JP7484251B2 - IMAGE ANALYSIS DEVICE, IMAGE ANALYSIS METHOD, AND PROGRAM - Google Patents

Description

The present invention relates to an image analysis device, an image analysis method, and a program.

Various technologies have been proposed to recognize the indicated values shown on the indicators of various meters.

For example, the following Patent Document 1 discloses a technology for a meter having an indicator where a value is indicated by a needle, and for recognizing the indicated value by performing image recognition processing on an image of the indicator of the meter.

JP 2011-81715 A

Patent Document 1 only discloses image recognition processing for recognizing the indicated value of a meter having an indicator where the indicated value is indicated by a needle. However, there are many other types of meters other than meters having an indicator where the indicated value is indicated by a needle. Therefore, it is desirable to be able to obtain the indicated value of each meter with higher accuracy by image analysis according to the type of meter.

In view of the above-mentioned problems, the object of the present invention is to provide an image analysis device, an image analysis method, and a program that can obtain the indicated values of various meters with high accuracy by performing image analysis according to the type of meter.

In order to solve the above-mentioned problems, an image analysis device according to one aspect of the present invention includes an image processing unit that generates a first binary image by binarizing an image of a meter in which an indicated value is indicated by a pointer, and generates a second binary image by binarizing a histogram relating to the distribution of pixels indicating the pointer in the first binary image, and an image analysis unit that identifies the pointer position where the pointer is pointing to the indicated value based on the second binary image, and acquires the indicated value at the pointer position.

An image analysis method according to one aspect of the present invention includes an image processing unit generating a first binary image by binarizing an image of a meter in which an indicated value is indicated by a pointer, and generating a second binary image by binarizing a histogram relating to the distribution of pixels indicating the pointer in the first binary image, and an image analysis unit identifying a pointer position where the pointer is pointing to the indicated value based on the second binary image, and acquiring the indicated value at the pointer position.

A program according to one aspect of the present invention causes a computer to function as an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a pointer, and generates a second binary image by binarizing a histogram relating to the distribution of pixels indicating the pointer in the first binary image, and an image analysis unit that identifies the pointer position where the pointer is pointing to the indicated value based on the second binary image, and acquires the indicated value at the pointer position.

In order to solve the above-mentioned problems, an image analysis device according to one aspect of the present invention includes an image processing unit that generates a first binary image by binarizing an image of a meter whose indication value is indicated by a float movable along an axis, and generates a second binary image based on pixel values of pixels indicating the axis in the first binary image and pixel values of pixels on a predetermined line parallel to the axis in the first binary image, and an image analysis unit that identifies the position of the float based on the second binary image and obtains the indication value at the position of the float.

An image analysis method according to one aspect of the present invention includes an image processing unit generating a first binary image by binarizing an image of a meter whose indication value is indicated by a float movable along an axis, and generating a second binary image based on pixel values of pixels indicating the axis in the first binary image and pixel values of pixels on a predetermined line parallel to the axis in the first binary image, and an image analysis unit identifying the position of the float based on the second binary image and obtaining the indication value at the position of the float.

A program according to one aspect of the present invention causes a computer to function as an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a float movable along an axis, and generates a second binary image based on pixel values of pixels in the first binary image that indicate the axis and pixel values of pixels on a predetermined line parallel to the axis in the first binary image, and an image analysis unit that identifies the position of the float based on the second binary image and obtains the indicated value at the position of the float.

According to the present invention, it is possible to obtain the readings of various meters with high accuracy by performing image analysis according to the type of meter.

1 is a block diagram showing an example of a configuration of an image analysis system according to an embodiment of the present invention. 2 is a block diagram showing an example of a functional configuration of an imaging control terminal according to the present embodiment. FIG. 2 is a block diagram showing an example of a functional configuration of an image analysis terminal according to the present embodiment. FIG. 10 is a flowchart showing an example of a processing flow in the image analysis terminal according to the present embodiment. 11 is a diagram showing an example of HSV masking for a 7-segment meter image according to the embodiment; FIG. 10 is a flowchart showing an example of the flow of an analysis process for a seven-segment meter according to the present embodiment. 11A and 11B are diagrams illustrating an example of HSV masking for a meter image of a circular meter according to the present embodiment. 11A and 11B are diagrams illustrating an example of noise removal for a binary image of a meter image of a circular meter according to the embodiment. 11 is a diagram showing an example of a histogram of angle information based on a binary image of a meter image of a circular meter according to the present embodiment. FIG. FIG. 11 is a diagram showing an example of binarization of a histogram of angle information according to the embodiment. 5 is a flowchart showing an example of the flow of an analysis process for a circular meter according to the present embodiment. 11 is a diagram showing an example of HSV masking for a meter image of an edgewise meter according to the embodiment; FIG. 11 is a diagram showing an example of a histogram of position information based on a binary image of a meter image of an edgewise meter according to the embodiment; FIG. FIG. 11 is a diagram showing an example of binarization of a histogram of position information according to the embodiment. 5 is a flowchart showing an example of a flow of an analysis process for an edgewise meter according to the present embodiment. 4 is a diagram showing an example of binarization of a meter image of a float meter according to the present embodiment; FIG. 10A to 10C are diagrams illustrating an example of generation of a feature value image according to the present embodiment. 5 is a flowchart showing an example of a flow of an analysis process for a float meter according to the present embodiment.

The following describes in detail an embodiment of the present invention with reference to the drawings.

1. Image analysis system configuration
First, an example of the configuration of an image analysis system according to this embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of the configuration of an image analysis system 1 according to this embodiment. The image analysis system 1 is composed of an imaging device 2, an imaging control terminal 4, and an image analysis terminal 5 shown in Fig. 1.

The imaging device 2 is a device (e.g., a camera) that captures an image of the meters 3 (3-1 to 3-n (n is a natural number)). The imaging area 20 is an area that can be captured by the imaging device 2. The imaging device 2 according to this embodiment is provided so that the imaging area 20 includes multiple meters 3. Note that the number of meters 3 included in the imaging area 20 is not particularly limited. The alignment markers 30 are markers used to correct the orientation of the captured image. The imaging area 20 may include multiple alignment markers 30.

The meter 3 includes a variety of meters. For example, the meter 3 is a meter having an indication portion in which an indicated value is shown by a seven-segment display (hereinafter, also referred to as a "seven-segment meter"). In a seven-segment display, one character is displayed by combining seven segments. The character includes numbers, letters, and symbols.

Meter 3 may also be a meter having an indicator in which a value is indicated by a needle (hereinafter also referred to as a "pointer meter"). Examples of a pointer meter include a circular meter and an edgewise meter.

