JP7254606B2 - Estimation work support system and estimation work support program - Google Patents

Description

TECHNICAL FIELD The present invention relates to an estimation work support system and an estimation work support program for assisting an estimation work by automatically picking up the names and numbers of electric appliances and the like from design drawings of image data.

Conventionally, various systems have been proposed for supporting estimation work (see, for example, Patent Document 1 (Japanese Unexamined Patent Application Publication No. 2006-113940)).
JP-A-2006-113940

The inventors extract (pick up) the product names and numbers of electrical appliances from the building design drawings provided by customers, and calculate the electrical work costs by accumulating them. By the way, the design drawing provided by the customer is characterized in that it is image data rather than CAD data or the like. This is because if CAD data etc. are received from the customer side, there is a possibility that they may be erroneously altered. I am trying to receive it. Therefore, conventionally, in the work of calculating the cost of electrical work, it is necessary to pick up symbols and texts of electrical appliances based on image data, and such picking work is substantially done manually. There was a problem that the work efficiency was very poor.

A specific description will be given with reference to FIGS. 17 and 18. FIG. (A) of each of FIGS. 17 and 18 is a diagram showing an example of a design drawing of a building. In relation to FIG. 17(A), the legend table in FIG. 17(B) defines symbols as symbol marks of electric appliances and names corresponding to the symbols. In the multiplication work from the design drawing, it is necessary to extract the symbols from the main drawing of the design drawing. However, until now, there has been no known technique for efficiently extracting a plurality of symbols defined in the table of legends by computer matching in the design drawings of the image data. (Task 1)
Furthermore, in the design drawing, for example, the figure in FIG. 18(B) is described in relation to FIG. 18(A). In such a drawing, the "product name" and the like are described. Such "product name" is converted into text data by character recognition technology such as OCR in figure drawings and design drawings, but conversion errors often occur at this time. Conventionally, however, there is no choice but to rely on people to correct such conversion errors, and there has been a demand for a technique for automatically and efficiently correcting them. (Task 2)

In order to solve such a problem, the estimation work support system according to the present invention extracts a plurality of symbol patterns related to electric appliances in the design drawing from the design drawing made up of image data, and extracts the product names in the design drawing. A recognition integration work support system comprising: a step of storing rectangular dot patterns of a plurality of symbols in a predetermined rectangular region; are compared, the correlation coefficient is calculated, and the correlation coefficient function for each symbol is obtained; A step of extracting a symbol as a correct answer for which there are no symbols exceeding a predetermined threshold, a step of calculating a maximum value function that gives the maximum value of the correlation coefficient function for all symbols in the specified range, a step of maximal value in the maximum value function A process of extracting a symbol that gives a value as a correct answer, a process of converting the product name in the design drawing into text based on character recognition technology, and if the converted text and the correct text of the product name do not match, the converted text is corrected to the correct text with the closest edit distance, and when the converted text and the correct text of the product name do not match, when there are multiple correct texts with the closest edit distance, the converted text Of these, a step of replacing similar characters with representative characters for those having similar characters, a step of replacing similar characters with representative characters for those having similar characters among the correct texts, and a conversion that has been replaced calculating an edit distance between the converted text and the replaced correct text; and correcting the converted text with a shorter edit distance to the correct text. do.

In addition, the estimation work support program according to the present invention executes a process of extracting a plurality of symbol patterns related to electric appliances in the design drawings from the design drawings made up of image data, and a process of recognizing product names in the design drawings on the computer. and a step of storing rectangular dot patterns of a plurality of symbols in a predetermined rectangular region, and a design drawing while shifting the rectangular dot patterns of each of the plurality of symbols by one dot. performing the comparison, calculating the correlation coefficient, and determining the correlation coefficient function for each symbol; A step of extracting a correct symbol for which there are no symbols exceeding a predetermined threshold, a step of calculating a maximum value function that gives the maximum value of the correlation coefficient function for all symbols in the specified range, and a maximum value in the maximum value function. a step of extracting the symbol that gives , in the process of correcting to the correct text with the closest edit distance, and when the converted text and the correct text of the product name do not match, when there are multiple correct texts with the closest edit distance, out of the converted text , a step of replacing similar characters with representative characters for those having similar characters, a step of replacing similar characters with representative characters for those having similar characters among the correct texts, and calculating the edit distance between the replaced text and the replaced correct text; and correcting the converted text with the shorter edit distance to the correct text. do.

