start point to the search direction with a predetermined search width in the direction perpendicular to the search direction If the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the minimum ruled line detection minimum length that extends in a direction perpendicular to the search direction and that is equal to or greater than the search width Defines a rectangular area in which a predetermined small area with the long side of the minute line segment is arranged in a matrix, and from the short area column on one short side to the other short side line Sequentially determine whether the second image is included in the column, and the second image is included If it is determined, counts up the predetermined counter, ratio borders detected minimum length of the counter value, if it is more than a predetermined ratio, it is assumed that borders health insurance card is detected, the detectedBased on the coordinates of the ruled line of the health insurance card, the position where the predetermined item of the health insurance card is described is detected, and based on the position where the item is described, the area where the character is described is detected and detected. Based on the area where the health insurance card letter is written, the letter written in the item is recognized, and based on the recognition result, the medical treatment of the outpatient is accepted and the health insurance card is confirmed. It is characterized by.
[0017]
In addition, the recognition of the characters written in the item detects the inclination of the health insurance card from the inclination of the left side of the insurance card frame of the health insurance card, and if the health insurance card is inclined upward, the insurer A rectangular frame of a predetermined size with the lower left point of the rectangular frame numbered as the apex and the horizontal and vertical line segments passing through the apex as two sides as the character area, the health insurance card goes up to the left Is a rectangular frame of a predetermined size with the upper left point of the rectangular frame with the insurer number as the apex and the horizontal and vertical line segments passing through the apex as two sides. This can be performed by reducing the character area by a predetermined length vertically and horizontally, and recognizing characters in the reduced character area.
[0018]
Further, the position of the seal stamp stamped on the health insurance card is recognized based on the color component constituting the image, and the upper and lower sides of the health insurance card are determined based on the recognized position of the seal stamp. Can do.
[0019]
Further, the item can be at least one of a health insurance card symbol, number, insurance association name, insurer number, effective date, date of birth, and date of acquisition.
[0020]
In addition, the detection of the area where the health insurance card characters are described is made by binarizing the image with a predetermined threshold value to divide the image into a first image and a second image, and a predetermined small area on the image. Sets character recognition areas arranged in a matrix, and determines whether or not the second image is included in the columns in order from the left column to the right of the small region constituting the character recognition region. It is determined whether or not the sequence not included is more than the predetermined character boundary determination number, and if it is determined that the sequence not including the second image is more than the character boundary determination number, The first column where two images are detected is the left end of the character, the last column where the second image is detected is the right end of the character, and the position of the second image detected at the bottom of the character recognition area is the lower end of the character, The position of the second image detected at the top of the character recognition area is the upper end of the character, and from the left end to the right Is the character width, the distance from the lower end to the upper end is the character height, the character width is equal to or greater than the minimum character width defined as the minimum character width, and the character height is the character length. If the character height is equal to or higher than the predetermined minimum character height defined as the minimum value, it can be performed by assuming that a character has been detected.
[0021]
Further, it is possible to perform thinning processing on the characters in the character area and recognize the characters subjected to thinning processing using the sonde method.
[0022]
The health insurance card automatic recognition apparatus according to claim 7 is an image reading means (for example, color scanner 2 in FIG. 17) for reading an image of a health insurance card, and an image read by the image reading means at a predetermined threshold value. By binarization, binarization means (for example, the processing unit 4 in FIG. 17) that divides the first image and the second image, and search start point setting means (for example, FIG. 17) using a predetermined position of the image as a search start point. 23, the search direction is the vertical direction or the left-right direction with respect to the health insurance card, the direction of the ruled line to be searched is the direction perpendicular to the search direction with respect to the health insurance card, and is perpendicular to the search direction. Second image search means (for example, step S42 in FIG. 23) for searching for a second image of the image in the search direction from the search start point with a predetermined search width in the direction, and the second image and the first image included in the search width The ratio of If the ratio is greater than or equal to the ratio, predetermined small areas extending in a direction perpendicular to the search direction and having a long side corresponding to the minimum length of a predetermined ruled line detection minimum length greater than the search width are arranged in a matrix. A determining unit that defines a rectangular region and determines whether or not the second image is included in the column in order from the column of the small region on one short side of the region to the column on the other short side. Step S48), and when it is determined that the second image is included, the counting means for counting up a predetermined counter (for example, step S49 in FIG. 23), and the ratio of the counter value to the minimum ruled line detection length is a predetermined value. If the ratio is greater than or equal to the ratio, the coordinate detection means (for example, step S51 in FIG. 23) that assumes that the ruled line of the health insurance card has been detected, and the health insurance card ruled line coordinates detected by the coordinate detection means, Insurance card Based on position detection means (for example, step S6 in FIG. 18) for detecting a position where a fixed item is described, and a position where the item detected by the position detection means is described, an area where characters are described Recognition for recognizing a predetermined item described in the health insurance card based on the character area detecting means to be detected (for example, step S8 in FIG. 18) and the area in which the character detected by the character area detecting means is described Based on the means (for example, the processing unit 4 in FIG. 17) and the recognition result by the recognition unit, the confirmation unit (for example, the processing unit 4 in FIG. 17) that accepts outpatient insurance medical treatment and confirms the health insurance card. It is characterized by providing.
[0023]
The recognition means includes an inclination detection means (for example, step S3 in FIG. 18) for detecting the inclination of the health insurance card from the inclination of the left side of the health insurance card frame, and the health insurance card is inclined upward to the right. A rectangular frame of a predetermined size with the lower left point of the rectangular frame in which the insurer number is written as a vertex, and a horizontal and vertical line segment passing through the vertex as two sides, and a character area, When the health insurance card is slanting upward to the left, a predetermined size with the top left corner of the rectangular frame containing the insurer number as the apex and the horizontal and vertical line segments passing through the apex as two sides A character area setting means (for example, the processing unit 4 in FIG. 17) having a rectangular frame as a character area, and a reduction means (for example, the processing unit 4 in FIG. 17) for reducing the character area vertically and horizontally by a predetermined length. , Character recognition to recognize characters in the character area reduced by the reduction means (Processor in example 17 4) means and may further provided.
[0024]
Further, based on the color components constituting the image, based on the seal position recognition means (for example, step S117 in FIG. 70) for recognizing the position of the seal stamped on the health insurance card, Up / down determination means (for example, step S118 and step S119 in FIG. 70) for determining the up / down of the health insurance card may be further provided.
[0025]
Further, the item can be at least one of a health insurance card symbol, number, insurance association name, insurer number, effective date, date of birth, and date of acquisition.
[0026]
Further, the character area detection means binarizes the image into a first image and a second image by binarizing the image with a predetermined threshold value (for example, the processing unit 4 in FIG. 17), and on the image. The setting means (for example, step S71 in FIG. 34) for setting the character recognition area in which the predetermined small areas are arranged in a matrix, and the columns in the order from the left column to the right of the small areas constituting the character recognition area. First determination means for determining whether or not two images are included (for example, step S75 in FIG. 34), and whether or not a sequence not including the second image has continued for a predetermined number of character boundary determinations. When it is determined that the second determination means for determining (for example, step S79 in FIG. 34) and the sequence not including the second image are more than the number of character boundary determinations, the first sequence in which the second image is detected Is the last column where the second image is detected Character area setting with the right edge of the character, the position of the second image detected at the bottom of the character recognition area as the bottom edge of the character, and the position of the second image detected at the top of the character recognition area as the top edge of the character Means (for example, step S80 in FIG. 34), the distance from the left end to the right end is the character width, the distance from the lower end to the upper end is the character height, and the predetermined character width minimum defined as the minimum character width. If the character height is equal to or greater than the value and the character height is equal to or greater than the predetermined minimum character height defined as the minimum value of the character length, character detection means that determines that a character has been detected (eg, step S84 in FIG. 34). Can be provided.
[0027]
Further, a thinning unit (for example, step S103 in FIG. 49) for performing a thinning process on characters in the character area is further provided, and the character recognizing unit applies the character subjected to the thinning process by the thinning unit. It can be recognized using the sonde method.
[0030]
Claim13The method for automatically recognizing a geriatric health insurance card described in 1 reads an image of a geriatric health insurance card, andBy binarizing with a threshold value, the image is divided into a first image and a second image, the search position is the predetermined position of the image, and the search direction is the vertical or horizontal direction for the geriatric health insurance card. The direction of the target ruled line is a direction perpendicular to the search direction with respect to the geriatric health insurance card, and the second image of the image is searched in the search direction from the search start point with a predetermined search width in the direction perpendicular to the search direction. When the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the length corresponding to the minimum ruled line detection minimum length that extends in the direction perpendicular to the search direction and is equal to or greater than the search width Define a rectangular area in which predetermined small areas with the long side of the line segment are arranged in a matrix, and in order from the column of the small area on one short side of the area to the column on the other short side Is determined to include the second image, and if it is determined that the second image is included Counts up the predetermined counter, ratio borders detected minimum length of the counter value, if it is more than a predetermined ratio, it is assumed that borders Medicare card is detected, the detectedBased on the number of ruled lines existing in a predetermined area of the image, it is identified whether the image is an image of a geriatric medical insurance card, and a position where a predetermined item of the geriatric medical insurance card is described from the read image Is detected, the character written in the item is recognized, insurance medical treatment is accepted based on the recognition result, and a geriatric medical insurance card is confirmed.
[0031]
The item may be at least one of a municipality number, a beneficiary number, and an expiration date of the geriatric medical insurance card.
[0032]
In addition, the recognition of the characters described in the item is based on detecting the inclination of the geriatric health insurance card from the inclination of the left side of the insurance card frame of the geriatric health insurance card, and the geriatric health insurance card is tilted upward to the right A rectangular frame of a predetermined size with the lower left point of each rectangular frame containing the municipality number and beneficiary number as a vertex and horizontal and vertical line segments passing through the vertex as two sides is defined as a character area. When the medical insurance card for the elderly is tilted to the left, the top left corner of each rectangular frame containing the municipality number and the beneficiary number is the apex, and the horizontal and vertical line segments passing through the apex are two sides. A rectangular frame of a predetermined size is used as a character area, and the character area is reduced vertically and horizontally by a predetermined length, and the character in the reduced character area is recognized. .
[0033]
The geriatric health insurance card automatic recognition device according to claim 16 includes image reading means (for example, color image input unit 12 in FIG. 72) for reading an image of the geriatric health insurance card,
By binarizing the image read by the image reading unit with a predetermined threshold value, the binarizing unit (for example, the confirmation processing unit 13 in FIG. 72) that divides the image into a first image and a second image, The search start point setting means (for example, step S181 in FIG. 79) with the position of the search start point as the search start point and the search direction as the search direction in the vertical direction or the horizontal direction with respect to the geriatric health insurance card, and the direction of the ruled line as the search target A second image search means that searches the second image of the image in the search direction from the search start point with a predetermined search width in the direction perpendicular to the search direction relative to the insurance card (for example, FIG. 79). If the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio in step S183), the minimum ruled line detection that extends in a direction perpendicular to the search direction and is equal to or greater than the search width A long line segment A rectangular region in which predetermined small regions having long sides are arranged in a matrix is defined, and a second image is arranged in a row in order from a column of small regions on one short side of the region to a column on the other short side. Determining means (for example, step S183 in FIG. 79) and counting means for counting up a predetermined counter when it is determined that the second image is included (for example, step S183 in FIG. 79) When the ratio of the counter value to the minimum ruled line detection length is equal to or greater than a predetermined ratio, the detection means (for example, step S184 in FIG. 79) that considers that the ruled line of the geriatric medical insurance card has been detected, and the detection means detect Identification means (for example, the confirmation processing unit 13 in FIG. 72) for identifying whether or not the image is an image of a geriatric medical insurance card based on the number of ruled lines existing in a predetermined area of the image, Based on the recognition result (for example, the confirmation processing unit 13 in FIG. 72) for recognizing a predetermined item described in the geriatric health insurance card from the image read by the recognition means, and the recognition result by the recognition means, the insurance treatment of the outpatient And confirming means (for example, the confirmation processing unit 13 in FIG. 72) for confirming the geriatric health insurance card.
[0034]
The item may be at least one of a municipality number, a beneficiary number, and an expiration date of the geriatric medical insurance card.