The pointer of a pointer meter can move in a restricted direction. For example, it moves in either a first direction or a second direction on a plane with a central axis as a reference. This causes the pointer of the pointer meter to indicate an indicated value. For example, the pointer of a circular meter moves in a clockwise direction (first direction) or counterclockwise direction (second direction) with a central axis as a reference.

In this embodiment, the dial of the pointer meter is fixed, and the pointer is a type of meter that moves. However, this is not limited to this example, and the pointer meter may be a type that has a moving dial.

The meter 3 may also be a meter having an indication portion in which an indicated value is indicated by a float (hereinafter, also referred to as a "float meter"). The float of the float meter may move in a regulated direction. For example, the float moves in either a first direction or a second direction along an axis. In this way, the float of the float meter indicates an indicated value.

The imaging control terminal 4 is a terminal that controls the imaging device 2. The imaging control terminal 4 is, for example, a terminal such as a PC (personal computer). The control of the imaging device 2 is, for example, control of the timing at which the imaging device 2 captures an image of the meter 3.

The image analysis terminal 5 is a terminal that analyzes the captured image captured by the imaging device 2. The image analysis terminal 5 is an image analysis device according to this embodiment. The image analysis terminal 5 is, for example, a terminal such as a PC. The image analysis terminal 5 obtains the indicated value shown on the indication section of the meter 3 by analyzing the captured image.

The imaging device 2 and the imaging control terminal 4 are connected so that they can communicate with each other. For example, the imaging device 2 and the imaging control terminal 4 are connected by a signal cable and perform wired communication. Through this communication, the imaging device 2 transmits captured images to the imaging control terminal 4 and receives control signals from the imaging control terminal 4. Through this communication, the imaging control terminal 4 transmits control signals to the imaging device 2 and receives captured images from the imaging device 2. Note that the imaging device 2 and the imaging control terminal 4 may also perform wireless communication.

The imaging control terminal 4 and the image analysis terminal 5 are connected so that they can communicate with each other. For example, the imaging device 2 and the imaging control terminal 4 are connected via a network NW and communicate with each other by wire or wirelessly. The imaging control terminal 4 transmits captured images to the image analysis terminal 5 through this communication. The image analysis terminal 5 receives captured images from the imaging control terminal 4 through this communication.

2. Functional configuration of the imaging control terminal 4
An example of the configuration of the image analysis system 1 according to the present embodiment has been described above. Next, an example of the functional configuration of the imaging control terminal 4 according to the present embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example of the functional configuration of the imaging control terminal 4 according to the present embodiment. As shown in Fig. 2, the imaging control terminal 4 includes a communication unit 410, a control unit 420, a storage unit 430, and an output unit 440.

The communication unit 410 has a function of transmitting and receiving various information. For example, the communication unit 410 transmits a control signal to the imaging device 2 and receives captured images from the imaging device 2. The communication unit 410 outputs the received captured images to the control unit 420. The communication unit 410 also transmits the captured images input from the control unit 420 to the image analysis terminal 5.

The control unit 420 controls the overall operation of the imaging control terminal 4. The control unit 420 is realized by causing a CPU (Central Processing Unit) provided as hardware in the imaging control terminal 4 to execute a program. As shown in FIG. 2, the control unit 420 includes an imaging control unit 4202 and an output processing unit 4204.

The imaging control unit 4202 controls the timing at which the imaging device 2 captures an image of the meter 3. The imaging control unit 4202 causes the communication unit 410 to transmit a control signal to the imaging device 2 to cause the imaging device 2 to capture an image of the meter 3.

The output processing unit 4204 controls the output in the imaging control terminal 4. For example, the output processing unit 4204 outputs the captured image to the communication unit 410 and transmits it to the image analysis terminal 5. The output processing unit 4204 may also output the captured image to the storage unit 430 and store it. The output processing unit 4204 may also output the captured image to the output unit 440 and display it.

The storage unit 430 is configured by a storage medium, for example, a hard disk drive (HDD), a flash memory, an electrically erasable programmable read only memory (EEPROM), a random access read/write memory (RAM), a read only memory (ROM), or any combination of these storage media. The storage unit 430 can be, for example, a non-volatile memory.
The storage unit 430 stores various types of information, such as captured images input from the output processing unit 4204.

The output unit 440 is, for example, a display device such as a display.
The output unit 440 outputs various information, such as a captured image input from the output processing unit 4204.

<3. Functional configuration of image analysis terminal 5>
An example of the functional configuration of the imaging control terminal 4 according to this embodiment has been described above. Next, an example of the functional configuration of the image analysis terminal 5 according to this embodiment will be described with reference to Fig. 3. Fig. 3 is a block diagram showing an example of the functional configuration of the image analysis terminal 5 according to this embodiment. As shown in Fig. 3, the image analysis terminal 5 includes a communication unit 510, a control unit 520, a storage unit 530, and an output unit 540.

The communication unit 510 has a function of transmitting and receiving various information. For example, the communication unit 510 receives captured images transmitted from the imaging control terminal 4. The communication unit 510 outputs the received captured images to the control unit 520.

The control unit 520 controls the overall operation of the image analysis terminal 5. The control unit 520 is realized by causing a CPU (Central Processing Unit) provided as hardware in the image analysis terminal 5 to execute a program. As shown in FIG. 3, the control unit 520 includes a correction unit 5202, a determination unit 5204, an image processing unit 5206, an image analysis unit 5208, and an output processing unit 5210.

The correction unit 5202 performs a correction process on the captured image input from the communication unit 510. For example, the correction unit 5202 performs a facing correction on the captured image based on the alignment marker 30 in the captured image.

Specifically, first, the correction unit 5202 detects multiple alignment markers 30 from the captured image. The correction unit 5202 calculates the distance between the detected multiple alignment markers 30, and then calculates the aspect ratio of the distance between the multiple alignment markers 30. Based on the calculated aspect ratio, the correction unit 5202 performs keystone correction on the captured image so that the captured image becomes a front-facing image.

As a result, the correction unit 5202 can obtain a captured image that has become a front-facing image. The correction unit 5202 outputs the captured image after the front-facing correction to the determination unit 5204.

The correction unit 5202 may also perform distortion correction for each meter image obtained by cutting out the area of each meter from the captured image in the analysis process described below. For example, the correction unit 5202 applies a correction matrix for distortion correction prepared in advance by the user to the meter image. The distortion of the meter image is corrected by the distortion correction. The correction unit 5202 outputs the meter image after distortion correction to the image analysis unit 5208. The image analysis unit 5208 described below can improve the analysis accuracy in the analysis process by performing analysis processing using the meter image for which the distortion correction has been performed by the correction unit 5202.