In the estimation work support system and the estimation work support program according to the present invention, the obtained correlation coefficient function is a symbol exceeding a predetermined threshold, and there is no nearby symbol having a correlation coefficient function exceeding the predetermined threshold. Since the symbol is extracted as the correct answer, according to the estimation work support system and the estimation work support program according to the present invention, a plurality of symbols similar to the symbols defined in the legend table are extracted from the design drawing of the image data. From this figure, it is possible to perform computer matching and extract efficiently.

In addition, the estimation work support system and the estimation work support program according to the present invention, when the converted text does not match the correct text of the product name described in the figure drawing, converts the converted text to the closest edit distance. Since the text is corrected to the correct text, according to the estimation work support system and the estimation work support program according to the present invention, it is possible to efficiently correct the text data erroneously converted by the character recognition technology.

1 is a diagram showing a system configuration example of a computer that executes an estimation work support system 1 according to an embodiment of the present invention; FIG. It is a figure which illustrates a part of (A) legend table in a design drawing, and (B) figure drawing. It is a figure which shows three types of symbols simplified for description. 4 is a diagram showing a flowchart of symbol extraction processing in the estimation work support system 1 according to the embodiment of the present invention; FIG. 4 is a diagram showing a flowchart of a correlation coefficient function calculation/storage processing subroutine in the integration work support system 1 according to the embodiment of the present invention; FIG. FIG. 3 is a diagram showing a symbol dot pattern of w×h dots having reference dots, and a main design drawing (plan view) of W×H dots, which is a scan target range; FIG. 10 is a diagram schematically illustrating processing for calculating a correlation coefficient; FIG. 10 is a diagram schematically illustrating the processing of calculating correlation coefficients for various symbols; 4 is a diagram showing a flowchart of a first extraction processing subroutine in the estimation work support system 1 according to the embodiment of the present invention; FIG. FIG. 4 is a diagram showing changes in values of correlation coefficients for the three types of symbols shown in FIG. 3; FIG. 5 is a diagram showing a flowchart of a second extraction processing subroutine in the estimation work support system 1 according to the embodiment of the present invention; FIG. 3 is a diagram illustrating a maximum value function c(x, y) and a symbol that gives a maximum value of the maximum value function c(x, y); FIG. 10 is a diagram illustrating automatic generation of symbols with different orientations as matching source patterns; 4 is a diagram showing a flow chart of product name recognition processing in the estimation work support system 1 according to the embodiment of the present invention. FIG. FIG. 10 is a diagram showing a case where there are multiple correct texts with close edit distances; It is a figure explaining the determination by the replacement process by a representative character, and edit distance calculation. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows an example of the design drawing of a building, (A) has shown the top view, (B) has shown the legend table. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows an example of the design drawing of a building, (A) has shown the top view, (B) has shown the floor plan.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing an example system configuration of a computer that executes a program of an estimation work support system according to an embodiment of the present invention.

In FIG. 1, 10 is a system bus, 11 is a CPU (Central Processing Unit), 12 is a RAM (Random Access Memory), 13 is a ROM (Read Only Memory), 14 is a communication control unit that controls communication with external information equipment, 15 is an input control unit such as a keyboard controller; 16 is an output control unit; 17 is an external storage device control unit; 18 is an input unit including input devices such as a keyboard, pointing device and mouse; Reference numeral 20 denotes an external storage device such as an HDD (Hard Disk Drive); 21, a graphics control unit; and 22, a display device.

In FIG. 1, the CPU 11 searches and acquires data by communicating with external devices according to programs stored in the program ROM in the ROM 13 or the large-capacity external storage device 20. It performs arithmetic processing such as processing output data in which figures, images, characters, tables, etc. are mixed, and managing the database stored in the external storage device 20.

Further, the CPU 11 comprehensively controls each device connected to the system bus 10 . A program ROM in the ROM 13 or an external storage device 20 stores an operating system program (OS), which is a basic program for controlling the CPU 11, and the like. In addition, the ROM 13 or the external storage device 20 stores various data used when performing output data processing and the like. A RAM 12, which is a main memory, functions as a main memory for the CPU 11, a work area, and the like.