[0035]
The identifying means includes an inclination detecting means for detecting the inclination of the geriatric medical insurance card from the inclination of the left side of the insurance card frame of the geriatric medical insurance card (for example, step S146 in FIG. 74), and the geriatric medical insurance card rising to the right. A rectangle with a predetermined size, with the lower left point of each rectangular frame containing the municipality number and beneficiary number as the apex, and horizontal and vertical line segments passing through the apex as two sides If the geriatric health insurance card is slanting to the left, the upper left point of each rectangular frame with the municipality number and the beneficiary number is the apex, and the horizontal and vertical passes through the apex. Character area setting means (for example, the confirmation processing unit 13 in FIG. 72) having a rectangular frame of a predetermined size with two line segments as sides, and the character area reduced vertically and horizontally by a predetermined length Reduction means (for example, the confirmation process of FIG. 72) And 13), it is possible to recognize the character recognition means the reduced character area (e.g., confirmation processing unit 13 of FIG. 72) and further provided as a by reduction means.
[0038]
The health insurance card automatic recognition method according to claim 1, and the claim7In the health insurance card automatic recognition device described in, the image of the health insurance cardIs read, and based on the first image and the second image obtained by binarizing the read image of the health insurance card, the ruled line of the health insurance card is detected, and the ruled line of the detected health insurance card is Based on this position, the position where the predetermined item of the health insurance card is written is detected, and based on the position where the item is written, the area where the character is written is detected. A predetermined item is recognized and the health insurance card is confirmed.Therefore, the health insurance card confirmation operation can be automated.Moreover, the position where the number and the insurer number are described can be detected.
[0039]
In the health insurance card automatic recognition method according to claim 2 and the health insurance card automatic recognition device according to claim 8, when the health insurance card is inclined, characters included in each frame of the insurer number are recognized. A rectangle of a predetermined size consisting of horizontal or vertical sides to be used as a character area is reduced by a predetermined length vertically and horizontally, and characters in the reduced character area are recognized. . Therefore, the accuracy of character recognition can be increased.
[0040]
In the health insurance card automatic recognition method according to claim 3 and the health insurance card automatic recognition device according to claim 9, the position of the seal stamped on the health insurance card is determined based on the color component constituting the image. Based on this position, the top and bottom of the health insurance card are determined. Therefore, it is possible to quickly make a vertical determination.
[0041]
Claim5Health insurance card automatic recognition method and claims11In the health insurance card automatic recognition apparatus described in 1), the position of the character described therein is detected based on the first image and the second image included in the character recognition area. Therefore, it is possible to specify a region where characters to be recognized exist.
[0043]
Claim13The method for automatically recognizing a geriatric health insurance card according to claim 1, and claims16In the geriatric health insurance card automatic recognition device described in 1Is read, and based on the first image and the second image obtained by binarizing the read image, the ruled line of the geriatric health insurance card is detected, and the number of ruled lines existing in a predetermined region of the image is detected. Based on this, it is identified whether or not the image is an image of a geriatric medical insurance card, a predetermined item of the geriatric medical insurance card is recognized from the image, and the geriatric medical insurance card is confirmed.Therefore, it is possible to automate the confirmation work for the geriatric health insurance card.Moreover, the position where the municipality number and the recipient number are described can be detected.
[0044]
In the geriatric health insurance card automatic recognition method according to claim 15 and the geriatric health insurance card automatic recognition device according to claim 18, when the geriatric health insurance card is tilted, each frame of a municipality number and a recipient number A rectangle of a predetermined size consisting of horizontal or vertical sides for recognizing the characters contained in the character area is defined as a character area. This character area is reduced by a predetermined length vertically and horizontally, and within the reduced character area Are recognized. Therefore, the accuracy of character recognition can be increased.
[0046]
DETAILED DESCRIPTION OF THE INVENTION
The confirmation items of the health insurance card currently managed by the hospital are the health insurance association number (insurer number), insured name, number, symbol, date of birth, insurance qualification date, and validity It is a deadline. The health insurance union number indicates the type of insurance. These items are entered in a medical record, and a large hospital creates a database using a computer.
[0047]
FIG. 1 shows the ratio of the population of insured persons who have insurance cards that can be automatically verified among various types of insurance to the population covered by medical guarantee. As shown in FIG. 1, the ratio of the population having an insurance card that can be automatically matched to the population subject to medical guarantee is 99.32%. However, the criteria for collation determination are based on the insurance card specifications as of August 29, 1994.
[0048]
FIG. 2 is a diagram showing types of health insurance insured cards (commonly called insurance cards). As shown in FIG. 2, the health insurance card is a health insurance insured card based on the Health Insurance Act, a one-year special insurance insured person certificate, and a special insurance insured person day 5 Annual health insurance special treatment invoices, national health insurance insured card based on the National Health Insurance Act, and national health insurance retirement insured ID card, seafarer insurance insured ID card based on the Seafarer Insurance Act, And a sailor insurance dependent's ID, further a national public servant's mutual aid association card based on the Mutual Aid Association Act, a national public servant's mutual aid association remote area dependent certificate, a local public servant's mutual aid association card, a local public servant's mutual aid association remote area, etc. There is a dependent certificate, a private school faculty and mutual aid membership card, and a private school faculty and mutual aid association remote area dependent card. In this way, there are currently 14 types of insurance cards, and the respective formats are determined by the Ministry of Health and Welfare.
[0049]
FIGS. 3 to 16 show the formats of the above 14 types of insurer cards in order.
[0050]
When the health insurance card is submitted to the hospital at the first visit, the hospital prepares a medical record and inputs it to the computer. The health insurance association number of the health insurance card, the insured person's name, number, symbol, date of birth, insurance Confirmation date such as date of qualification and expiry date is input to computer, and database of medical records and insurance card data is created. At the time of returning, the insurance card is required to be verified once a month under the guidance of the Ministry of Health and Welfare.
[0051]
If the patient is a revisited patient, it is checked against insurance card data registered in advance. First, it is determined whether or not the copied insurance card. The image data of the insurance card is extracted with a color scanner or color CCD camera, and it is input to a computer to determine whether it is genuine. In the case of an insurance card that has been color-copied, color elements are detected as RGB data in a white background portion, whereas in a genuine case, the white background portion is not detected because it is printed on colored paper. From this, the copied counterfeit can be identified.
[0052]
In addition, under the guidance of the Ministry of Health and Welfare, the color of the insurance card print is specified and is changed every 2 or 3 years, so whether it is genuine or not by judging whether it is the specified color or not Judgment can be made.
[0053]
Next, it is determined whether or not the patient's insurance card. There are four items common to all types of health insurance cards: health insurance association number (unique number given to the insurance association), insured number, insured date of birth, and issue date. Among these, the health insurance association number and the insured number are identified, and it is determined whether or not the patient's insurance card.
[0054]
Furthermore, the position where the health insurance association number and the insured number are written is identified from the image data of the insurance card and recognized. As mentioned above, when analyzing the format of the insurance card, the insured number is listed in the second or third frame from the top, and the health insurance association number is listed at the bottom. Recognize the position, and then recognize the number on it. Details of this recognition method will be described later.
[0055]
FIG. 17 is a block diagram showing a configuration of an embodiment of the health insurance card automatic recognition apparatus of the present invention. The color scanner 2 reads an image of the insurance card 1 set therein and supplies corresponding image data to the processing unit 4.
[0056]
From the image data supplied from the color scanner 2, the processing unit 4 includes a health insurance association number, an insured person number, a symbol, a date of birth, an insurance qualification date, an expiration date, etc. At least the health insurance association number and the insured number are recognized, and it is determined whether the insurance card is the patient's own or valid.
[0057]
The keyboard 5 is configured to be able to give a predetermined command to the processing unit 4. The storage unit 6 is configured to store information corresponding to a patient's medical record, for example. The output unit 7 outputs predetermined data supplied therefrom under the control of the processing unit 4.
[0058]
Therefore, for example, the processing unit 4 searches for a predetermined patient chart stored in the storage unit 6 and displays corresponding characters, figures, or images on the display unit 3 including a CRT or a liquid crystal display, or the like. The data is output from an output unit 7 such as a printer.
[0059]
Next, the operation will be described with reference to the flowchart of FIG. First, in step S1, the insurance card 1 is read by the color scanner 2. At this time, the insurance card 1 is read with a resolution of a predetermined accuracy or higher, and a bitmap file (for example, the file extension is bmp) is created. The red (R), green (G), and blue (B) components of the read image data are binarized by a predetermined threshold value.
[0060]
As a result, pixels with a pixel level smaller than the predetermined threshold are set as white dots (pixels corresponding to the first image), and pixels larger than or equal to the predetermined threshold are set as black dots (pixels corresponding to the second image). ). As a result, image data composed of white dots and black dots is obtained for each of the R, G, and B components.
[0061]
Here, the threshold for binarization can be set to a predetermined value in advance, but can also be set dynamically corresponding to the color of the insurance card 1 mount. is there. For example, the insurance card 1 is scanned by the color scanner 2, and a threshold value for binarization can be set by a predetermined calculation based on each color component of the read pixel data. Thereby, it is possible to set an appropriate threshold value corresponding to the color of the card of the insurance card 1 and binarize based on the threshold value.
[0062]
In step S2, it is determined whether the insurance card 1 set in the color scanner 2 has been copied. As described above, whether or not the insurance card 1 has been copied is determined based on the image data of the insurance card 1 read by the color scanner 2. For example, a color-copyed health insurance card has a color element detected as RGB data in a white background portion, so that it can be identified whether or not it is genuine.
[0063]
Alternatively, an image corresponding to a seal of a predetermined color (for example, vermilion) stamped on the health insurance card 1 in the color image captured by the color scanner 2, as in the case of collation processing for a medical insurance card for the elderly described later. It is also possible to identify whether or not it is genuine by determining whether or not it exists.
[0064]
If it is determined that the insurance card 1 has been copied, an error is generated, and the message “Health insurance card cannot be recognized. Please come to the reception desk” is displayed on the display unit 3 and the process is terminated. On the other hand, if it is determined that the health insurance card 1 set in the color scanner 2 is not copied, the process proceeds to step S3.
[0065]
In step S3, the inclination of insurance card 1 is checked. For this purpose, first, a search for insurance policy frames is performed. FIG. 19 shows an example of image data obtained when the insurance card 1 is read by the color scanner 2. As described above, the image data of the insurance card 1 is read as image data of a predetermined size indicated by the image data boundary including the insurance card 1. At this time, if it is desired to shorten the computer processing time, the intermediate portion of the image data is unnecessary and can be avoided.
[0066]
Next, a method for checking the inclination of the insurance card 1 will be described with reference to the flowchart of FIG. First, in step S21, a predetermined point A is selected as shown in FIG. For this point A, for example, the Y coordinate can be set to a quarter position from the top of the image data of the insurance card 1 shown in FIG. 19, and the X coordinate can be set to the left end position of the image data.
[0067]
In step S22, when the dots are searched in order from the point A to the right, black dots detected immediately after a predetermined number of consecutive white dots appear are displayed in the left frame of the insurance card 1. Judge that there is. In this way, as shown in FIG. 21, the intersection 1P is detected.
[0068]
Similarly, in step S23, the Y coordinate is set to the position of the third quarter from the top of the image data of the insurance card 1 shown in FIG. 19, and the point B having the X coordinate as the left end position of the image data (FIG. 21). Is selected, the process proceeds to step S24, and the intersection 2P is detected by searching for dots in the right direction from the point B.
[0069]
Next, in step S25, a straight line L obtained by connecting the intersection 1P and the intersection 2P.₁Is the left frame of insurance card 1, and the inclination is the inclination of insurance card 1. In step S26, it is determined whether or not the inclination is within a predetermined degree. If it is determined that the inclination is not within the predetermined degree, an error is determined, and the process proceeds to step S27. After the message “It is tilted. Please try again” is displayed on the display unit 3, the process returns to step S 1, and the subsequent processing is repeated. On the other hand, when it is determined that the inclination of the insurance card 1 is within the predetermined degree, the process returns to step S4, and the coordinates of the four corners are retrieved from the outer frame of the insurance card 1.
[0070]
Next, a method for searching for the coordinates of the four corners of the insurance card will be described with reference to the flowchart of FIG. First, in step S31, on the insurance card 1 shown in FIG. 21, the Y coordinate is set to a quarter position from the top of the image data of the insurance card 1 shown in FIG. 19, and the X coordinate is the right end of the image data. When the point C is selected, the process proceeds to step S32, and when dots are searched sequentially in the left direction from there, the black dots detected immediately after a predetermined number of consecutive white dots appear are detected by this insurance. It is determined that it is the right frame of Certificate 1. Then, the intersection 3P (FIG. 21) is detected.