The correction unit 5202 may also perform enlargement processing for each meter image in the analysis processing described below. As an example, the correction unit 5202 determines whether the number of pixels in the meter image is equal to or greater than a predetermined threshold, and performs enlargement processing on each meter image depending on the determination result. As an example of enlargement processing, when the size of the meter image is smaller than an image of 300 pixels square, that is, when the predetermined threshold is not equal to or greater than 90,000 pixels (300 pixels x 300 pixels), the correction unit 5202 performs enlargement processing on the meter image. Note that the value of the predetermined threshold is not limited to this example, and any value may be set.

The correction unit 5202 may also perform bilinear interpolation for each meter image in the analysis process described below. For example, the correction unit 5202 performs bilinear interpolation on the meter image after the enlargement process. By using bilinear interpolation, the correction unit 5202 can obtain a meter image that is more smoothly expressed than before the bilinear interpolation. The correction unit 5202 outputs the meter image after bilinear interpolation to the image analysis unit 5208. The image analysis unit 5208 described below can improve the analysis accuracy in the analysis process by performing analysis using the meter image on which the correction unit 5202 has performed bilinear interpolation.

The determination unit 5204 determines the type of the meter 3. For example, the determination unit 5204 determines the type of the meter 3 for each meter image based on the meter image and the meter information. The meter information includes, for example, information indicating the characteristics of each type of meter 3. As an example, the determination unit 5204 determines the type of the meter 3 by comparing the characteristics of each indicator obtained by performing image recognition processing for each meter image with the meter information. The determination unit 5204 outputs the determination result to the image analysis unit 5208.

The image processing unit 5206 performs image processing according to various processes. For example, the image processing unit 5206 generates a meter image used in the correction process of the correction unit 5202, the determination process of the determination unit 5204, and the analysis process of the image analysis unit 5208. Specifically, the image processing unit 5206 cuts out the area of each meter 3 from the captured image and generates a meter image for each meter 3 that appears in the captured image. The image processing unit 5206 outputs the generated meter image to the correction unit 5202.

The image processing unit 5206 also performs image processing related to the analysis processing of the image analysis unit 5208. Details of the image processing related to the analysis processing will be described later.

Based on the determination result, the image analysis unit 5208 performs an analysis process on the meter image according to the type of meter 3, and obtains the indication value of each of the multiple meters 3. Details of the analysis process for each type of meter 3 will be described later. The image analysis unit 5208 outputs the obtained indication value to the output processing unit 5210.

The output processing unit 5210 performs output processing related to information obtained by processing in the image analysis terminal 5. For example, the output processing unit 5210 outputs the indication value obtained by the image analysis unit 5208 through the analysis processing to the storage unit 530 and stores it.

The output processing unit 5210 also outputs the captured image received by the communication unit 510 from the imaging control terminal 4 to the storage unit 530 for storage. The captured image is stored in the storage unit 530 as an archive, for example. The output processing unit 5210 may also output the captured image to the output unit 540 for display.

The storage unit 530 is configured by a storage medium, for example, a hard disk drive (HDD), a flash memory, an electrically erasable programmable read only memory (EEPROM), a random access read/write memory (RAM), a read only memory (ROM), or any combination of these storage media. The storage unit 530 may be, for example, a non-volatile memory.
The storage unit 530 stores various types of information, such as an instruction value or a captured image input from the output processing unit 5210.

The output unit 540 is, for example, a display device such as a display.
The output unit 540 outputs various types of information, such as the captured image input from the output processing unit 5210.

<4. Processing flow>
An example of the functional configuration of the image analysis terminal 5 according to the present embodiment has been described above. Next, an example of the flow of processing in the image analysis terminal 5 according to the present embodiment will be described with reference to Fig. 4. Fig. 4 is a flowchart showing an example of the flow of processing in the image analysis terminal 5 according to the present embodiment.

As shown in FIG. 4, first, the image analysis terminal 5 receives a captured image from the imaging control terminal 4 (S102).

The image analysis terminal 5 detects the alignment markers 30 from the received captured image (S104). The image analysis terminal 5 performs orientation correction on the captured image based on the alignment markers 30 (S106).

The image analysis terminal 5 cuts out the area of each meter 3 from the captured image after orientation correction, and generates a meter image for each meter 3 (S108).

The image analysis terminal 5 determines the type of meter 3 shown in the meter image based on the generated meter image and meter information (S110).

The image analysis terminal 5 performs an analysis process on the meter image according to the determined type of meter 3, and obtains the indicated value of each meter 3 (S112). Details of the analysis process will be described later.

The image analysis terminal 5 performs an output process related to the acquired instruction value (S114). For example, the image analysis terminal 5 outputs the received captured image and the acquired instruction value to the storage unit 530 and stores them.

5. Specific Examples
An example of the flow of processing in the image analysis terminal 5 according to the present embodiment has been described above. Next, a specific example according to the present embodiment will be described with reference to Figs. 5 to 18. As described above, the image analysis terminal 5 performs analysis processing according to the type of the meter 3 and acquires the indicated value of the meter 3. A specific example of analysis processing according to the type of the meter 3 will be described in detail below.

(1) First Specific Example In the first specific example, an example of the analysis process when the meter 3 is a seven-segment meter will be described with reference to Fig. 5 and Fig. 6. First, when the meter 3 is a seven-segment meter, the details of the processes performed in the analysis process by the correction unit 5202, the image processing unit 5206, and the image analysis unit 5208 will be described.

(Processing of correction unit 5202)
The correction unit 5202 performs distortion correction on, for example, a seven-segment meter image. The correction unit 5202 also performs bilinear interpolation on the seven-segment meter image after the distortion correction.

Note that 7-segment displays include types in which each character of the indicated value is displayed in italics, and types in which the indicated value includes a decimal point. The correction unit 5202 performs correction on the meter image so that the image analysis unit 5208 described below can recognize the indicated value displayed in italics or the indicated value including a decimal point. For example, the correction unit 5202 corrects the distortion of the meter image so that when each character constituting the indicated value is surrounded by a circumscribing rectangle, only one character is included in one circumscribing rectangle.

(Processing of image processing unit 5206)
The image processing unit 5206 performs, for example, masking based on the HSV of the seven-segment meter image (hereinafter also referred to as "HSV masking") on the seven-segment meter image after bilinear interpolation. The image processing unit 5206 generates a binarized image in which the seven-segment meter image is binarized by HSV masking the seven-segment meter image. The image processing unit 5206 performs HSV masking on the seven-segment meter image so that, for example, the color of pixels of a segment indicating an indicated value becomes white, and the color of pixels other than the pixels of the segment becomes black.