The input control unit 15 controls an input unit 18 from a keyboard or pointing device (not shown). Also, the output control unit 16 controls the output of an output unit 19 such as a printer.

The external storage device control unit 17 is an external storage device such as an HDD (Hard Disk Drive) or a solid state drive (SSD) that stores boot programs, various applications, font data, user files, edit files, printer drivers, and the like. 20. A system program for realizing the estimation work support system of the present invention is stored in the external storage device 20 as described above. Also, the graphics control unit 21 is a component for drawing processing of information to be displayed on the display device 22 .

Further, the communication control unit 14 controls communication with external devices via a network. It is configured to be able to transmit information to an external device.

In addition to the operating system program (OS), which is the control program for the CPU 11, the external storage device 20 stores a system program for operating the integration work support system 1 of the present invention on the CPU 11, and data used in this system program. is installed and stored. Spreadsheet software such as Excel (trademark) for executing spreadsheet processing is installed in the external storage device 20 of the present invention, and is programmed to operate in conjunction with the integration work support system 1 of the present invention. ing.

It is assumed that the data used by the system program that implements the estimation work support system 1 of the present invention is basically stored in the external storage device 20. It is also possible to obtain the information from an external device on the Internet or an intranet via the communication control unit 14 . Also, the data used by the system program that realizes the estimation work support system 1 of the present invention can be configured to be acquired from various media such as a USB memory, CD, and DVD. The estimation work support system according to the present invention is implemented by executing a program on the computer configured as described above.

In the following, the present embodiment will be described based on an example of picking up and extracting symbols, product names, numbers, etc. of electric appliances from design drawings. In addition, the image data of the design drawing consists of a table of legends, a plan view (this drawing), and a figure. To explain, the data that can be used in the present invention is not limited to a combination of a legend table, a plan view (this view), and a profile view.

The inventors extract (pick up) the product names and numbers of electrical appliances from the building design drawings provided by customers, and calculate the electrical work costs by accumulating them. Here, the design drawings provided by the customer are image data, not CAD data, etc., and in conventional electrical work cost calculation work, lighting fixtures are picked up based on image data. This was essentially manual work.

Therefore, in the estimation work support system 1 according to the present invention, it is intended to extract symbols of electric appliances and text data necessary for estimating electrical work costs from image data of design drawings provided by customers. be.

FIG. 2 is a diagram exemplifying a part of (A) a legend table and (B) a drawing in a design drawing. In the legend table of FIG. 2A, "symbols" as symbol marks of electric appliances and "names" corresponding to the symbols are defined. In the multiplication work from the design drawing, it is necessary to extract such symbols from the main drawing of the design drawing. However, until now, there has been no known technique for efficiently extracting a plurality of symbols defined in the table of legends by computer matching in the design drawings of the image data. (Task 1)
Also, in the drawing of FIG. 2(B), "product name" and the like are described. Such figure drawings and design drawings are converted into text data by character recognition technology such as OCR, but at this time conversion errors often occur. Conventionally, however, there is no choice but to rely on people to correct such conversion errors, and there has been a demand for a technique for automatically and efficiently correcting them. (Task 2)
It should be noted that the drawing is characterized by the inclusion of a perspective view in the figure drawing. Illustrations etc. are omitted.

Specific methods for solving these problems 1 and 2 by the estimation work support system 1 and the estimation work support program according to the present invention will be described below.

First, a technique for extracting a "symbol" from a main drawing of a design drawing by matching processing will be described. FIG. 3 is a diagram showing three simplified symbols for explanation. Such a symbol is assumed to have a matrix of w dots horizontally and h dots vertically in computer dot coordinates. Note that the symbol to be matched in this drawing of the design drawing also has a matrix of w dots in the horizontal direction and h dots in the vertical direction. Also, the symbols shown in FIG. 3 and the symbols in this drawing of the design drawing are composed of binary values.

The dot at the upper left corner of w dots in the horizontal direction and h dots in the vertical direction is defined as the reference dot. When extracting the dot pattern of the symbol as shown in FIG. 3 by performing matching processing on the main drawing of the design drawing, while moving the dot pattern of the symbol along with the reference dot by one dot, Take the correlation coefficient with Then, a location where the correlation coefficient is higher than a predetermined threshold value is treated as if there is a symbol. As a result, symbols can be extracted from the drawing.