[0071]
Next, in step S33, in the same manner, a point D having the X coordinate as a position one third from the left of the image data of the insurance card 1 shown in FIG. 19 and the Y coordinate as the position of the upper end of the image data is set. Select and proceed to step S34, and search for dots in order from there downward, and determine that the black dot detected immediately after a predetermined number of consecutive white dots is the upper frame of this insurance card. The intersection 4P (FIG. 21) is detected. Further, in step S35, the point E is selected with the X coordinate set to a position one third from the left of the image data of the insurance card 1 shown in FIG. 19, and the Y coordinate set to the lower end position of the image data, and step S36. When the dots are searched in order upward from there, the black dot detected immediately after the number of consecutive white dots appears more than the specified number of dots is determined to be the frame under this insurance card, and the intersection 5P (FIG. 21) is detected.
[0072]
Next, in step S37, perpendicular lines are drawn from the intersections 1P to 5P, and the vertices 1C to 4C of the insurance card 1 are obtained. For example, a straight line L obtained by connecting the intersection 1P and the intersection 2P₁Next, a perpendicular line is drawn from the intersection point 4P, and a straight line L connecting the perpendicular leg and the intersection point 4P.₂Get. Next, a straight line L obtained by connecting the intersection 1P and the intersection 2P from the intersection 5P.₁A straight line connecting the vertical line and the intersection 5P_ThreeGet. Furthermore, from the intersection 3P, a straight line L₂Or L_ThreeA straight line connecting the vertical line and the intersection 3P_FourGet. The straight line L thus obtained₁Thru L_FourAre the vertices 1C to 4C of the insurance card 1.
[0073]
Next, the process returns to step S5, and the coordinates of the horizontal ruled line in the insurance card 1 are searched. Here, a horizontal ruled line detection method will be described with reference to the flowchart of FIG. First, in step S41, as shown in FIG. 24, a search start point p1 is set at a predetermined position of the insurance card 1. Next, the process proceeds to step S42, and as shown in FIG. 25, black dots are searched in order with the search width (w) as a predetermined dot in the downward direction (search direction (vector 1)) with respect to the insurance card 1. In step S43, it is determined whether or not one-third or more black dots are detected with respect to the search width. If it is determined that a black dot of 1/3 or more dots has not been detected with respect to the search width, the process returns to step S42, and the subsequent processing is repeated.
[0074]
On the other hand, if it is determined that one-third or more black dots have been detected, the process proceeds to step S44, and as shown in FIG. 26, the detected black dots (detection points) are included, and from there, k1 dots to the left A point having a distance of a certain number (a preset number of dots) is set as the upper left vertex, and a rectangle having a vertical length of, for example, 2 dots is set. Next, in step S45, an initial value 0 is set in the black counter.
[0075]
In step S46, the upper left vertex of the rectangle is set as the search position. Next, the process proceeds to step S47, where black dots are searched downward (vector 1 direction) from the search position. Next, proceeding to step S48, it is determined whether or not a black dot has been detected. If it is determined that a black dot has been detected, the process proceeds to step S49, and after adding 1 to the black counter, the process proceeds to step S50. On the other hand, if it is determined that no black dot has been detected, the process proceeds to step S50 without doing anything.
[0076]
Next, in step S50, the value of the black counter is a predetermined value (value N_FourIt is determined whether or not it is larger. If it is determined that the value of the black counter is greater than the predetermined value, the process proceeds to step S51 and it is considered that a ruled line has been detected. That is, if tens of percent of the dots with the preset number of dots, which is the length of the horizontal side of the rectangle set in step S44, are black dots, it is considered that a ruled line has been detected, and the process returns. On the other hand, when it is determined that the value of the black counter is smaller than or equal to the predetermined value, the process proceeds to step S52, and the right next to the search position is set as a new search position.
[0077]
Next, proceeding to step S53, it is determined whether or not the search position extends beyond the range of the image data, or whether or not the search has been completed to the right end of the rectangle. If the search position does not protrude from the image data range and the search has not been completed to the right end of the rectangle, the process returns to step S47, and the subsequent processing is repeated. On the other hand, if the search position is outside the range of the image data, or if the search is completed up to the right end of the rectangle, the process proceeds to step S54, where it is assumed that no ruled line has been detected, and the process returns.
[0078]
In step S5, by repeating the process shown in FIG. 23, all the coordinates of the horizontal ruled line of the insurance card 1 existing downward from the search start point p1 are detected and stored in the storage unit 6. Similarly, as shown in FIG. 27, a ruled line is searched downward (in the direction of vector 1) from a predetermined search start point p2 in the upper left part of insurance card 1. However, in this case, the ruled lines to be searched are only horizontal ruled lines extending in the right direction (vector 3 direction).
[0079]
In this way, the horizontal ruled lines K4 to K8 are detected from the search start point p1, and the horizontal ruled lines K1 to K3 are searched from the search start point p2.
[0080]
In step S6, the number recognition position is extracted. Taking out the number position of the insurance card is currently performed in the following procedure because two types of insurance card patterns are possible as shown in FIG.
[0081]
FIG. 30 is a flowchart showing this procedure. First, in step S61, as shown in FIG. 31, three horizontal ruled lines are detected downward from the detection start point p1 on the right side of the insurance card 1. Next, the process proceeds to step S62, and a vertical ruled line in the upward direction (vector 4 direction) is searched in the left direction (vector 2 direction) from a predetermined point p3 between the two as upper and lower side surfaces.
[0082]
The method of searching for vertical ruled lines is basically the same as the method of searching for horizontal ruled lines described with reference to the flowchart of FIG. 23, but when searching for vertical ruled lines, the following method is used. It is also possible. That is, as shown in FIG. 32, a search is performed in the right direction with a predetermined search width dot. For example, the total number of black dots included in the region for two columns is the number of dots existing there (in this case 10 (= When the ratio to 5 × 2) dots) is several tens of percent or more, it can be regarded as a ruled line. In the case of FIG. 32, the proportion of black dots present in the first and second columns from the left is several tens of percent or less, and at this point in time, it is not regarded as a ruled line. Next, since the ratio of black dots present in the second and third columns from the left is close to 100%, it is regarded as a ruled line at this point.
[0083]
In the case of FIG. 31, the search is started in the left direction from the search start point p3, for example, with a predetermined number of search width dots, and the total number of black dots for two horizontal rows is the number of dots 10 for two horizontal rows (= 5 × 2). If it is more than several tens of%, it is regarded as a vertical ruled line.
[0084]
In this way, the left side surface shown in FIG. 31 is searched, and in step S63, the insurance card is specified from the positional relationship, and the process proceeds to step S64, where the coordinates of the number frame as shown in FIG. That is, the number recognition position is extracted and the process returns.
[0085]
Next, in step S7, the number size and pitch described in the number frame are extracted. As shown in FIG. 33, the total number of vertical black dots in the number frame at a predetermined X coordinate position in the number frame is obtained in the horizontal direction in the number frame, and a histogram is created. And based on this, the size and pitch of the character (number) described in the number frame are determined.
[0086]
Next, the process proceeds to step S8, and a character position search is first performed. Processing during this period will be described with reference to the flowchart shown in FIG. First, in step S71, a search target area (area1) is set as shown in FIG. In step S72, the left side of area1 is set as the search position.
[0087]
Next, in step S73, an initial value 0 is set as a white count value indicating the number of consecutive white dots, and the process proceeds to step S74. In step S74, the number of black dots in the column including the search position is counted, and in step S75, it is determined whether or not the number of black dots is zero.
[0088]
If it is determined that the number of black dots is not 0, the process proceeds to step S76, and the right next to the search position is set as a new search position, and the process proceeds to step S77. In step S77, it is determined whether or not the search position is outside the range of the image data, or whether or not the search has been completed to the right end of area1. If it is determined that the search position does not protrude from the image data range and the search has not been completed to the right end of area1, the process returns to step S73, and the subsequent processing is repeated.
[0089]
On the other hand, if it is determined that the search position is outside the range of the image data or the search is completed to the right end of area 1, the process proceeds to step S85, where it is determined that no character is detected, and the process returns.
[0090]
On the other hand, if it is determined in step S75 that the number of black dots is 0, the process proceeds to step S78, and the value 1 is added to the white count. In step S79, it is determined whether the white count value is greater than or equal to a preset number of character boundary white dots, in this case, a predetermined dot.
[0091]
If it is determined that the white count value is smaller than the number of white dots on the character boundary, the process returns to step S74, and the subsequent processing is repeated. On the other hand, if it is determined that the value of the white count is greater than or equal to the number of white dots on the character boundary, the process proceeds to step S80, and the coordinates of the left end, right end, lower end, and upper end of the character are determined.
[0092]
Next, in step S81, the distance from the left end to the right end of the character determined in step S80 is the character width, and then the distance between the upper end and the lower end of the character determined in step S80 is the character height. Next, the process proceeds to step S82, where it is determined whether or not the character width is greater than or equal to a preset character width minimum value, in this case, a predetermined dot. If it is determined that the character width is smaller than the minimum character width value, the process proceeds to step S85, the character is not detected, and the process returns. On the other hand, if it is determined that the character width is greater than or equal to the minimum character width value, the process proceeds to step S83.
[0093]
In step S83, it is determined whether or not the character height is a preset minimum character height, in this case greater than or equal to a predetermined dot. If it is determined that the character height is smaller than the minimum character height, the process proceeds to step S85, where it is regarded that no character is detected, and the process returns. On the other hand, if it is determined that the character height is greater than or equal to the minimum character height value, the process proceeds to step S84, where it is assumed that the character has been detected, and the process returns.
[0094]
In this manner, after the character position is searched, the characters described in the searched character position, here, the numbers 0 to 9 are recognized using the sonde method described later. Here, if the number is not recognized normally, an error is assumed and the process is terminated. On the other hand, if the number is recognized normally, the process proceeds to step S9.
[0095]
In step S9, for example, the frame coordinates of the insurer number as shown in FIG. 36 are extracted. Details of the processing performed in step S9 will be described with reference to the flowchart shown in FIG. First, in step S91, as shown in FIG. 38, the search start point p4 located at the lower left part of the insurance card 1 is selected.
[0096]
Next, in step S92, a horizontal ruled line extending from the search start point p4 upward (vector 4 direction) to the right (vector 3 direction) is searched. Since the search method is the same as that described with reference to FIG. 23, the description thereof is omitted here, but in this case, the search width is set wide for identification with characters. In this way, the first horizontal ruled line and the second horizontal ruled line are searched as shown in FIG. These two horizontal ruled lines are used as the upper and lower side surfaces of the insurer number area.
[0097]
At this time, as shown in FIG. 39, if there is another horizontal ruled line between the search start point p4 and the insurer number area, the horizontal ruled line searched first is the lower side of the insurer number area. There is a possibility of being recognized. In order to solve this, the horizontal ruled line that matches the horizontal ruled line previously searched in step S5 is excluded.
[0098]
In addition, as shown in FIG. 40, when a horizontal ruled line search is performed, if there is a dense area of black dots, for example, if a character string is described, the character string is recognized as a horizontal ruled line. It may end up. In order to solve this, a case where the interval between the detected horizontal ruled lines seems to be appropriate, for example, about 4 to 9 millimeters (mm) is excluded. In the case of FIG. 40, the interval between the first and second horizontal ruled lines is 2 mm and is excluded. The interval between the second and third horizontal ruled lines is 2.5 mm, which is also excluded. The distance between the third and fourth horizontal ruled lines is 7 mm, and these two horizontal ruled lines are to be processed, and it is determined that these two horizontal ruled lines constitute the upper and lower side surfaces of the insurer number area. The
[0099]
In this manner, after the upper and lower side surfaces of the insurer number area are searched, the process proceeds to step S93, and then the left and right side surfaces are searched. First, as shown in FIG. 41, a search position is set at a predetermined position between the lower side surface and the upper side surface searched in step S92, and then the left side (vector 2 direction) and the right side (vector 3 direction) are moved upward. The vertical ruled line in the direction (vector 4 direction) is searched. Since the search method is basically the same as that described with reference to FIG. 32, the description thereof is omitted here.
[0100]
In the case of this insurance card 1, as a result of the search, a total of 13 vertical ruled lines are searched. The intervals are 12, 10, 24, 30, 32, 30, 31, 33, 33, 32, 31, and 139 from the left. Next, values obtained by rounding these intervals into predetermined dot units are calculated, and a table as shown in FIG. 42 is created. Here, INDEX is an identification number of the interval between the vertical ruled lines, and a value obtained by rounding the actual interval in units of a predetermined dot is described in the interval column.