In binarization, which converts an image into black and white, the pixel values of the pixels in the image are converted to either a value of 0 or 1. For example, in binarization of an image, if 0 is a pixel value that indicates white, then 1 is a pixel value that indicates black. On the other hand, if 1 is a pixel value that indicates white, then 0 is a pixel value that indicates black. In this embodiment, in binarization of an image, 1 is a pixel value that indicates white, and 0 is a pixel value that indicates black.

The H in HSV stands for Hue, S for Saturation, and V for Value. With HSV masking, the pixel value of each pixel in a seven-segment meter image is converted to either 1 (white) or 0 (black).

This causes the seven-segment meter image to be binarized. In this embodiment, the image processing unit 5206 converts the pixel values of pixels showing the indicated value in the seven-segment meter image to 1 (white), and converts the pixel values of other pixels to 0 (black).

Here, an example of HSV masking for a 7-segment meter image will be described with reference to FIG. 5. FIG. 5 is a diagram showing an example of HSV masking for a 7-segment meter image according to this embodiment.

The left image in FIG. 5 is a seven-segment meter image 40. The seven-segment meter image 40 shows "1523" as the indicated value. The image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the seven-segment meter image 40.

As a result, the image processing unit 5206 can generate a binary image 41 in which the seven-segment meter image 40 is binarized, as shown in the right diagram of Figure 5. In the binary image 41, the color of pixels that indicated the indicated value in the seven-segment meter image 40 is shown in white, and the color of pixels that indicated values other than the indicated value is shown in black.

(Processing of image analysis unit 5208)
The image analysis unit 5208 acquires an indication value indicated by the seven-segment display by performing character recognition processing on the binary image, for example. The binary image is an image generated by binarizing a seven-segment meter image by the image processing unit 5206. As an example of character recognition processing, the image analysis unit 5208 performs SSOCR (Seven Segment Optical Character Recognition) on the binary image.

The image analysis unit 5208 calls SSOCR with default parameters, for example. SSOCR called with default parameters detects black pixels in the binary image as pixels that make up the background (masked area), and detects white pixels in the binary image as pixels that make up the indication value. SSOCR also detects the number of digits in the indication value.

If the SSOCR does not obtain a recognition result that indicates that black pixels are background pixels, the image analysis unit 5208 changes the parameter settings and calls SSOCR again. For example, the image analysis unit 5208 changes the parameters so that SSOCR detects white pixels as background pixels and black pixels as pixels of the specified value.

In addition, in SSOCR, misrecognition may occur due to variations in notation on the seven-segment display. For example, in SSOCR, the number "6" may be mistakenly recognized as the alphabet "b." In this case, the image analysis unit 5208 replaces the incorrectly recognized alphabet "b" with the number "6," which indicates the correct recognition result.

The image analysis unit 5208 also converts the recognition result by SSOCR into a numerical value and acquires the instruction value as a numerical value. The image analysis unit 5208 outputs the acquired instruction value to the output processing unit 5210. Note that if the image analysis unit 5208 is unable to convert the recognition result by SSOCR into a numerical value, it outputs the recognition result by SSOCR to the output processing unit 5210 without converting it.

(Flow of analysis process)
Next, an example of the flow of the analysis process for a seven-segment meter will be described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of the flow of the analysis process for a seven-segment meter according to this embodiment.

As shown in FIG. 6, first, the correction unit 5202 of the image analysis terminal 5 performs distortion correction on the 7-segment meter image (S202).

Next, the correction unit 5202 performs bilinear interpolation on the distortion-corrected seven-segment meter image (S204).

Next, the image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the 7-segment meter image after bilinear interpolation (S206).

Next, the image analysis unit 5208 of the image analysis terminal 5 performs SSOCR on the binarized image (S208).

Finally, the image analysis unit 5208 of the image analysis terminal 5 obtains the numerical value converted from the recognition result by SSOCR as the instruction value (S210).

(2) Second Specific Example In the second specific example, an example in which the meter 3 is a pointer meter will be described with reference to Figs. 7 to 15. Examples of types of pointer meters include a circular meter and an edgewise meter. Below, an example of analysis processing for each of the circular meter and the edgewise meter will be described.

7 to 11, an example of the analysis process when the meter 3 is a circular meter will be described. First, when the meter 3 is a circular meter, the details of the processes performed by the correction unit 5202, the image processing unit 5206, and the image analysis unit 5208 in the analysis process will be described.

(Processing of correction unit 5202)
The correction unit 5202 performs distortion correction on, for example, a meter image of a circular meter. The correction unit 5202 performs distortion correction so that, for example, a preset reference position and the position of the central axis of the needle of the circular meter coincide with each other.

The correction unit 5202 also performs bilinear interpolation on the meter image of the circular meter after distortion correction.

(Processing of image processing unit 5206)
The image processing unit 5206 performs HSV masking on the meter image of the circular meter after bilinear interpolation, for example. The image processing unit 5206 generates a binary image (first binary image) in which the meter image of the circular meter is binarized by HSV masking the meter image of the circular meter. The image processing unit 5206 performs HSV masking on the meter image of the circular meter so that the color of the pixels of the needle of the circular meter indicating the indicated value is white, and the color of the pixels other than the pixels of the needle is black.

Here, an example of HSV masking for a meter image of a circular meter will be described with reference to FIG. 7. FIG. 7 is a diagram showing an example of HSV masking for a meter image of a circular meter according to this embodiment.

The left image in Figure 7 is a meter image 50 of a circular meter. The circular meter shown in the left image in Figure 7 is a meter in which a pointer 52 rotates around an axis 53. A scale indicating the range of indicated values is printed on the dial 54 of the circular meter. Of the two ends of the pointer 52 of the circular meter, the end with the thinner pointer 52 is the tip side, and the end with the thicker pointer 52 is the end side. In the circular meter shown in the left image in Figure 7, the value of the scale indicated by the tip side of the pointer 52 is the indicated value. For example, the indicated value in the left image in Figure 7 is 0.7.

The image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the meter image 50. As a result, the image processing unit 5206 can generate a binary image 51 in which the meter image 50 is binarized, as shown in the right diagram of FIG. 7. In the binary image 51, the color of the pixels of the needle 52 that indicated the indicated value in the meter image 50 is shown in white, and the color of the pixels that indicated anything other than the needle 52 is shown in black. Hereinafter, in the binary image of the meter image, the area shown by white pixels is also referred to as a "white pixel area".