However, in the main drawing of the design drawing, the pattern of the symbol of the electrical appliance may overlap with the line of the architectural drawing, or the symbols of the electrical appliance may be similar to each other in the first place, Since there are situations where determination is difficult, in order to deal with this, the concept of a maximum value function is introduced in the present invention, and a symbol that gives a maximum value to such a maximum value function is extracted. There is a major feature of the present invention.

Now, based on the above, an example of an algorithm for extracting symbols from the main drawing of the design drawing will now be described. FIG. 4 is a diagram showing a flowchart of symbol extraction processing in the estimation work support system 1 according to the embodiment of the present invention.

In FIG. 4, when the symbol extraction process is started in step S100, the process proceeds to step S101, and a correlation coefficient function calculation/storage process subroutine is executed. Such a subroutine will be described with reference to FIG. Such a subroutine is executed for all symbols to be extracted (three types of symbols (A) to (C) in FIG. 3 in this example).

In FIG. 5, when the correlation coefficient function calculation/storage processing subroutine is started in step S200, the process advances to step S201 to set the coordinates of the reference dot pattern of the target symbol as the initial coordinates. As such initial coordinates, for example, the dot coordinates of the upper left corner of the main drawing of the design drawing are selected. Any coordinates can be used as long as they can be scanned dot by dot from the initial coordinates to the final coordinates without gaps.

FIG. 6 shows a symbol dot pattern of w×h dots having a reference dot in the upper left corner, and a design drawing (plan view) of W×H dots, which is the scan target range. In FIG. 6, the x-coordinate is defined horizontally to the right and the y-coordinate is defined vertically downward. Basically, for the symbol dot pattern of w×h dots, the correlation coefficient is calculated while scanning the design drawing main drawing (plan view) sequentially in the x-coordinate direction from the initial coordinates (1, 1). When the right end (W-w, 1) is reached, the y-coordinate is set to +1, and the correlation coefficient is calculated while scanning the design drawing main figure (plan view) sequentially from (1, 2) in the x-coordinate direction. continue. In the same manner, such steps are repeated to obtain the correlation coefficient while scanning the main design drawing (plan view) with the symbol dot pattern.

In step S202, the correlation coefficient between the dot pattern of the symbol and the dot pattern in the figure overlapping this dot pattern is calculated and stored.

In step S203, it is determined whether or not correlation coefficients have been calculated and stored for all areas (scan target range) of the design drawing. (That is, it is determined whether or not the reference dot has reached (Ww, Hh).) If the determination is NO, the process advances to step S204 to shift the coordinates of the reference dot pattern by one dot. Then, step S202 is executed again. On the other hand, if the determination is YES, the process proceeds to step S205 and returns to the original routine.

FIG. 7 is a diagram schematically explaining the processing of calculating the correlation coefficient. In FIG. 7, hatched symbols indicate symbols in the drawing of the design drawing. The x-coordinate is defined as a coordinate whose coordinate value increases as it proceeds horizontally to the right, and the y-coordinate is defined as a coordinate whose coordinate value increases as it proceeds vertically downward.

Also _, R ₁ is defined as the correlation coefficient between the dot pattern of the symbols shown in FIG. , and R ₃ is defined as the correlation coefficient between the symbol dot pattern shown in FIG. 3C and the dot pattern in FIG.

FIG. 7 shows how the steps in the correlation coefficient calculation/storage processing subroutine proceed in the order of (A)→(B)→(C)→(D). As can be seen from FIG. 7, the value of the correlation coefficient between the dot pattern of the symbol and the dot pattern of this figure changes depending on the overlapping degree. Note that the maximum value of the correlation coefficient is 1. For example, the correlation coefficient R ₁ changes from 0 to 0.16 to 0.98 to 0.30 in the order of (A)→(B)→(C)→(D) in FIG.

Also, FIG. 8 is a diagram for schematically explaining the processing of calculating the correlation coefficients for the three types of symbols (A) to (C) in FIG. In this embodiment, correlation coefficients are obtained in the form shown in FIG. 8 for all three types of symbols (A) to (C) in FIG. be taken away.