[0101]
From this table, the mode value of the interval is obtained. In this example, since it is 6, the value obtained by dividing the total of the actual number of real dots between ruled lines belonging to the mode value zone having the interval value of 6 by the number of intervals belonging to the mode value zone (character width ( Character frame width). In this example, 31 (= (30 + 32 + 30 + 31 + 33 + 33 + 32 + 31) / 8) dots.
[0102]
Next, the process proceeds to step S95, and the vertical ruled line on the left side is specified. In other words, the horizontal ruled lines are searched in the vertical direction of the vertical ruled lines in order from the left one of the previously searched vertical ruled lines. If the retrieved horizontal ruled line matches the horizontal ruled line corresponding to the upper and lower sides of the insurer number area searched in step S92, the vertical ruled line is determined as the left side.
[0103]
Next, proceeding to step S96, from the left side obtained in step S95, the left side is obtained by multiplying the number of digits of the insurer number, in this case, the number of dots obtained by multiplying 8 digits by the character width determined in step S94. The ruled line at the position moved in the right direction from is determined as the right side. In this way, the frame coordinates of the insurer number are extracted.
[0104]
Next, the process proceeds to step S10, and as shown in FIG. 43, the number size described as the insurer number is extracted. This is done by creating a histogram of the numbers described in one square unit of the insurer number frame, and determining the size in the same manner as described with reference to FIG.
[0105]
Next, it progresses to step S11 and number recognition is performed. Recognition of the numbers starts from the left square of the insurer number area. At this time, as shown in FIG. 44, when the insurance card 1 is inclined upward to the right, the lower left point of the actual frame is the center. Create horizontal and vertical frames (character recognition areas). As shown in FIG. 45, when the insurance card 1 is inclined upward to the left, a horizontal / vertical frame (character recognition area) is created around the upper left point of the actual frame.
[0106]
Next, as shown in FIG. 46, the character recognition area created in FIG. 44 or 45 is reduced by a predetermined length, for example, about 1 mm vertically and horizontally. Then, with respect to the reduced character recognition area, as described later, the characters described there, in this case, numbers are recognized. Each time one character is recognized, the character recognition area to be recognized is shifted to the right by one square, and the same processing as described above is repeated. In this way, all the insurer numbers described in the insurer number area are recognized, and the process is terminated. If an error occurs during the recognition process at this time, a predetermined error process is performed and the process is terminated.
[0107]
FIG. 47 summarizes the above-described thresholds when searching for insurance policy frames. When the search target is the left side of the insurance policy frame, the X coordinate of the search start position is the left end of the image data, and the Y coordinate is the position of 1/4 and 3/4 from the top of the image data length. The number of consecutive white bits is the predetermined number of dots. When the search target is the right side of the insurance policy frame, the X coordinate of the search start position is the right end of the image data, the Y coordinate is a position of a quarter from the top of the image data length, and the number of consecutive white bits is The predetermined number of dots.
[0108]
In addition, when the search target is the upper side of the insurance policy frame, the X coordinate of the search start position is, for example, a third position from the left of the image data width, the Y coordinate is the upper end of the image data, and the white bit The continuous number is a predetermined number of dots. Further, when the search target is the lower side of the insurance policy frame, for example, the X coordinate of the search start position is a position one third from the left of the image data width, the Y coordinate is the lower end of the image data, and the white bit The continuous number is a predetermined number of dots.
[0109]
FIG. 48 summarizes the thresholds when searching for the ruled lines of the above-mentioned insurance card. When the search target is a horizontal ruled line in the insurance card, the ratio when proceeding to the ruled line search when black dots are detected, that is, the ratio of the number of black dots existing in the search width (scanning line width) of a predetermined number of dots is The search target ruled line minimum length (number of dots) is a predetermined number of dots, and the black dot ratio (percentage) in the ruled line when determining whether or not the line is a ruled line is Dozens of percent. When the search target is an insurer number upper and lower frame, the ratio when proceeding to the ruled line search when black dots are detected is one third of the scanning line width, and the minimum ruled line length (number of dots) is the predetermined number of dots. The black dot ratio (percentage) in the ruled line when determining whether it is a ruled line is several tens of percent.
[0110]
In addition, when the search target is the left and right frame of the insurer number, the ratio when proceeding to the ruled line search when black dots are detected is one third of the scanning line width, and the minimum ruled line length (number of dots) is the insurance It is a predetermined number of dots corresponding to the interval (6 to 7 mm) of the upper and lower frames of the person number, and the black dot ratio (percentage) in the ruled line when determining whether it is a ruled line is several tens%. When the search target is the left side of the insurer number, the ratio when proceeding to ruled line search when black dots are detected is one third of the scanning line width, and the minimum length (number of dots) of the search target ruled line is insurance. The number of dots corresponding to the character number character width (about 5 mm), and the black dot ratio (percentage) in the ruled line when determining whether it is a ruled line is several tens of percent.
[0111]
Furthermore, when the search target is the left side of the insured number, the ratio when proceeding to ruled line search when black dots are detected is one third of the scanning line width, and the minimum length (number of dots) of the search target ruled line is The predetermined number of dots corresponding to the interval (8 to 10 mm) between the upper and lower sides of the insured person number frame, and the black dot ratio (percentage) in the ruled line when determining whether or not it is a ruled line is several tens of percent.
[0112]
Next, a character recognition method will be described with reference to the flowchart of FIG. First, in step S101, the location where the target number is described is determined from the image data of the insurance card 1 taken in by the color scanner 2 as described above, and a histogram of that portion is created. 50 to 58 show histograms corresponding to the numbers 0, 1, 2, 3, 4, 5, 6, 7, and 9 as an example. Next, the process proceeds to step S102, and it is determined which number is to some extent from such a histogram pattern. For example, some candidates are raised.
[0113]
Next, in step S103, the image data is thinned. FIG. 59 shows image data of number 6 before thinning. In this case, the character height (CH) from the upper end to the lower end of the character is 12 dots, and the character width (CW) from the left end to the right end of the character is 14 dots. FIG. 60 shows the image data after the number 6 shown in FIG. 59 has been thinned. 61 and 62 show changes in image data before and after the thinning process, using the numeral 0 as an example. 63 and 64 show changes in the image data before and after the thinning process, taking the numeral 4 as an example. 63A shows image data before the insurer number thinning process, and FIG. 63B shows image data before the insurer number thinning process. On the other hand, FIG. 64A shows the image data after the insurer number thinning process, and FIG. 64B shows the image data after the insurer number thinning process.
[0114]
FIG. 65 and FIG. 66 show changes in image data before and after the thinning process, taking the number 6 as an example. Here, FIG. 65A shows image data before thinning the insurer number, and FIG. 65B shows image data before thinning the insured number. On the other hand, FIG. 66A shows the image data after the insurer number thinning process, and FIG. 66B shows the image data after the insured person number thinning process.
[0115]
In this manner, image data obtained by performing thinning processing on the image data of the insurer number and the insured number described on the insurance card 1 is obtained.
[0116]
Next, in step S104, the character recognition process by the sonde method is performed on the image data subjected to the thinning process obtained in step S103. FIG. 67 is a diagram illustrating the principle of the sonde method. In the sonde method, as shown in FIG. 67, a predetermined number of points (sonde base points) are set at predetermined positions in advance. In this example, two sonde base points are set. Then, as shown in FIG. 68, numbers are written on the basis of this sonde base point. At this time, for each number, the relative positional relationship between the stroke of each number and the sonde base point, that is, the rule of writing is determined in advance. FIG. 68 shows a standard way of writing numbers.
[0117]
Next, as shown in FIG. 67, white line sondes are set radially from the sonde base point. If the stroke of the number passes this sonde, it is set to 1, and the number is read with this code combination. This technique is applicable only to simple characters of the order of numbers, and in addition, as shown in FIG. 68, there is a restriction that a writing rule and a writing place are specified in advance. However, in spite of such a simple method, the character deformation as shown in FIG. 69 can be absorbed.
[0118]
In the case of the present embodiment, the numbers to be recognized are printed not in handwriting but in a predetermined size and a typeface typeface. Therefore, it is possible to recognize a number with high accuracy by setting the probe base point to an appropriate position and emitting the probe radially in an appropriate direction therefrom.
[0119]
In this way, the insurer number and the insured number of the insurance card 1 can be recognized. Thereby, it can be determined whether or not the presented health insurance card belongs to the patient.
[0120]
As described above, the health insurance card confirmation process, which requires confirmation once a month, is automatically performed simultaneously with the reception process at the time of re-entry, thereby preventing uncollected medical expenses in advance. Significant improvement in paperwork can be achieved.
[0121]
In the above embodiment, before recognizing characters or the like written on the insurance card, it is possible to make an up / down decision and perform a recognition process based on the result. FIG. 70 is a flowchart for explaining a processing example for determining the vertical direction of the insurance card.
[0122]
First, in step S111, a value 0 is set to a predetermined seal counter described later and is initialized.
[0123]
Next, in step S112, the color information of the insurance card 1 is taken in by the color scanner 2. Here, since only the insurance card up / down judgment is performed, the insurance card image captured by the color scanner 2 is divided into two upper and lower areas for convenience, and two types of seal counters corresponding to the divided upper and lower areas. A (corresponding to the upper area) and a seal counter B (corresponding to the lower area) are provided.
[0124]
Next, in step S113, predetermined color identification information (color data value) for discriminating a color from each of the red (R), green (G), and blue (B) color components constituting the captured color information. Is calculated. Here, the color data value is information obtained by, for example, adjusting each color component, or multiplying each color component by a predetermined coefficient and then adding or subtracting them.
[0125]
Next, in step S114, based on the color data value obtained in step S113, it is determined whether or not the color of the pixel of interest is a seal color, for example, vermilion. If it is determined that the color of the pixel currently focused on is vermilion, the process proceeds to step S115, and either the seal stamp counter A or the seal stamp counter B corresponding to the position of the pixel currently focused on is incremented by one. That is, when the pixel of interest is in the upper area, the value of the seal counter A is increased by 1. When the pixel of interest is in the lower area, the value of the seal counter B is increased by 1. Let Thereafter, the process proceeds to step S116.
[0126]
On the other hand, if it is determined that the color of the pixel of interest is not vermilion, no processing is performed and the process proceeds to step S116.
[0127]
In step S116, it is determined whether or not the entire image of insurance card 1 has been captured and the color thereof has been determined. If it is determined that the acquisition of all images and the determination of the color have not been completed yet, the process returns to step S112, and the processes of steps S112 to S116 are repeatedly executed. On the other hand, if it is determined that all images have been captured and the color has been determined, the process proceeds to step S117.
[0128]
In step S117, it is determined whether or not the value of the seal counter A corresponding to the upper area is smaller than the value of the seal counter B corresponding to the lower area. When it is determined that the value of the seal counter A is greater than or equal to the value of the seal counter B, the process proceeds to step S118, where the insurance card is inserted into the color scanner 2 with the top and bottom reversed, that is, the insurance card 1 In the state where the vertical direction of the image of FIG. 5 is opposite to the vertical direction of the color scanner 2, it is considered that the insurance card 1 is installed in the color scanner 2, and the processing is terminated.
[0129]
On the other hand, if it is determined that the value of the stamp counter A is smaller than the value of the stamp counter B, the process proceeds to step S119, and the insurance card 1 is in a correct state in which the vertical direction of the insurance card image and the vertical direction of the color scanner 2 match. Is installed in the color scanner 2, and the process is terminated.
[0130]
Since the data corresponding to the image of the insurance card 1 read by the color scanner 2 is stored in the storage unit 6 and can be read out by designating an address, the up / down determination of the insurance card 1 is performed as described above. If, for example, it is determined that the insurance card 1 is installed in the color scanner 2 in a state where the vertical direction of the insurance card 1 is opposite to the vertical direction of the color scanner 2, the memory is stored. Address when data is rearranged or read out in the storage unit 6 so that the data corresponding to the image of the health insurance card 1 stored in the unit 6 is rotated 180 degrees Can be read out in order from the upper part to the lower part of the image of the insurance card 1 by converting the data by a predetermined conversion function.
[0131]
Thus, since the data can be read from the upper part of the insurance card 1 regardless of the vertical direction of the insurance card 1 installed in the color scanner 2, the number or character recognition process described in the insurance card 1 can be recognized. Can be done correctly.