The image processing unit 5206 also performs noise removal on the binary image of the meter image of the circular meter. For example, the image processing unit 5206 performs noise removal based on the white pixel areas in the binary image of the meter image.

For example, the image processing unit 5206 removes white pixels outside a predetermined area as noise. The predetermined area is, for example, a circular area centered on the axis of the circular meter. The size of the circular area is, for example, 70% of the size of the circular meter. In other words, the radius of the circular area is 70% of the radius of the circular meter.

Specifically, the image processing unit 5206 calculates the position of the center of gravity of the white pixel region, and determines whether or not to perform noise removal based on whether or not the position of the center of gravity is included in a predetermined region. If the position of the center of gravity is included in the predetermined region, the image processing unit 5206 determines not to perform noise removal of the binarized image of the meter image. This is because the position of the center of gravity of the white pixel region can be considered to indicate the pointer.
On the other hand, if the center position is not included in the predetermined region, the image processing unit 5206 determines to perform noise removal on the binarized image of the meter image.

Here, an example of noise removal from a binary image of a meter image of a circular meter will be described with reference to FIG. 8. FIG. 8 is a diagram showing an example of noise removal from a binary image of a meter image of a circular meter according to this embodiment.

In the binary image 51 of the meter image shown in FIG. 8, the circular area indicated by the dashed line indicates the circular meter area 54. Also, the circular area indicated by the dashed line indicates the specified area 55.

The image processing unit 5206 first detects the contours of white pixels from the binarized image 51 to detect white pixel regions. In the example shown in FIG. 8, for example, the contours of white pixel region 56 and white pixel region 57 are detected. The image processing unit 5206 calculates the positions of the centers of gravity of the detected white pixel regions 56 and 57. The image processing unit 5206 then performs noise removal based on whether the calculated center positions are included in the specified region 55.

The image processing unit 5206 also generates a histogram of angle information based on the binary image of the meter image of the circular meter. The angle information is, for example, information indicating the angle of a white pixel in the binary image. Specifically, the angle of a white pixel is the angle between a line connecting a reference point and the position of the axis of the circular meter and a line connecting a target white pixel and the position of the axis of the circular meter. The image processing unit 5206 calculates the angle for each white pixel for all white pixels in the binary image of the meter image, and generates a histogram relating to the distribution of the angles.

Now, with reference to Figure 9, we will explain an example of histogramming of angle information based on a binary image of a meter image of a circular meter. Figure 9 is a diagram showing an example of histogramming of angle information based on a binary image of a meter image of a circular meter according to this embodiment.

In the binary image 51 in the upper diagram of Figure 9, straight line L1 is a line that passes through reference point P1 and the position of axis 53 of the circular meter. Straight line L2 is a line that passes through point P2, which indicates the position of the target white pixel, and the position of axis 53 of the circular meter. Angle θ is the angle made by lines L1 and L2. This angle θ is the angle of the target white pixel. The image processing unit 5206 calculates the angle θ of all white pixels in the binary image 51. After the calculation, the image processing unit 5206 generates a histogram 58 that shows the distribution of all the calculated angles θ.

Histogram 58 shown in the lower diagram of FIG. 9 shows the distribution of angle θ calculated by image processing unit 5206. The horizontal axis of histogram 58 shows the angle, and the vertical axis shows the number of white pixels. Note that 0 degrees in histogram 58 shows the position of line L1 in binarized image 51 in the upper diagram of FIG. 9.

The image processing unit 5206 also binarizes a histogram relating to the distribution of pixels indicating the pointer in the binary image of the meter image of the circular meter, and generates a binary image of the histogram (second binary image). For example, the image processing unit 5206 normalizes the number of white pixels present at each angle indicated by the histogram of the angle information.

Specifically, the image processing unit 5206 normalizes the range of the number of white pixels that can exist at each angle indicated by the histogram of the angle information so that it is represented by pixel values ranging from 0 to 255. This allows the image processing unit 5206 to grayscale the histogram of the angle information and generate a grayscale image of the histogram.

The image processing unit 5206 binarizes the grayscale image of the histogram of the angle information based on a predetermined threshold. As an example, the image processing unit 5206 determines whether or not a pixel value in the grayscale image of the histogram of the angle information is equal to or greater than a predetermined threshold, and converts the pixel value of each pixel according to the determination result.

For example, in the grayscale image of the histogram of angle information, if the pixel value of a pixel is equal to or greater than a predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 1 (white). On the other hand, in the grayscale image of the histogram of angle information, if the pixel value of a pixel is not equal to or greater than the predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 0 (black).

This allows the image processing unit 5206 to generate a binary image of the histogram of angle information. When generating the binary image of the histogram of angle information, the image processing unit 5206 performs binarization based on a predetermined threshold value, thereby reducing noise in the binary image. Note that any value may be set as the predetermined threshold value.

Here, an example of binarization of a histogram of angle information will be described with reference to FIG. 10. FIG. 10 is a diagram showing an example of binarization of a histogram of angle information according to this embodiment.

The image processing unit 5206 grayscales the histogram 58 shown in FIG. 9 to generate the grayscale image 59 shown in the upper diagram of FIG. 10. The image processing unit 5206 further binarizes the grayscale image 59 to generate the binary image 60 shown in the lower diagram of FIG. 10. Note that the horizontal axis of the grayscale image 59 and the binary image 60 indicates the angle θ.

(Processing of image analysis unit 5208)
The image analysis unit 5208 acquires the indication value by analyzing the binary image of the histogram. For example, the image analysis unit 5208 identifies the position of the pointer where the pointer is pointing to the indication value based on the binary image of the histogram, and acquires the indication value at the pointer position.

Specifically, the image analysis unit 5208 determines the position of the tip side of the needle that is pointing to the indicated value and the position of the end side of the needle that is not pointing to the indicated value based on the binary image of the histogram. Then, the image analysis unit 5208 identifies the position of the tip side of the needle as the needle position.

When identifying the position of the tip side of the needle as the needle position, the image analysis unit 5208 first detects multiple areas (peaks) indicated by white pixels from the binary image of the histogram. Next, the image analysis unit 5208 searches for a combination of peaks from the multiple detected peaks that results in an angular relationship of 180 degrees. Here, an example of identifying the position of the tip side of the needle will be described with reference to FIG. 10 again.

The image analysis unit 5208 searches for peaks of white pixels in the histogram 60 shown in FIG. 10. In the example shown in FIG. 10, three peaks, peak 61, peak 62, and peak 63, are detected.