In step S205, the correlation coefficient function R _n (x, y) for each symbol is calculated based on the correlation coefficients calculated as described above (in this example, n=1, 2, 3). is stored in the storage unit. At step S206, the original routine is returned to.

After the correlation coefficient function calculation/storage processing subroutine is executed as described above, the process returns to the original routine shown in FIG. 4 and proceeds to step S102. Subsequently, the first extraction processing subroutine shown in FIG. 9 is executed. be.

After the subroutine of the first extraction process is started in step S300, the process proceeds to step S301 to acquire the correlation coefficient function calculated in the subroutine of FIG.

Subsequently, in step S302, the calculated correlation coefficient function is a symbol exceeding a predetermined threshold (Condition 1), and there is no nearby symbol having a correlation coefficient function exceeding a predetermined threshold (Condition 2 ) is determined.

FIG. 10 is a diagram showing changes in correlation coefficient values between the dot patterns of the three types of symbols shown in FIG. 3 and the main dot pattern shown in FIG. In FIG. 10, the horizontal axis indicates the x-coordinate, and the vertical axis indicates the respective correlation coefficient R values. Here, let R _th be the above predetermined threshold. The symbol corresponding to condition 1 above is the symbol exceeding the threshold value _Rth in the diagram shown in FIG. 10, and in this example, "Δ" shown in FIG.

Next, condition 2 will be examined. If the rectangular dot patterns of two symbols do not overlap, the correlation coefficient values of each symbol can be independently compared. On the other hand, when the rectangular dot patterns (w×h patterns) of two symbols overlap each other (this case is called “nearby” the two symbols), both of the threshold values R _th It is possible that it may be exceeded. In this case, highly accurate matching cannot be performed. Therefore, in the present invention, condition 2 is ANDed with condition 1 to avoid erroneous determination.

Now, if the above conditions 1 and 2 are satisfied and the determination in step S302 is YES, the process proceeds to step S303, and the target symbol is extracted as a matching correct answer.

On the other hand, if the determination in step S302 is NO, the process proceeds to step S303, and the symbol of interest is treated as an undetermined symbol for the time being.

When the process proceeds to step S305, the process returns to the original routine and returns to the flow chart of FIG. In the flowchart of FIG. 4, in step S103, it is determined whether or not an undetermined symbol exists in the target dot pattern of this figure.

Here, if there is no undetermined symbol and the determination in step S103 is NO, the process proceeds to step S105 and the symbol extraction process ends.

On the other hand, if there is an undetermined symbol and the determination in step S103 is YES, the process proceeds to step S104, and the second extraction processing subroutine is executed. This second extraction processing subroutine will be described with reference to FIG.

When the second extraction processing subroutine is started in FIG. 11, the process proceeds to step S401 to calculate the maximum value function. Here, the maximum value function c(x, y) will be described with reference to FIG. FIG. 12 is a diagram for explaining the maximum value function c(x, y) and the symbol giving the maximum value of the maximum value function c(x, y). In addition, in FIG. 12, (x, y) is simply illustrated one-dimensionally.

A maximum value function c(x,y) is defined as giving the maximum value of the correlation coefficient function for all symbols in the specified range. Here, the designated range is from -ε to +ε with (x, y) as a reference. ε is a positive integer that satisfies 0<ε/w<1 and 0<ε/h<1.

Such a maximum value function c(x, y) can be obtained from the stored correlation coefficient function of each symbol.

Now, after calculating the maximum value function c(x, y) in step S401 as described above, the process proceeds to step S402. Determine if there is any that gives a local maximum.

When the determination in step S402 is YES, the process advances to step S404 to extract the symbol that gives the maximum value as the correct answer.

On the other hand, when the determination in step S402 is NO, the process advances to step S404, and it is processed as an error that extraction of the correct answer has failed.

In step S405, return to the original flow chart. Also in the original flowchart, the process proceeds to step S105 and ends.

As described above, in the estimation work support system 1 and the estimation work support program according to the present invention, the calculated correlation coefficient function is a symbol exceeding a predetermined threshold value, and the correlation coefficient function near the predetermined threshold value symbols that do not exist are extracted as correct answers. can be efficiently extracted from the design drawings of the image data by computer matching.