[0132]
FIG. 71 is a flowchart for explaining another example of a method for determining whether the insurance card 1 is up or down. First, in step S121, a predetermined feature extraction operation is performed on the binarized data of each pixel constituting the image of the insurance card 1 read by the color scanner 2.
[0133]
Next, in step S122, it is determined whether or not a feature extraction operation has been performed on all pixels. If it is determined in step S122 that the feature extraction calculation has not been performed on all the pixels, the process returns to step S121, and the processes of steps S121 and S122 are repeatedly executed. On the other hand, when it is determined that the feature extraction calculation has been performed on all the pixels, the process proceeds to step S123.
[0134]
In step S123, the insurance card up / down determination is performed based on the result of the feature extraction process executed in steps S121 and S122.
[0135]
FIG. 72 is a block diagram showing a configuration of one embodiment of the geriatric health insurance card automatic recognition apparatus of the present invention.
[0136]
The color image input unit 12 is configured by, for example, a color image scanner or a color camera, and takes in image data of the geriatric health insurance card 12A. The patient card input unit 11 is configured by a magnetic card reader or a barcode reader, and reads the patient number recorded on the patient card 11A by magnetism or a barcode. In addition, when patient data such as a municipal number, a recipient number, and an expiration date are recorded on the patient card 11A together with the patient number, the patient data is also read.
[0137]
Based on the data input from the patient card input unit 11 and the image data input from the color image input unit 12, the confirmation processing unit 13 performs a verification process for the geriatric health insurance card 12A and controls each unit. Has been made.
[0138]
The display / input unit 14 is configured by combining, for example, a CRT (Cathode Ray Tube) or LCD (Liquid Crystal Display), a touch panel, and the like, and predetermined image data supplied from the confirmation processing unit 13. In addition to displaying images corresponding to text data, etc., by operating the touch panel, a predetermined instruction can be given to the confirmation processing unit from a patient or a system administrator (for example, a hospital administration manager) It has been made possible.
[0139]
The voice guide unit 15 outputs a predetermined voice under the control of the confirmation processing unit 13 and performs an operation guide. The printing unit 17 prints the processing result in the confirmation processing unit 13 on a predetermined sheet and outputs it.
[0140]
Next, the operation will be described with reference to the flowchart of FIG. First, in step S131, the patient card input unit 11 reads the patient number recorded on the patient card 11A. The read patient number is supplied to the confirmation processing unit 13.
[0141]
Next, in step S132, the confirmation processing unit 13 determines whether or not patient data such as a municipality number, a recipient number, and an expiration date exists in the patient card 11A via the patient card input unit 11. . If it is determined that the patient data exists in the patient card 11A, the process proceeds to step S133, the patient data is read from the patient card 11A, and the process proceeds to step S135. On the other hand, if it is determined that the patient data does not exist in the patient card 11A, the process proceeds to step SS134, and the patient data corresponding to the patient number is retrieved from the patient database 16 using the patient number as a search key and read.
[0142]
However, if the patient database 16 is located outside the apparatus, the patient database 16 is connected to each other via a communication network and communicates based on a predetermined communication protocol. Searched and read.
[0143]
At this time, together with the patient data, data indicating whether or not the monthly medical insurance card 12A has been confirmed is read. Thereafter, the process proceeds to step S135.
[0144]
In step S135, collation processing of the geriatric medical insurance card 12A is performed. That is, the content described in the geriatric health insurance card 12A and the content recorded in the patient card 11A are collated. That is, the confirmation process for the geriatric health insurance card 12A is performed on patients who have not yet been confirmed for the geriatric health insurance card 12A.
[0145]
Next, details of this collation processing will be described with reference to FIGS. Since the process here is basically the same as the case of the health insurance card described above, a detailed description thereof will be omitted as appropriate.
[0146]
In step S141 of the flowchart shown in FIG. 74, the color image input unit 12 takes in the image data of the geriatric health insurance card 12A. The captured image data is binarized in step S142.
[0147]
A detailed processing example of image data capture and binarization processing in steps S141 and S142 will be described with reference to the flowchart of FIG. First, in step S161, a value 0 is set to a predetermined seal counter and is initialized. Next, in step S162, as shown in FIG. 76, the color image input unit 12 scans the inner peripheral portion of the mount of the geriatric health insurance card 12A as indicated by the arrow L, for example. Each color component (color information) of red (R), green (G), and blue (B) corresponding to the pixel is captured by the confirmation processing unit 13.
[0148]
Next, proceeding to step S163, the confirmation processing unit 13 performs a predetermined calculation process on each captured color component, and calculates a predetermined color data value. In step S164, it is determined whether or not the color data value calculated in step S163 is a value corresponding to the vermilion color that has been calculated in advance. If it is determined that the color data value is not a value corresponding to vermilion, the process proceeds to step S166. On the other hand, if it is determined that the color data value is a value corresponding to vermilion, the process proceeds to step S165, and the value of the seal counter A or B corresponding to the position of the pixel of interest is incremented by 1 and updated. .
[0149]
Next, in step S166, enhancement processing for a predetermined color is performed. For example, when the color of a ruled line or a character is green, a color data value corresponding to green is obtained in advance, and the color data value calculated in step S163 matches the value obtained in advance Alternatively, if the difference between the color data value calculated in step S163 and the value obtained in advance is within a predetermined range, the value of the pixel of interest is set to value 1, for example. In this way, green ruled lines, characters, etc. can be emphasized.
[0150]
Alternatively, when the green component of the pixel of interest is larger than a predetermined value, it is determined that this pixel is a pixel constituting a green ruled line or a character, and the value of this pixel is set to 1. Even in this way, green ruled lines, characters, and the like can be emphasized. A predetermined color other than green can be similarly emphasized.
[0151]
Next, proceeding to step S167, the predetermined color is invalidated. For example, when the vermillion color that is the color of the seal is invalidated and the image of the seal is removed, the color data value corresponding to the vermilion is obtained in advance, and the color data value calculated in step S163 is obtained in advance. If it matches the color data value corresponding to vermilion, or if the difference between the color data value calculated in step S163 and the color data value corresponding to vermilion obtained in advance is within a predetermined range, The value of the pixel of interest is set to 0, for example. In this way, the vermilion pixels constituting the seal image can be sequentially removed.
[0152]
Alternatively, when the size of the red component of the pixel of interest is greater than or equal to a predetermined level, it is determined that this pixel is a pixel constituting the seal image, and the value of this pixel is set to 0. And Even in this way, it is possible to sequentially remove the vermilion pixels constituting the seal image. The predetermined colors other than vermilion can be invalidated in the same manner.
[0153]
In step S168, a predetermined identification level for binarization is calculated based on the color data value calculated in step S163. The calculation of the identification level can be performed by pixels on the scan line L of the geriatric medical insurance card shown in FIG. 76, and thereafter, binarization processing is performed according to the identification level. That is, when the color data value is equal to or higher than the identification level, the value of the pixel of interest is set to, for example, value 1, and when the color data value is smaller than the identification level, for example, the value is set to 0. At this time, for the pixel emphasized in step S166 and the pixel invalidated in step S167, the value set there is directly used as a result of the binarization process.
[0154]
In this way, since the identification level can be calculated in the initial stage after the start of scanning, the identification level is calculated by one pass (scanning the entire geriatric health insurance card 12A), Binarization processing can be performed. The binarized result is stored in a video RAM (not shown) of the confirmation processing unit 13 or the like.
[0155]
Next, the process proceeds to step S169, where all parts of the geriatric health insurance card 12A are scanned, and it is determined whether or not the binarization processing has been performed on all the pixels constituting the read image. When it is determined that the binarization processing has not been performed for all the pixels, the process returns to step S162, and the processing of steps S162 to S169 is repeatedly executed. On the other hand, if it is determined that binarization processing has been performed for all pixels, the process proceeds to step S170.
[0156]
In step S170, it is determined whether or not the value of the seal stamp counter A or B is larger than a predetermined designated value. In step S170, when it is determined that the value of the seal counter A or B is smaller than or equal to a preset value (specified value), the process proceeds to step S171 and the geriatric medical insurance card 12A set in the color image input unit 12 is entered. Is determined to be a duplicate and returns.
[0157]
On the other hand, if it is determined that the value of the seal stamp A or B is greater than the specified value, the process proceeds to step S172, where it is determined that the geriatric medical insurance card 12A set in the color image input unit 12 is genuine.
[0158]
Next, in step S173, it is determined whether the count value of the seal counter A corresponding to the upper half of the geriatric health insurance card 12A is smaller than the count value of the seal counter B corresponding to the lower half of the geriatric health insurance card 12A. Is done. When it is determined that the count value of the seal counter A corresponding to the upper half of the geriatric health insurance card 12A is smaller than the count value of the seal counter B corresponding to the lower half of the geriatric health insurance card 12A, the process proceeds to step S174. It is determined that the medical insurance card 12A is set in the color image input unit 12 in a state where the vertical direction of the medical insurance card 12A matches the vertical direction of the color image input unit 12, and the process returns.
[0159]
On the other hand, if it is determined that the count value of the seal counter A corresponding to the upper half of the geriatric health insurance card 12A is greater than or equal to the count value of the seal counter B corresponding to the lower half of the geriatric health insurance card 12A, step Proceeding to S175, it is determined that the geriatric health insurance card 12A is set in the color image input unit 12 with the up and down directions reversed, and the process returns.
[0160]
As described above, when the binarization process of the geriatric medical insurance card 12A is completed, the process proceeds to step S143 in FIG.
[0161]
In step S143, it is determined from the determination result in step S171 or S172 of FIG. 75 whether the geriatric medical insurance card 12A set in the color image input unit 12 is genuine. If it is determined that the geriatric health insurance card is not genuine, the process proceeds to step S154. On the other hand, when it is determined that the geriatric health insurance card is genuine, the process proceeds to step S144, and the up / down determination is performed.
[0162]
In step S144, based on the determination result in step S174 or S175 of FIG. 75, whether the vertical direction of the color image input unit 12 matches the vertical direction of the geriatric medical insurance card 12A set in the color image input unit 12 or not. The up / down judgment is made.
[0163]
If it is determined that the vertical direction of the color image input unit 12 and the vertical direction of the geriatric medical insurance card 12A set in the color image input unit 12 do not match, that is, the opposite direction, the process proceeds to step S145. The image data corresponding to the geriatric medical insurance card 12A binarized in S142 is rotated by 180 degrees, and the process proceeds to step S146. On the other hand, if it is determined that the vertical direction of the color image input unit 12 matches the vertical direction of the geriatric medical insurance card 12A set in the color image input unit 12, no processing is performed and the process proceeds to step S146. .
[0164]
In step S146, the frame line of the elderly medical insurance card 12A is detected. The processing performed in step S146 is basically the same as the case of detecting the health insurance card border described above with reference to FIGS. 20 to 22, and will be briefly described here with reference to FIG. 77. .
[0165]
That is, as shown in FIG. 77, for example, a search for a vertical ruled line is started in the right direction from a predetermined point A on the left side of the geriatric health insurance card 12A, and a point where the vertical ruled line is first searched is set as an intersection 11P. . Similarly, the search for the vertical ruled line is started in the right direction from a predetermined point B located on the left side of the medical insurance card 12A for the elderly, and the point where the vertical ruled line is first searched is set as the intersection 12P.
[0166]
Next, the search for the vertical ruled line is started in the left direction from the point C located on the right side of the geriatric health insurance card 12A, and the point where the vertical ruled line is first searched is set as the intersection 13P. Next, the search for the horizontal ruled line starts from the point D located above the geriatric health insurance card 12A, and the point where the horizontal ruled line is first searched is set as the intersection 14P. Further, the search for the horizontal ruled line is started from the point E located below the geriatric health insurance card, and the point where the horizontal ruled line is first detected is set as the intersection 15P.
[0167]
A straight line L passing through the intersection points 11P and 12P₁₁Is calculated as the inclination of the geriatric health insurance card 12A. When this inclination exceeds a predetermined range, an error message can be displayed on the display / input unit 14. When this inclination is within a predetermined range, a predetermined correction process is performed corresponding to this inclination when recognizing numbers and characters.
[0168]
Next, straight line L from intersection 14P₁₁The vertical line that descends from the point 15C is the vertex 11C, and the straight line L from the intersection 15P₁₁Let the leg of the perpendicular dropped down to be apex 12C. Next, a straight line L connecting the vertex 11C and the intersection 14P₁₂The straight line L connecting the vertex 12C and the intersection 15P₁₃Ask for. Furthermore, a straight line L from the intersection 13P₁₂The vertical leg that is lowered with respect to the vertex 13C, and from the intersection 13P to the straight line L₁₃Let the leg of the perpendicular dropped to the vertex 14C. In this way, the coordinates of the vertex of the frame line of the geriatric health insurance card 12A can be obtained.