The image analysis unit 5208 searches for a combination of peaks with an angular relationship of 180 degrees from the detected multiple peaks. For example, assume that there is a first peak and a second peak that indicates an angle larger than the angle indicated by the first peak among the multiple peaks. Assume that, for the first and second peaks, the difference between the maximum angle indicated by the second peak and the minimum angle indicated by the first peak is greater than 180 degrees, and the difference between the minimum angle indicated by the second peak and the maximum angle indicated by the first peak is less than 180 degrees. In this case, the image analysis unit 5208 regards the combination of the first and second peaks as a combination of peaks with an angular relationship of 180 degrees. In the case of the example shown in FIG. 10, the image analysis unit 5208 regards the combination of peak 61 (first peak) and peak 63 (second peak) as a combination of peaks with an angular relationship of 180 degrees.

The image analysis unit 5208 determines whether each peak of the combination of peaks with an angular relationship of 180 degrees is within the meter angle range. The meter angle range is the range within which the tip of the pointer points to the indicated value within the range of angles within which the pointer can rotate.

If one of the peaks is not within the meter angle range, the image analysis unit 5208 considers the peak within the meter angle range to be the peak on the tip side of the needle. For example, in the example shown in FIG. 10, if the meter angle range is -100 degrees to 100 degrees, the image analysis unit 5208 considers peak 63 to be the peak on the tip side of the needle.

When both peaks are within the meter angle range, the image analysis unit 5208 considers the peak with the narrower peak width to be the peak on the tip side of the needle. For example, in the example shown in FIG. 10, assume that both peaks 61 and 63 are within the meter angle range. Comparing the peak widths of each peak, it can be seen that the peak width of peak 63 is narrower than the peak width of peak 61. Therefore, the image analysis unit 5208 considers peak 63 to be the peak on the tip side of the needle. On the other hand, the image analysis unit 5208 considers peak 61 to be the peak on the end side of the needle.

Let us say that the image analysis unit 5208 is unable to identify the peak at the tip of the needle by analyzing the binary image of the histogram (binary image 60). In this case, the image analysis unit 5208 may identify the peak at the tip of the needle based on the histogram of the grayscaled angle information (grayscale image 59). For example, the image analysis unit 5208 considers the peak with the maximum brightness value in the histogram of the grayscaled angle information to be the peak at the tip of the needle.

The image analysis unit 5208 identifies the angle of the pointer based on the identified peak at the tip of the pointer. For example, the image analysis unit 5208 identifies the median value of the peak as the angle of the pointer. The position indicated by the identified angle of the pointer is the pointer position of the pointer.

The image analysis unit 5208 acquires the indication value based on the identified angle of the needle. For example, the image analysis unit 5208 calculates the indication value by the following formula (1). Note that the angle of the needle is θ, the minimum angle of the meter angle range is θ _min , the maximum angle is θ _max , the indication value is x ₁ , the minimum value of the meter scale is x _min , and the maximum value is x _max .
_x1 = _xmin +(θ- _θmin )×( _xmax - _xmin )/( _θmax - _θmin ) (1)

The image analysis unit 5208 may round the calculated instruction value to a predetermined digit precision. For example, the image analysis unit 5208 rounds the calculated instruction value to an integer.

(Flow of analysis process)
Next, an example of the flow of the analysis process for a circular meter will be described with reference to Fig. 11. Fig. 11 is a flowchart showing an example of the flow of the analysis process for a circular meter according to this embodiment.

As shown in FIG. 11, first, the correction unit 5202 of the image analysis terminal 5 performs distortion correction on the meter image of the circular meter (S302).

Next, the correction unit 5202 performs bilinear interpolation on the meter image after distortion correction (S304).

Next, the image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the meter image after bilinear interpolation (S306). As a result, the image processing unit 5206 generates a binary image of the meter image of the circular meter.

Next, the image processing unit 5206 performs noise removal on the binary image of the circular meter image (S308).

Next, the image processing unit 5206 creates a histogram of the angle information (S310). As a result, the image processing unit 5206 generates a histogram that shows the distribution of the angle information.

Next, the image processing unit 5206 binarizes the generated histogram of angle information (S312). As a result, the image processing unit 5206 generates a binarized image of the histogram of angle information.

Next, the image analysis unit 5208 analyzes the binarized image of the histogram of angle information (S314). As a result, the image analysis unit 5208 identifies the needle position (angle of the needle).

Finally, the image analysis unit 5208 obtains the indication value based on the identified pointer position (S316).

(2-2) Edgewise Meter An example of the analysis process when the meter 3 is an edgewise meter will be described with reference to Fig. 12 to Fig. 15. First, when the meter 3 is an edgewise meter, the details of the processes performed in the analysis process by the correction unit 5202, the image processing unit 5206, and the image analysis unit 5208 will be described.

(Processing of correction unit 5202)
The correction unit 5202 performs distortion correction on, for example, the meter image of the edgewise meter. The correction unit 5202 also performs bilinear interpolation on the meter image of the edgewise meter after the distortion correction.

(Processing of image processing unit 5206)
The image processing unit 5206 performs, for example, HSV masking on the meter image of the edgewise meter after bilinear interpolation. The image processing unit 5206 generates a binary image (first binary image) in which the meter image of the edgewise meter is binarized by HSV masking the meter image of the edgewise meter. The image processing unit 5206 performs HSV masking on the meter image of the edgewise meter so that, for example, the color of the pixels of the needle of the edgewise meter indicating the indicated value is white and the color of the pixels other than the pixels of the needle is black.

Here, an example of HSV masking for a meter image of an edgewise meter will be described with reference to FIG. 12. FIG. 12 is a diagram showing an example of HSV masking for a meter image of an edgewise meter according to this embodiment.

The left diagram in FIG. 12 is a meter image 70 of an edgewise meter. The edgewise meter shown in the left diagram in FIG. 12 is a meter that indicates an indicated value by moving a pointer 72 in either a first direction or a second direction. A scale indicating the range of indicated values is printed on a dial 73 of the edgewise meter. The pointer 72 moves, for example, between 0 and 200 on the scale in either the direction of 0 (first direction) or the direction of 200 (second direction). In the edgewise meter shown in the left diagram in FIG. 12, the value of the scale at the position of the pointer 72 is the indicated value.

The image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the meter image 70. As a result, the image processing unit 5206 can generate a binary image 71 in which the meter image 70 is binarized, as shown in the right diagram of FIG. 12. In the binary image 71, the color of the pixels of the pointer 72 that indicated the indicated value in the meter image 70 is shown in white, and the color of the pixels that indicated anything other than the pointer 72 is shown in black.