Further, in the estimation work support system 1 and the estimation work support program according to the present invention, when accurate extraction cannot be performed only with the correlation coefficient function, symbol extraction processing is performed based on the maximum value function c(x, y). is performed, extraction processing can be performed with higher accuracy.

3(A) (same as FIG. 13(A)) is the symbol shown in FIG. All patterns of the symbol shown in FIG. 13(C) rotated to the right and the symbol shown in FIG. 13(D) rotated by 90° to the right can be used as original symbols to be matched. preferable. This is because it is not known in what direction the above symbol "Δ" is arranged in the image data of this drawing of the design drawing. It is convenient to automatically generate the matching patterns of FIGS. 13B to 13D from the pattern of FIG. 13A.

Next, a specific method for solving problem 2 by the estimation work support system 1 and the estimation work support program according to the present invention will be described. Problem 2 was that there was no established technology for automatically and efficiently correcting conversion errors made by conventional character recognition technology.

Hereinafter, description will be made based on an example of reading the "product name" in a drawing similar to FIG. 2(B). Also, the product names "AdOI", "AcOx", and "AeOy" in FIG. In addition, the description will be made on the assumption that such correct text data is stored in advance in the system or program.

FIG. 14 is a diagram showing a flow chart of product name recognition processing in the estimation work support system 1 according to the embodiment of the present invention. In this example, the algorithm is called "product name recognition processing", but the object for which recognition processing can be performed in the present invention is not limited to the product name in the figure drawing.

In FIG. 14, when the flow chart of product name recognition processing is started at step S500, the flow advances to step S501 to convert the product name column of the target figure drawing into text based on the character recognition technology. A specific example of the figure to be read is not particularly shown here. Further, as a character recognition technique for converting read data into text data, conventionally well-known techniques can be appropriately used.

Subsequently, in step S502, the text converted in step S501 is compared with the correct text (list in FIG. 2(B)).

In step S503, it is determined whether or not there is a match in the list of correct texts. If the determination in step S503 is YES, the process proceeds to step S507, the converted text is treated as the correct answer without modification, the process proceeds to step S513, and the product name recognition process ends.

On the other hand, if the determination in step S503 is NO, the process advances to step S504 to calculate the edit distance between the converted text and the correct text.

Here, when converting a character string to another character string, when one character is inserted, deleted, or replaced as one time, the edit distance is the one character string to another character string. It can be defined as the (minimum) number of times that conversion can be performed by the process before conversion. A shorter edit distance indicates that the two character strings are closer.

Subsequently, in step S505, it is determined whether or not there is only one correct text with the closest edit distance. When the determination in step S505 is YES, the process proceeds to step S506, and the text converted by character recognition is corrected to the correct text. That is, automatic correction is performed in this step.

On the other hand, when the determination in step S505 is NO, that is, when there are a plurality of correct texts with close edit distances, the process proceeds to step S508.

FIG. 15 is a diagram showing a case where there are a plurality of correct texts with close edit distances. In step S505, the text " The edit distance of "ABOl" is considered. The edit distance from the character string “ABOl” of the converted text to the character strings of the correct text (“AdOI”, “AcOx”, and “AeOy”) is 2.

In the case shown in FIG. 15, in step S508, similar characters are replaced with representative characters. Here, it is assumed that 0 (zero) is a representative character of O (o), and 1 (ichi) is a representative character of I (eye) and l (el). In the example of FIG. 15, the character string "ABOl" of the converted text is replaced with "AB01" in the process of step S508 as shown in FIG.

In step S508, processing for replacing similar characters with representative characters is also executed for the correct text. The correct texts "AdOI", "AcOx", and "AeOy" are replaced with "Ad01", "Ac0x", and "Ae0y" by the replacement processing in this step. ”.

Subsequently, the process advances to step S509 to calculate the edit distance between the converted text (“AB01”) that has been replaced and the correct text that has been replaced (“Ad01”, “Ac0x”, and “Ae0y”). As shown in FIG. 16, the edit distance from the replaced converted text "AB01" to the replaced correct texts "Ac0x" and "Ae0y" is two. On the other hand, the edit distance from the replaced converted text "AB01" to the replaced correct text "Ad01" is one. (The latter is the only combination in step S510)
Subsequently, in step S510, it is determined whether or not there is only one combination with a shorter edit distance. If the determination result is NO, it is determined that which is the correct answer cannot be determined, and the process proceeds to step S512 to perform error processing.