[0169]
Of course, the positions of points A to D with respect to the geriatric health insurance card 12A do not have to be positions as shown in FIG. 77, and can be changed to arbitrary positions where a frame line can be detected.
[0170]
Next, the process proceeds to step S147, and a horizontal ruled line is detected to identify the elderly medical insurance card 12A. That is, as shown in FIG. 78, the number of horizontal ruled lines existing in the upper right half area of the geriatric health insurance card 12A is detected, and whether or not it is the geriatric health insurance card 12A is determined from the number. ing.
[0171]
The detailed processing will be described with reference to the flowchart of FIG. First, in step S181, as shown in FIG. 78, a search start point p11 is set at a predetermined position on the upper right on the mount outside the frame of the geriatric health insurance card 12A. Next, in step S182, the horizontal ruled line counter is initialized with a value of zero.
[0172]
Next, proceeding to step S183, a horizontal ruled line is searched for one dot (pixel) in the downward direction (search direction (vector 1)) from the search start point p11. The horizontal ruled line search method is basically the same as that of the health insurance card described above with reference to the flowchart of FIG. 23, and thus detailed description thereof is omitted here. In step S184, it is determined whether the line is a horizontal ruled line. If it is determined that the line is a horizontal ruled line, the process proceeds to step S185, and the horizontal ruled line counter is incremented by one. Next, in step S186, the coordinates of the detected horizontal ruled line are temporarily stored in a memory (not shown) or the like of the confirmation processing unit 13.
[0173]
On the other hand, if it is determined in step S184 that the line is not a horizontal ruled line, no processing is performed and the process proceeds to step S187.
[0174]
In step S187, it is determined whether or not the horizontal ruled line search has been completed for the upper half of the geriatric health insurance card 12A. If it is determined that the search for the horizontal ruled line is not completed for the upper half of the geriatric health insurance card 12A, the process returns to step S183, and the processes of steps S183 to S187 are repeatedly executed. On the other hand, if it is determined that the search for the horizontal ruled line has been completed for the upper half of the geriatric health insurance card 12A, the process proceeds to step S188.
[0175]
In step S188, it is determined whether or not the value of the horizontal ruled line counter is 3 or less, for example. If it is determined that the value of the horizontal ruled line counter is 3 or less, the process advances to step S189 to determine that the one set in the color image input unit 12 is the geriatric medical insurance card 12A, and the process returns. On the other hand, if it is determined that the value of the horizontal ruled line counter is greater than 3, the process proceeds to step S190, where it is determined that the color image input unit 12 is not the geriatric health insurance card 12A, and the process returns.
[0176]
Next, the process proceeds to step S148 in FIG. 74, and whether or not the health insurance card 12A is set in the color image input unit 12 based on the determination result in step S189 or S190 in the flowchart shown in FIG. Is determined. If it is determined that the medical insurance card for the elderly is not 12A, the process proceeds to step S154. On the other hand, if it is determined that the medical card for the elderly is 12A, the process proceeds to step SS149, and the collation process of the municipality number is performed.
[0177]
FIG. 80 is a flowchart for explaining the procedure of the collation processing of the municipality number and the recipient number of the geriatric medical insurance card 12A. Since the municipality number verification process and the recipient number verification process are basically performed in the same manner, only the municipality number verification process will be described here. First, in step S201, a search is performed for an area in which a municipality number is described. That is, among the horizontal ruled lines searched in step S183 of the flowchart shown in FIG. 79, as shown in FIG. 81, the first horizontal ruled line is the upper side of the municipality number field, and the second horizontal ruled line is the lower side. And Next, a predetermined search start point p21 located between the upper side surface and the lower side surface is set, and from there, the vertical ruled line is searched in the left direction (vector 2 direction) and the right direction (vector 3 direction), whereby the municipality number The column area is searched.
[0178]
In step S202, the thinning process is performed on the area of the municipality number of the image data stored in the video RAM. Next, in step S203, a one-character area in which each municipality number is described is searched. Since the processing here is basically the same as that described above with reference to FIGS. 37 to 42, detailed description thereof is omitted here.
[0179]
In step S204, character image data in one character area is taken from the video RAM. At this time, in consideration of the inclination of the geriatric health insurance card 12A obtained previously, the position of each dot constituting the character image data is corrected to minimize the adverse effect on the recognition processing due to the inclination. be able to.
[0180]
Next, in step S205, the character corresponding to the read character image data is recognized, for example, by the above-described sonde method. The recognized character is converted into a character code. Next, in step S206, the first letter of the municipality number written on the recognized health insurance card 12A, the letter of the municipality number recorded on the patient card 11A, and the municipality number are not recorded on the patient card 11A. In this case, the patient number recorded on the patient card 11A is used as a search key, and collation with the character of the patient's municipality number read from the patient database 16 is performed by comparing character codes.
[0181]
As a result of the collation, if it is determined that both are the same, the count value of the collation character number counter is incremented by 1, and the process proceeds to step S209. On the other hand, if it is determined that the two are not the same as a result of the collation, no processing is performed and the process proceeds to step S209.
[0182]
In step S209, it is determined whether or not the matching process has been completed for all the digits of the municipality number. If it is determined that the matching process has not been completed for all the digits of the municipality number, the process returns to step S203. The processes in steps S203 to S209 are repeatedly executed. On the other hand, if it is determined that the collation process has been completed for all digits of the municipality number, the process proceeds to step S210.
[0183]
In step S210, after the count value of the collation character number counter counted in step S208 is divided by the number of characters (digits) of the municipality number, the collation rate is calculated, and the calculation result is stored in a predetermined memory. Return.
[0184]
Next, the process proceeds to step S150 in FIG. 74, and it is determined whether or not the collation rate calculated in step S210 in FIG. 80 is higher than a preset allowable collation rate. When it is determined that the collation rate calculated in step S210 is lower than or equal to the preset allowable collation rate, the process proceeds to step S154. On the other hand, when it is determined that the collation rate calculated in step S210 is higher than the preset allowable collation rate, the process proceeds to step S151.
[0185]
In step S151, a collation process of the recipient number of the geriatric medical insurance card 12A set in the color image input unit 12 is performed. In the case of a beneficiary number, the lower side of the municipality number column is the upper side of the beneficiary number column, and the horizontal ruled line existing below is the lower side of the beneficiary number column. Then, as shown in FIG. 81, a predetermined search start point p22 is set between the upper side and the lower side of the beneficiary number field, and from there, the left (vertical 2 direction) and the right (vertical 3 direction) are vertically By searching the ruled line, the area of the recipient number column is searched. The subsequent processing procedure is basically the same as that in the case of the municipality number collating process described above with reference to FIG. 80, and therefore detailed description thereof is omitted here.
[0186]
When the collation of the recipient number is completed, the process proceeds to step S152, and it is determined whether or not the collation result of the recipient number is higher than a predetermined collation rate set in advance. As a result, if it is determined that the collation rate of the recipient number is higher than a predetermined collation rate set in advance, the process proceeds to step S153, where it is determined that both the municipality number and the recipient number have been collated normally, and the process returns. On the other hand, when it is determined that the collation rate of the recipient number is lower than or equal to a predetermined collation rate set in advance, the process proceeds to step S154. In step S154, it is determined that collation is impossible, and the process returns.
[0187]
Next, the process proceeds to step S136 of the flowchart shown in FIG. 73, and the collation result performed in step S135 is registered. That is, when the patient data is recorded on the patient card, the content of the patient card 11A is updated based on the result of the collation, and when the patient data is registered in the patient database, the patient data is recorded based on the collation result. The contents of the database 16 are updated using the patient number as a search key.
[0188]
In step S137, the confirmation processing unit 13 converts the collation result into predetermined data for printing, and then supplies the data to the printing unit 17. The printing unit 17 prints data corresponding to the collation result supplied from the confirmation processing unit 13 on a predetermined sheet and outputs it as a confirmation table, and then ends the process.
[0189]
Thus, since the identification level for binarization is calculated corresponding to the color of the mount of the geriatric health insurance card 12A, even if the color of the mount of the geriatric health insurance card 12A is changed, it is appropriate. Binary processing can be performed. The color data value corresponding to the color of the mount of the geriatric health insurance card 12A can be assumed to be within a predetermined level range, and the color corresponding to the image data of a predetermined area of the geriatric health insurance card 12A. Based on the data value, an identification level for binarization can be obtained.
[0190]
Further, since the identification level is calculated based on the color data value corresponding to the scanned pixel in the inner peripheral portion of the mount, the identification level for binarization is determined at an early stage after the start of scanning. , Binarization processing based on it can be performed. As a result, it is possible to perform binarization processing of all image data with one pass (scanning the old medical insurance card 12A as a whole), and the processing speed can be improved.
[0191]
When the verification process for the geriatric medical insurance card is completed in this way, the confirmation processing unit displays selection items such as internal medicine and surgery on the display / input unit 14, and operates the touch panel on the patient to perform a predetermined selection. Display a message prompting you to select an item. At this time, when no operation is performed by the patient for a certain period, the voice guide unit 15 is controlled to output a voice for prompting the patient to select a predetermined selection item. Thereby, even a patient who is not accustomed to the operation can perform a predetermined operation according to the voice guide.
[0192]
Next, an example of a method for changing the allowable collation rate will be described with reference to the flowchart in FIG. This allowable collation rate is used when determining whether or not the collation result is correct in steps S150 and S152 of FIG. This allowable verification rate may need to be changed depending on the operation content.
[0193]
First, in step S221, an allowable collation rate to be changed is determined. Next, in step S222, the portion corresponding to the allowable collation rate in the source code of a predetermined collation program for determining the collation result (application program executed in the confirmation processing unit 13 in FIG. 72) is corrected. Next, proceeding to step S223, the source code is compiled, and after the debugging process is performed at step S224, the changed collation program is installed at step S225. Thereby, an allowable collation rate can be changed.
[0194]
Next, another example of the method for changing the allowable collation rate will be described with reference to the flowchart in FIG. In the case of this example, the collation program executed in the confirmation processing unit 13 shown in FIG. 72 reads the allowable collation rate from a predetermined file such as a storage unit (for example, hard disk) (not shown) of the confirmation processing unit 13, It is programmed to make a determination of the matching result based on it.
[0195]
Therefore, by changing the allowable verification rate written in this file with another predetermined application program (hereinafter referred to as a file rewriting program), the allowable verification rate can be set to an arbitrary value without changing the source code of the verification program. It becomes possible to set to.
[0196]
The operation of the file rewriting program will be described below. When the file rewriting program is activated, the confirmation processing unit 13 controlled by the file conversion program outputs a signal corresponding to a predetermined menu screen for changing the allowable collation rate to the display / input unit 14. Thereby, for example, a menu screen as shown in FIG. 84 is displayed. This menu screen displays items corresponding to a plurality of allowable collation rate values that can be set (in this case, items corresponding to 30% (%) to 100% (in increments of 10%)).
[0197]
In this state, a system administrator (for example, a hospital office manager or the like) starts a process for changing the allowable verification rate. First, in step S231, the value of the allowable collation rate to be changed is determined. Next, in step S232, the system administrator changes the allowable verification rate by selecting an item corresponding to the allowable verification rate to be changed from the predetermined items displayed on the display / input unit 14. Can do. That is, by pressing a part where an item corresponding to a predetermined allowable verification rate is displayed with a finger, the item can be selected and a corresponding allowable verification rate can be set.
[0198]
In addition, it is possible to clearly indicate which item has been selected by changing the display color of the selected item. In this case, the item displaying the character string “80%” is selected, and the background color thereof is changed.
[0199]
Next, proceeding to step S233, the confirmation processing unit 13 opens a predetermined file for writing data corresponding to the allowable verification rate (allowable verification data) stored in the internal memory. Next, in step S234, the file pointer is moved to a position on the file where the allowable verification data is to be written, and the process proceeds to step S235, where the allowable verification data corresponding to the allowable verification rate set in the display / input unit 14 is set to the file pointer. Write to a predetermined position on the file indicated by.
[0200]
Next, in step S236, the file is closed and the process is terminated.
[0201]
Thereafter, the collation program reads the allowable collation data written in this file and sets the corresponding permissible collation rate. Based on this, the processing in steps S150 and S152 of FIG. 74, that is, the collation result is evaluated. Will be done.