The image processing unit 5206 also generates a histogram of position information based on the binary image of the meter image of the edgewise meter. The position information is, for example, information indicating the position of a white pixel in the binary image. The position of the white pixel is, for example, any position in the movement direction of the pointer 72. The specific value of the position of the white pixel is indicated, for example, by the value of the scale in the movement direction of the pointer 72. Thus, the position of the white pixel is indicated by a value in the range from the minimum value to the maximum value of the scale.

The image processing unit 5206 acquires the position of each white pixel for all white pixels in the binary image of the meter image, and generates a histogram relating to the distribution of the positions. The image processing unit 5206 acquires the positions of white pixels in the binary image of the meter image, for example, based on the meter information of the edgewise meter. Specifically, the image processing unit 5206 identifies at which position on the scale the white pixels in the binary image of the meter image are located, based on information such as the size of the indicator, the size of the needle, and the interval of the scale, contained in the meter information of the edgewise meter.

Now, with reference to FIG. 13, an example of histogramming of position information based on a binary image of the meter image of an edgewise meter will be described. FIG. 13 is a diagram showing an example of histogramming of position information based on a binary image of the meter image of an edgewise meter according to this embodiment.

First, the image processing unit 2506 identifies the positions of white pixels in the binary image 71 in the upper diagram of FIG. 13 based on the meter information. After identification, the image processing unit 5206 generates a histogram 74 that shows the distribution of all the identified positions.

The histogram 74 shown in the lower diagram of FIG. 13 indicates the distribution of positions identified by the image processing unit 5206. The horizontal axis of the histogram 74 indicates the positions of white pixels, and the vertical axis indicates the number of white pixels.

The image processing unit 5206 also binarizes a histogram relating to the distribution of pixels indicating the pointer in the binary image of the meter image of the edgewise meter, and generates a binary image of the histogram (second binary image). For example, the image processing unit 5206 normalizes the number of white pixels present at each position indicated by the histogram of position information.

Specifically, the image processing unit 5206 normalizes the range of the number of white pixels that can exist at each position indicated by the histogram of the position information so that it is represented by pixel values ranging from 0 to 255. This allows the image processing unit 5206 to grayscale the histogram of the position information and generate a grayscale image of the histogram.

The image processing unit 5206 binarizes the grayscale image of the histogram of the position information based on a predetermined threshold. As an example, the image processing unit 5206 determines whether or not a pixel value in the grayscale image of the histogram of the position information is equal to or greater than a predetermined threshold, and converts the pixel value of each pixel according to the determination result.

For example, in the grayscale image of the histogram of position information, if the pixel value of a pixel is equal to or greater than a predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 1 (white). On the other hand, in the grayscale image of the histogram of position information, if the pixel value of a pixel is not equal to or greater than the predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 0 (black).

This allows the image processing unit 5206 to generate a binary image of the histogram of the position information. When generating the binary image of the histogram of the position information, the image processing unit 5206 performs binarization based on a predetermined threshold value, thereby reducing noise in the binary image. Note that any value may be set as the predetermined threshold value.

Here, an example of binarization of a histogram of location information will be described with reference to FIG. 14. FIG. 14 is a diagram showing an example of binarization of a histogram of location information according to this embodiment.

The image processing unit 5206 generates the grayscale image 75 shown in the upper diagram of FIG. 14 by grayscaling the histogram 74 shown in FIG. 13. The image processing unit 5206 further binarizes the grayscale image 75 to generate the binary image 76 shown in the lower diagram of FIG. 13. Note that the horizontal axis of the grayscale image 75 and the binary image 76 indicates the position.

(Processing of image analysis unit 5208)
The image analysis unit 5208 acquires the indication value by analyzing the binary image of the histogram. For example, the image analysis unit 5208 identifies the position of the pointer where the pointer is pointing to the indication value based on the binary image of the histogram, and acquires the indication value at the pointer position.

When identifying the needle position, the image analysis unit 5208 first detects an area (peak) indicated by white pixels from the binary image of the histogram. For example, the image analysis unit 5208 detects peak 77 shown in FIG. 14. Note that if the binary image of the histogram contains noise, multiple peaks may be detected. In that case, the image analysis unit 5208 selects, from among the multiple peaks, the peak with the widest peak width in the axial direction of the binary image of the histogram.

The image analysis unit 5208 identifies the position indicated by the detected or selected peak as the needle position of the needle. The image analysis unit 5208 then obtains the median value of the peaks at the identified needle position as the indication value.

The image analysis unit 5208 may round the calculated instruction value to a predetermined digit precision. For example, the image analysis unit 5208 rounds the calculated instruction value to an integer.

The image analysis unit 5208 may apply an offset value to the acquired indication value. Due to the angle of view of the imaging device and the effects of processing, a difference may occur between the indication value calculated by the image analysis unit 5208 and the indication value recognized by the user by visually observing the meter 3.

Therefore, when a difference of a predetermined threshold or more occurs between the indication value calculated by the image analysis unit 5208 and the indication value visually recognized by the user, the image analysis unit 5208 adds or subtracts an offset value to the calculated indication value. This allows the image analysis unit 5208 to obtain an indication value closer to the indication value visually recognized by the user. Note that the offset value to be added or subtracted from the indication value and the predetermined threshold are not particularly limited.

(Flow of analysis process)
Next, an example of the flow of analysis processing for an edgewise meter will be described with reference to Fig. 15. Fig. 15 is a flowchart showing an example of the flow of analysis processing for an edgewise meter according to this embodiment.

As shown in FIG. 15, first, the correction unit 5202 of the image analysis terminal 5 performs distortion correction on the meter image of the edgewise meter (S402).

Next, the correction unit 5202 performs bilinear interpolation on the meter image after distortion correction (S404).

Next, the image processing unit 5206 of the image analysis terminal 5 performs HSV masking on the meter image after bilinear interpolation (S406). As a result, the image processing unit 5206 generates a binary image of the meter image of the edgewise meter.

Next, the image processing unit 5206 creates a histogram of the position information (S408). As a result, the image processing unit 5206 generates a histogram that shows the distribution of the position information.

Next, the image processing unit 5206 binarizes the generated histogram of the position information (S410). As a result, the image processing unit 5206 generates a binarized image of the histogram of the position information.

Next, the image analysis unit 5208 analyzes the binarized image of the histogram of the position information (S412). As a result, the image analysis unit 5208 identifies the needle position of the needle.

Finally, the image analysis unit 5208 obtains the indication value based on the identified pointer position (S414).

(3) Third Specific Example In the third specific example, an example of the analysis process when the meter 3 is a float meter will be described with reference to Fig. 16 to Fig. 18. First, when the meter 3 is a float meter, the details of the processes performed in the analysis process by the correction unit 5202, the image processing unit 5206, and the image analysis unit 5208 will be described.