On the other hand, when the determination result in step S510 is YES, the process advances to step S511 to perform so-called automatic correction, in which the converted text is corrected to the correct text. FIG. 16 shows an example of proceeding to step S510. According to this step, the character string "ABOl" of the converted text is automatically corrected to the correct text "AdOI".

In step S513, the product name recognition process is terminated. As explained above, by using both processing based on edit distance and, if necessary, replacement processing with representative characters, it is possible to increase the probability of automatically correcting text that has been erroneously converted by character recognition. becomes possible.

As described above, the estimation work support system 1 and the estimation work support program according to the present invention edit the converted text when the converted text does not match the correct text of the product name described in the figure drawing. Since the text is corrected to the correct text with the closest distance, according to the estimation work support system 1 and the estimation work support program according to the present invention, it is possible to efficiently correct the text data erroneously converted by the character recognition technology. It becomes possible.

Reference Signs List 1: Totalization work support system 10: System bus 11: CPU (Central Processing Unit)
12 RAM (Random Access Memory)
13 ROM (Read Only Memory)
Reference numeral 14: communication control unit 15: input control unit 16: output control unit 17: external storage device control unit 18: input unit 19: output unit 20: external storage device 21 ... graphic control unit 22 ... display device

Claims (2)

An estimation work support system that extracts a plurality of symbol patterns related to electrical appliances in the design drawings from the design drawings made up of image data and recognizes product names in the design drawings,
storing a rectangular dot pattern of a plurality of symbols in a predetermined rectangular area;
A step of comparing with the design drawing while shifting the rectangular dot patterns of each of the plurality of symbols by one dot, calculating the correlation coefficient, and obtaining the correlation coefficient function for each symbol;
a step of extracting, as a correct symbol, a symbol whose correlation coefficient function exceeds a predetermined threshold and for which there is no nearby symbol whose correlation coefficient function exceeds a predetermined threshold;
calculating a maximum value function that gives the maximum value of the correlation coefficient function for all symbols in the specified range;
A step of extracting a symbol that gives a maximum value in the maximum value function as a correct answer;
A process of converting the product name in the design drawing into text based on character recognition technology;
If the converted text and the correct text of the product name do not match, correcting the converted text to the correct text with the closest edit distance;
When the converted text and the correct text of the product name do not match, and there are multiple correct texts with the closest edit distance,
a step of replacing the similar characters with representative characters for those having similar characters among the converted text;
a step of replacing the similar characters with representative characters for those having similar characters among the correct texts;
calculating an edit distance between the replaced converted text and the replaced correct text;
correcting the converted text with the shorter edit distance to the correct text;
is executed. A cost estimation support program that causes a computer to execute a process of extracting a plurality of symbol patterns related to electrical appliances in the design drawing from the design drawing consisting of image data and a process of recognizing the item name in the design drawing,
storing a rectangular dot pattern of a plurality of symbols in a predetermined rectangular area;
A step of comparing with the design drawing while shifting the rectangular dot patterns of each of the plurality of symbols by one dot, calculating the correlation coefficient, and obtaining the correlation coefficient function for each symbol;
a step of extracting, as a correct symbol, a symbol whose correlation coefficient function exceeds a predetermined threshold and for which there is no nearby symbol whose correlation coefficient function exceeds a predetermined threshold;
calculating a maximum value function that gives the maximum value of the correlation coefficient function for all symbols in the specified range;
A step of extracting a symbol that gives a maximum value in the maximum value function as a correct answer;
A process of converting the product name in the design drawing into text based on character recognition technology;
If the converted text and the correct text of the product name do not match, correcting the converted text to the correct text with the closest edit distance;
When the converted text and the correct text of the product name do not match, and there are multiple correct texts with the closest edit distance,
a step of replacing the similar characters with representative characters for those having similar characters among the converted text;
a step of replacing the similar characters with representative characters for those having similar characters among the correct texts;
calculating an edit distance between the replaced converted text and the replaced correct text;
correcting the converted text with the shorter edit distance to the correct text;
Calculation work support program characterized by causing a computer to execute

Images