[0202]
In addition, as shown in FIG. 85, the menu screen displays icons corresponding to the numbers 0 to 9 on the display / input unit 14 and points to a portion where the predetermined icon is displayed on the display / input unit 14. It is also possible to set an arbitrary allowable verification rate by pressing with. In this case, the display / input unit 14 presses the portion corresponding to the number “8” and the portion corresponding to the number “0” with the finger in this order, thereby setting the allowable collation rate to 80. A screen example when set to% is shown.
[0203]
As described above, it is possible to change and set the allowable verification rate to an arbitrary value without changing the source code of the verification program simply by rewriting the allowable verification data recorded in the predetermined file.
[0204]
Alternatively, by using a predetermined editor program (editing software), accessing the file directly and rewriting the portion where the allowable verification data is recorded, thereby changing or setting the allowable verification rate. Is also possible.
[0205]
In the above embodiment, the insurance card image data is captured using a color scanner. However, it is also possible to capture image data using a color CCD camera or other imaging device.
[0206]
In addition, it is needless to say that various specific numerical values used in the above-described embodiment are one example, and other numerical values may be used.
[0207]
【The invention's effect】
The health insurance card automatic recognition method according to claim 1, and the claim12According to the health insurance card automatic recognition device described in, the image of the health insurance card is read and the read imageBased on the first image and the second image obtained by binarizing, the ruled line of the health insurance card is detected and detectedBased on the coordinates of the ruled line, the position where the predetermined item of the health insurance card is described is detected, the area where the character is described is detected based on the position where the item is described, and based on this area Since the health insurance card is confirmed by recognizing a predetermined item of the health insurance card, the health insurance card confirmation operation can be automated. Therefore, it is possible to easily and quickly determine whether the insurance card is valid, reduce uncollected medical expenses, and rationalize paperwork such as reception work in a hospital or the like.In addition, numbers such as an insured number and an insurer number can be reliably and promptly recognized, and a health insurance card confirmation process can be efficiently performed.
[0208]
According to the health insurance card automatic recognition method according to claim 2 and the health insurance card automatic recognition device according to claim 8, when the health insurance card is inclined, the characters included in each frame of the insurer number are A rectangle of a predetermined size consisting of horizontal or vertical sides for recognition is defined as a character area, this character area is reduced by a predetermined length vertically and horizontally, and characters in the reduced character area are recognized, Since the health insurance card is confirmed, the accuracy of character recognition can be improved. Therefore, the health insurance card confirmation process can be performed efficiently.
[0209]
According to the health insurance card automatic recognition method according to claim 3 and the health insurance card automatic recognition device according to claim 9, the position of the seal stamped on the health insurance card based on the color component constituting the image Since the health insurance card is judged based on this position, it is possible to make a prompt judgment. Therefore, the processing speed can be improved.
[0210]
According to the health insurance card automatic recognition method according to claim 5 and the health insurance card automatic recognition device according to claim 11, based on the first image and the second image included in the character recognition area, Since the position of the described character is detected, the region where the character to be recognized exists can be specified. Accordingly, it is possible to reliably and promptly recognize characters, and to efficiently check the health insurance card.
[0212]
Claim13The method for automatically recognizing a geriatric health insurance card according to claim 1, and claims16According to the geriatric health insurance card automatic recognition device described in the above, it reads the geriatric health insurance card image,Based on the first and second images obtained by binarizing the read image, the ruled line of the geriatric health insurance card was detected and detected.Based on the number of ruled lines existing in a predetermined area of the image, it is determined whether the image is an image of a geriatric medical insurance card, and a predetermined item of the geriatric medical insurance card is recognized from the image. Since the health insurance card is confirmed, the confirmation work for the geriatric health insurance card can be automated. Therefore, it is possible to easily and quickly determine whether the geriatric health insurance card is valid, reduce uncollected medical expenses, and rationalize paperwork such as reception work in a hospital or the like.Moreover, the position where the municipality number and the recipient number are described can be detected. Accordingly, it is possible to reliably and promptly recognize numbers such as municipality numbers and beneficiary numbers, and it is possible to efficiently perform the process for confirming a geriatric medical insurance card.
[0213]
According to the geriatric health insurance card automatic recognition method according to claim 15 and the geriatric health insurance card automatic recognition device according to claim 18, when the geriatric health insurance card is inclined, it is included in each frame of the insurer number. The character area is a rectangle with a predetermined size consisting of a horizontal or vertical side for recognizing the character to be recognized, and the character area is reduced by a predetermined length vertically and horizontally, and the characters in the reduced character area are Since it is made to recognize, the accuracy of character recognition can be improved. Accordingly, it is possible to reliably and promptly recognize characters, and it is possible to efficiently perform a geriatric health insurance card confirmation process.
[Brief description of the drawings]
FIG. 1 is a diagram showing a relationship between a medical insurance application population and a population that can be automatically verified.
FIG. 2 is a diagram showing types and shapes of health insurance insured cards.
FIG. 3 is a diagram showing a format of a health insurance insured card.
FIG. 4 is a diagram showing a format of a special day certificate for insured persons (one year).
FIG. 5 is a diagram showing a format of a special daily medical treatment receipt slip for special day insured persons.
FIG. 6 is a diagram showing the format of a special day certificate for insured persons (5 years).
FIG. 7 is a diagram showing a format of a national health insurance insured card.
FIG. 8 is a diagram showing a format of a national health insurance retirement insured card.
FIG. 9 is a diagram showing a format of a sailor insurance insured card.
FIG. 10 is a diagram showing a sailor insurance dependent card format.
FIG. 11 is a diagram showing a format of a national public servant mutual aid association card.
FIG. 12 is a diagram showing a format of a national public employee mutual aid association remote area dependent certificate.
FIG. 13 is a diagram showing a format of a local public servant mutual aid association card.
FIG. 14 is a diagram showing a format of a local public employee mutual aid association remote area dependent certificate.
FIG. 15 is a diagram showing a format of a private school faculty and staff mutual aid association card.
FIG. 16 is a diagram showing the format of a private school teacher / employee mutual aid association remote area dependent certificate;
FIG. 17 is a block diagram showing the configuration of an embodiment of the insurance card automatic recognition apparatus of the present invention.
FIG. 18 is a flowchart for explaining a method of recognizing a number of an insurance card.
FIG. 19 is a diagram for explaining insurance frame search processing;
FIG. 20 is a view for explaining insurance frame search processing;
FIG. 21 is a flowchart showing a processing example when checking the inclination of a health insurance card.
FIG. 22 is a flowchart showing a processing example when detecting the coordinates of the four corners of the insurance card;
FIG. 23 is a flowchart showing a processing example when detecting the coordinates of the horizontal ruled line in the insurance card.
FIG. 24 is a diagram for explaining a horizontal ruled line search method;
FIG. 25 is a diagram for explaining a horizontal ruled line search method;
FIG. 26 is a diagram for explaining a horizontal ruled line search method;
FIG. 27 is a diagram for explaining a horizontal ruled line search method;
FIG. 28 is a diagram showing a predetermined insurance card pattern at a number position in the insurance card.
FIG. 29 is a diagram showing another insurance card pattern of the number position in the insurance card.
FIG. 30 is a flowchart showing an example of processing when a recognized position of an insured person number in an insurance card is taken out.
FIG. 31 is a diagram for explaining a method of specifying a recognized position of an insured person number;
FIG. 32 is a diagram for explaining a method of recognizing vertical ruled lines.
FIG. 33 is a diagram for explaining a method of extracting a character size and a pitch.
FIG. 34 is a flowchart showing an example of processing when searching for a character position;
FIG. 35 is a diagram for explaining a character recognition method;
FIG. 36 is a diagram showing a description example of an insurer number.
FIG. 37 is a flowchart showing an example of processing for extracting frame coordinates of an insurer number.
FIG. 38 is a diagram for explaining a method for specifying an insurer number area;
FIG. 39 is a diagram showing a case where a horizontal ruled line other than the insurer number frame is detected.
FIG. 40 is a diagram illustrating a case where a character string is regarded as a horizontal ruled line.
FIG. 41 is a diagram for explaining a search method for the left and right side surfaces of the insurer number area;
FIG. 42 is a diagram showing the interval between vertical ruled lines existing between the upper and lower side surfaces of the insurer number in units of 5 dots.
FIG. 43 is a diagram for explaining a method of extracting a character size of an insurer number.
FIG. 44 is a diagram for describing a method for setting a character area when the health insurance card is inclined upward.
FIG. 45 is a diagram for explaining a method for setting a character area when the health insurance card is inclined upward to the left.
FIG. 46 is a diagram for explaining a method of reducing a character area.
FIG. 47 is a table summarizing thresholds when searching for insurance policies.
FIG. 48 is a table summarizing threshold values when detecting a ruled line of an insurance card.
FIG. 49 is a flowchart illustrating an example of a number recognition process.
FIG. 50 is a diagram showing a histogram of number 0;
FIG. 51 is a diagram showing a histogram of numeral 1;
FIG. 52 is a diagram showing a histogram of numeral 2;
FIG. 53 is a diagram showing a histogram of numeral 3;
FIG. 54 is a diagram showing a histogram of numeral 4;
FIG. 55 is a diagram showing a histogram of numeral 5;
FIG. 56 is a diagram showing a histogram of numeral 6;
FIG. 57 is a diagram showing a histogram of numeral 7;
FIG. 58 is a diagram showing a histogram of numeral 9;
FIG. 59 is a diagram showing image data of number 6 before thinning processing;
FIG. 60 is a diagram showing image data of number 6 after thinning processing.
FIG. 61 is a diagram illustrating image data of a number 0 before thinning processing.
FIG. 62 is a diagram showing image data with a number 0 after thinning processing;
FIG. 63 is a diagram showing image data of number 4 before thinning processing;
FIG. 64 is a diagram showing image data of number 4 after thinning processing;
FIG. 65 is a diagram showing image data of number 6 before thinning processing;
FIG. 66 is a diagram showing image data of number 6 after thinning processing.
FIG. 67 is a diagram showing the principle of the sonde method.
FIG. 68 is a diagram showing a standard number writing method based on recognition by the sonde method.
FIG. 69 is a diagram showing deformed characters that can be recognized by the sonde method for the number 3;
FIG. 70 is a flowchart illustrating a processing example of up / down determination.
FIG. 71 is a flowchart illustrating another processing example of up / down determination.
FIG. 72 is a block diagram showing a configuration example of an elderly medical insurance card automatic recognition apparatus of the present invention.
FIG. 73 is a flowchart showing a processing example of an automatic collation process for a geriatric medical insurance card;
74 is a flowchart showing an example of processing in step S135 of FIG. 73. FIG.
75 is a flowchart showing an example of processing in step S142 of FIG. 74. FIG.
FIG. 76 is a diagram for explaining a method of scanning a geriatric medical insurance card;
FIG. 77 is a diagram for explaining a method for obtaining a frame line for a medical insurance card for the elderly.
FIG. 78 is a diagram for explaining a method for identifying whether or not a medical insurance card for the elderly.
FIG. 79 is a flowchart showing an example of a process for detecting a horizontal ruled line and identifying whether or not the medical card is a geriatric medical insurance card.
FIG. 80 is a flowchart showing an example of processing in steps S149 and S151 of FIG.
FIG. 81 is a diagram showing a method of searching for a municipality number field and a recipient number field.
FIG. 82 is a flowchart showing a processing example when changing the allowable collation rate.
FIG. 83 is a flowchart showing another processing example when changing the allowable collation rate.
FIG. 84 is a diagram showing an example of an input screen when changing the allowable collation rate.
FIG. 85 is a diagram showing another example of an input screen when changing the allowable collation rate.
FIG. 86 is a flowchart for explaining a conventional procedure for confirming a geriatric medical insurance card.
[Explanation of symbols]
1 insurance card
2 Color scanner
3 Display section
4 processing section
5 Keyboard
6 storage unit
7 Output section
11 Patient card input section
11A patient card
12 Color image input section
12A Health insurance card for the elderly
13 Confirmation processing part
14 Display / input section
15 Voice guide
16 Patient database
17 Printing department

Claims (18)

Read the health insurance card image
By binarizing the read image with a predetermined threshold, it is divided into a first image and a second image,
The search start point is a predetermined position of the image,
A vertical direction or a horizontal direction with respect to the health insurance card is set as a search direction, a ruled line direction to be searched is set as a direction perpendicular to the search direction with respect to the health insurance card, and a predetermined direction perpendicular to the search direction is set. Search the second image of the image in the search direction from the search start point with a search width of
When the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the minimum ruled line detection that extends in a direction perpendicular to the search direction and is equal to or greater than the search width Defines a rectangular region in which predetermined small regions having a long side of a long line segment are arranged in a matrix, and the other short side from the row of small regions on one short side of the region In order of the column on the side, it is determined whether or not the second image is included in the column,
When it is determined that the second image is included, a predetermined counter is counted up,
When the ratio of the counter value to the minimum ruled line detection length is equal to or greater than a predetermined ratio, it is considered that the ruled line of the health insurance card has been detected,
Based on the coordinates of the ruled line of the detected health insurance card, a position where a predetermined item of the health insurance card is described is detected,
Based on the position where the detected item is described, the area where the character is described is detected,
Based on the area where the detected character is described, recognize the character described in the item,
A health insurance card automatic recognition method characterized in that, based on the recognition result, accepting insurance treatment for outpatients and confirming the health insurance card. The recognition of the characters described in the above item is
Detecting the inclination of the health insurance card from the inclination of the left side of the insurance policy frame of the health insurance card;
When the health insurance card is inclined upward to the right, the lower left point of the rectangular frame in which the insurer number is written is a vertex, and the horizontal and vertical line segments passing through the vertex are two sides. If the health insurance card is tilted upward to the left, the top left point of the rectangular frame on which the insurer number is written is the vertex, the horizontal passing through the vertex, And a rectangular frame of a predetermined size with a vertical line segment as two sides as a character area,
The character area is reduced by a predetermined length vertically and horizontally,
2. The health insurance card automatic recognition method according to claim 1, wherein the method is performed by recognizing characters in the reduced character region. Recognizing the position of the seal stamped on the health insurance card based on the color components constituting the image,
2. The health insurance card automatic recognition method according to claim 1, wherein upper and lower of the health insurance card are determined based on the recognized position of the seal. The item is at least one of a health insurance card symbol, number, insurance association name, insurer number, effective date, date of birth, and date of acquisition. Item 4. The health insurance card automatic recognition method according to any one of Items 1 to 3 . Detection of the area where the letters of the health insurance card are written
By binarizing the image with a predetermined threshold value, it is divided into a first image and a second image,
A character recognition area in which predetermined small areas are arranged in a matrix on the image is set,
Determining whether the second image is included in the column in order from the left column of the small region constituting the character recognition region in the right direction;
It is determined whether or not the sequence that does not include the second image continues for more than a predetermined number of character boundary determinations,
When it is determined that a column that does not include the second image is more than the number of character boundary determinations, the first column in which the second image is detected is the left end of the character, and the second image is detected. The last column is the right end of the character, the position of the second image detected at the bottom of the character recognition area is the bottom of the character, and the position of the second image detected at the top of the character recognition area is The top of the character,
The distance from the left end to the right end is a character width, the distance from the lower end to the upper end is a character height, and the character width is not less than a predetermined character width minimum value defined as the minimum value of the character width, Further, when the character height is equal to or more than a predetermined minimum character height defined as the minimum value of the character length, it is performed by assuming that a character is detected . The health insurance card automatic recognition method according to any one of the above. Thinning processing is performed on the character in the character region,
3. The health insurance card automatic recognition method according to claim 2 , wherein the character subjected to the thinning process is recognized using a sonde method. Image reading means for reading an image of a health insurance card;
Binarizing means for binarizing the image read by the image reading means into a first image and a second image by binarizing with a predetermined threshold value;
Search start point setting means having a predetermined position of the image as a search start point;
A vertical direction or a horizontal direction with respect to the health insurance card is set as a search direction, a ruled line direction to be searched is set as a direction perpendicular to the search direction with respect to the health insurance card, and a predetermined direction perpendicular to the search direction is set. Second image search means for searching the second image of the image in the search direction from the search start point with a search width of
When the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the minimum ruled line detection that extends in a direction perpendicular to the search direction and is equal to or greater than the search width Defines a rectangular region in which predetermined small regions having a long side of a long line segment are arranged in a matrix, and the other short side from the row of small regions on one short side of the region Determination means for sequentially determining whether or not the second image is included in the column in the column on the side;
Counting means for counting up a predetermined counter when it is determined that the second image is included;
If the ratio of the counter value to the ruled line detection minimum length is equal to or greater than a predetermined ratio, coordinate detection means that considers that the ruled line of the health insurance card has been detected;
A position detecting means on the basis of the coordinates of the ruled line of said detected health insurance card, detecting a position of a predetermined item is described in the health insurance card by the coordinate detection means,
A character area detecting means for detecting an area where a character is written based on a position where the item detected by the position detecting means is written;
Recognizing means for recognizing a predetermined item described in the health insurance card based on an area where the character detected by the character area detecting means is described;
A health insurance card automatic recognition apparatus, comprising: a confirmation means for accepting outpatient medical care based on a recognition result by the recognition means and confirming the health insurance card. The recognition means is
Inclination detecting means for detecting the inclination of the health insurance card from the inclination of the left side of the insurance policy frame of the health insurance card,
When the health insurance card is inclined upward to the right, the lower left point of the rectangular frame in which the insurer number is written is a vertex, and the horizontal and vertical line segments passing through the vertex are two sides. If the health insurance card is tilted upward to the left, the top left point of the rectangular frame on which the insurer number is written is the vertex, the horizontal passing through the vertex, And a character area setting means having a rectangular frame of a predetermined size with two vertical lines as two sides, and a character area,
Reduction means for reducing the character area by a predetermined length vertically and horizontally;
Character recognition means for recognizing characters in the character area reduced by the reduction means;
Health insurance card automatic recognition apparatus according to claim 7, characterized in that it comprises a. Seal position recognition means for recognizing the position of the seal stamped on the health insurance card based on the color components constituting the image;
8. The health insurance card automatic recognition apparatus according to claim 7 , further comprising an up / down determination means for determining up / down of the health insurance card based on the recognized position of the seal. The item is at least one of a health insurance card symbol, number, insurance association name, insurer number, effective date, date of birth, and date of acquisition. Item 10. The health insurance card automatic recognition device according to any one of Items 7 to 9 . The character area detecting means includes
Binarizing means for binarizing the image into a first image and a second image by binarizing the image with a predetermined threshold;
Setting means for setting a character recognition area in which predetermined small areas are arranged in a matrix on the image;
First determination means for determining whether or not the second image is included in the column in order from the left column of the small region constituting the character recognition region in the right direction;
Second determination means for determining whether or not the sequence not including the second image has continued for more than a predetermined number of character boundary determinations;
When it is determined that the sequence not including the second image is more than the number of character boundary determinations, the first column in which the second image is detected is the left end of the character, and the second image is detected. The last column is the right end of the character, the position of the second image detected at the bottom of the character recognition area is the bottom of the character, and the position of the second image detected at the top of the character recognition area is A character area setting means for the upper end of the character;
The distance from the left end to the right end is a character width, the distance from the lower end to the upper end is a character height, and the character width is equal to or greater than a predetermined character width minimum value defined as the minimum value of the character width, And when the character height is not less than a predetermined minimum character height defined as the minimum value of the character length, a character detection means that considers that a character has been detected;
Health insurance card automatic recognition apparatus according to any one of claims 7 to 9, characterized in that it comprises a. Further comprising a thinning means for performing a thinning process on the character in the character region;
9. The health insurance card automatic recognition apparatus according to claim 8 , wherein the character recognizing unit recognizes the character subjected to the thinning process by the thinning unit using a sonde method. Read the image of a medical insurance card for the elderly,
By binarizing the read image with a predetermined threshold, it is divided into a first image and a second image,
The search start point is a predetermined position of the image,
The search direction is the vertical direction or the left-right direction with respect to the geriatric health insurance card, the direction of the ruled line to be searched is the direction perpendicular to the search direction with respect to the geriatric health insurance card, and the direction perpendicular to the search direction Searching the second image of the image in the search direction from the search start point with a predetermined search width of
When the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the minimum ruled line detection that extends in a direction perpendicular to the search direction and is equal to or greater than the search width Defines a rectangular region in which predetermined small regions having a long side of a long line segment are arranged in a matrix, and the other short side from the row of small regions on one short side of the region In order of the column on the side, it is determined whether or not the second image is included in the column,
When it is determined that the second image is included, a predetermined counter is counted up,
When the ratio of the counter value to the ruled line detection minimum length is equal to or greater than a predetermined ratio, it is considered that the ruled line of the geriatric medical insurance card has been detected,
Based on the number of ruled lines present in the predetermined area of the detected image, whether the image is an image of a geriatric medical insurance card,
Detecting a position where a predetermined item of the geriatric health insurance card is described from the read image,
Recognizing the characters described in the item,
A method for automatically recognizing a geriatric medical insurance card, which receives insurance medical treatment based on the recognition result and confirms the geriatric medical insurance card. The method for automatically recognizing a geriatric medical insurance card according to claim 13 , wherein the item is at least one of a municipality number, a beneficiary number, and an expiration date of the geriatric medical insurance card. The recognition of the characters described in the item is to detect the inclination of the geriatric medical insurance card from the inclination of the left side of the insurance card frame of the geriatric medical insurance card,
When the geriatric health insurance card is inclined upward, the horizontal and vertical line segments passing through the vertex are defined as the top left corner of each rectangular frame in which the municipality number and the beneficiary number are written. When a rectangular frame of a predetermined size as one side is used as a character area and the medical insurance card for the elderly is inclined upward to the left, the upper left point of each rectangular frame in which the municipality number and the recipient number are described As a character area, a rectangular frame of a predetermined size having two sides of a horizontal and vertical line segment passing through the vertex,
The character area is reduced by a predetermined length vertically and horizontally,
The method for automatically recognizing a geriatric medical insurance card according to claim 13 , wherein the recognition is performed by recognizing a character in the reduced character region. Image reading means for reading an image of a medical insurance card for the elderly,
Binarizing means for binarizing the image read by the image reading means into a first image and a second image by binarizing with a predetermined threshold value;
Search start point setting means having a predetermined position of the image as a search start point;
The search direction is the vertical direction or the left-right direction with respect to the geriatric medical insurance card, the direction of the ruled line to be searched is the direction perpendicular to the search direction with respect to the geriatric medical insurance card, and the direction perpendicular to the search direction Second image search means for searching for the second image of the image in the search direction from the search start point with a predetermined search width of
When the ratio between the second image and the first image included in the search width is equal to or greater than a predetermined ratio, the minimum ruled line detection that extends in a direction perpendicular to the search direction and is equal to or greater than the search width Defines a rectangular region in which predetermined small regions having a long side of a long line segment are arranged in a matrix, and the other short side from the row of small regions on one short side of the region Determination means for sequentially determining whether or not the second image is included in the column in the column on the side;
Counting means for counting up a predetermined counter when it is determined that the second image is included;
When the ratio of the counter value to the ruled line detection minimum length is equal to or greater than a predetermined ratio, detection means that considers that the ruled line of the geriatric medical insurance card has been detected;
Identification means for identifying whether or not the image is an image of a geriatric medical insurance card based on the number of ruled lines present in a predetermined region of the image detected by the detection means;
Recognizing means for recognizing a predetermined item described in the medical insurance card for the elderly from the image read by the image reading means;
A geriatric health insurance card automatic recognition apparatus, comprising: a confirmation means for accepting outpatient medical care based on a recognition result by the recognition means and confirming the geriatric health insurance card. The apparatus for automatically recognizing a geriatric health insurance card according to claim 16 , wherein the item is at least one of a municipality number, a beneficiary number, and an expiration date of the geriatric health insurance card. The identification means includes
Inclination detecting means for detecting the inclination of the geriatric medical insurance card from the inclination of the left side of the insurance card frame of the geriatric medical insurance card,
When the geriatric health insurance card is inclined upward to the right, the lower left point of each rectangular frame in which the municipality number and the beneficiary number are written is a vertex, and a horizontal and vertical line segment passing through the vertex is represented by 2 When a rectangular frame of a predetermined size as one side is used as a character area and the medical insurance card for the elderly is inclined upward to the left, the upper left point of each rectangular frame in which the municipality number and the beneficiary number are described A character area setting means having a rectangular frame of a predetermined size having two sides of a horizontal and vertical line segment passing through the vertex,
Reduction means for reducing the character area by a predetermined length vertically and horizontally;
Character recognition means for recognizing characters in the character area reduced by the reduction means;
The apparatus for automatically recognizing a geriatric medical insurance card according to claim 16 .

Images