(Processing of correction unit 5202)
The correction unit 5202 performs distortion correction on, for example, the meter image of a float meter. The correction unit 5202 also performs bilinear interpolation on the meter image of the float meter after the distortion correction.

(Processing of image processing unit 5206)
The image processing unit 5206 performs binarization based on the luminance value on the float meter meter image after bilinear interpolation, for example. The image processing unit 5206 generates a binarized image (first binarized image) in which the float meter meter image has been binarized by binarizing the float meter meter image. The image processing unit 5206 binarizes the float meter meter image based on a predetermined threshold, for example.

As an example, the image processing unit 5206 determines whether the luminance value of each pixel of the float meter meter image is equal to or greater than a predetermined threshold, and converts the pixel value of each pixel depending on the determination result. For example, if the luminance value of a pixel is equal to or greater than the predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 1 (white). On the other hand, if the luminance value of a pixel is not equal to or greater than the predetermined threshold, the image processing unit 5206 converts the pixel value of the pixel to 0 (black). In this way, the image processing unit 5206 can generate a binary image of the float meter meter image. Note that any value may be set as the predetermined threshold.

The image analysis unit 5208, which will be described later, identifies the position of the float based on the area of pixels that indicate the float in the binarized image, and obtains the indication value at the position of the float. For this reason, it is desirable that the area of pixels that indicate the float is shown with continuity in the axial direction in the binarized image. Therefore, the image processing unit 5206 performs Gaussian blurring with a (5, 7) kernel on the meter image 80 before binarization. This further emphasizes the continuity in the axial direction of the area of pixels that indicate the float.

Here, an example of binarization of a meter image of a float meter will be described with reference to FIG. 16. FIG. 16 is a diagram showing an example of binarization of a meter image of a float meter according to this embodiment.

The left image in Fig. 16 is a meter image 80 of a float meter. The float meter shown in the left image in Fig. 16 is a meter in which a float 82 moves in the direction of an axis 83. In a float meter, an indicated value is shown based on the shape and position of the float 82. For example, the indicated value is the scale value at the position of the maximum diameter part of the float 82.

The image processing unit 5206 of the image analysis terminal 5 performs binarization based on the luminance value on the meter image 80. As a result, the image processing unit 5206 can generate a binarized image 81 in which the meter image 80 is binarized, as shown in the right diagram of FIG. 16.

The image processing unit 5206 also generates an image (second binary image) that indicates the feature values of the binary image of the float meter meter image. The image that indicates the feature values of the binary image of the float meter meter image is also referred to as a "feature value image" below. For example, the image processing unit 5206 generates a binary image as the feature value image by adding a predetermined feature value to the pixel value of a predetermined pixel in the binary image of the meter image.

Specifically, the image processing unit 5206 generates a feature value image based on the pixel values of pixels indicating the axis in the binary image of the meter image and the pixel values of pixels on a predetermined line parallel to the axis in the binary image of the meter image. Here, the pixel values of the pixels indicating the axis in the binary image of the meter image are the pixel values of the predetermined pixels. Also, the pixel values of the pixels on a predetermined line parallel to the axis in the binary image of the meter image are the predetermined feature values.

Now, an example of generating a feature value image will be described with reference to FIG. 17. FIG. 17 is a diagram showing an example of generating a feature value image according to this embodiment. The left diagram in FIG. 17 is a binarized image 81 of a float meter. The right diagram in FIG. 17 is a feature value image 84 in which feature values have been added to the binarized image 81. Note that the vertical sizes of the binarized image 81 and the feature value image 84 in FIG. 17 are the same.

The image processing unit 5206 generates a feature value image based on the pixel values of pixels indicating the axis in the binary image 81 and the pixel values of pixels on a predetermined line parallel to the axis in the binary image 81. The straight line L3 shown in the left diagram of FIG. 17 is a line drawn so as to pass through at least the pixels indicating the axis of the float. The straight lines L4 and L5 shown in the left diagram of FIG. 17 are lines drawn parallel to the straight line L3.

The image processing unit 5206 adds, to the pixel value of a pixel on line L3, the pixel values of pixels on lines L4 and L5 that are located at positions corresponding to the pixel on line L3, as feature values. For example, the image processing unit 5206 adds, to the pixel value of a pixel on line L3 at a position 200 in the vertical direction of the binarized image 81, the pixel value of a pixel on line L4 at a position 200 in the vertical direction of the binarized image 81 and the pixel value of a pixel on line L5 at a position 200 in the vertical direction of the binarized image 81, as feature values.

When adding feature values, the image processing unit 5206 adds the bit-inverted pixel values of the pixels on line L4 and line L5 to the pixel value of the pixel on line L3. Note that the image processing unit 5206 adds feature values by AND operation.

The image processing unit 5206 adds the feature values for each pixel on the line L3. The image processing unit 5206 generates the feature value image 84 shown in the right diagram of FIG. 17 based on the pixel values of the pixels after addition. For example, the image processing unit 5206 generates the feature value image 84 so that the color of the pixel at each vertical position in the feature value image 84 corresponds to the color indicated by the pixel value resulting from the addition of the feature values at each vertical position in the binarized image 81.

(Processing of image analysis unit 5208)
The image analysis unit 5208 acquires an indication value by analyzing the feature value image. For example, the image analysis unit 5208 identifies the position of the float based on the feature value image and acquires an indication value at the position of the float.

Specifically, the image analysis unit 5208 first detects from the feature value image an area composed of pixels having the same pixel values as pixels indicating a float in the binary image of the meter image. The image analysis unit 5208 identifies, from among the detected areas, an area having a size corresponding to a predetermined float size as an area indicating a float.

The predetermined float size is the size of the float in the axial direction. The area of the size corresponding to the predetermined float size is, for example, the area of the size closest to the size corresponding to the predetermined float size.

For example, in the feature value image 84 of FIG. 17, assume that the size of region 85 in the axial direction corresponds to a predetermined float size. In this case, the image analysis unit 5208 identifies region 85 as a float region.

The image analysis unit 5208 identifies the position of the identified float area as the position of the float. The image analysis unit 5208 acquires an indication value based on the identified float position. For example, the image analysis unit 5208 calculates the indication value by the following formula (2). Note that the indication value is x ₂ , the value of the scale of the reference meter is x _S , the minimum value of the scale is x _min , the maximum value of the scale is x _max , the value indicating the current position of the float is y _n , the position of the scale of the reference meter is y _S , the start position of the scale is y _min , and the end position of the scale is y _max .
_x2 = _xS + ( _yn - _yS ) x ( _ymax - _ymin ) / ( _xmax - _xmin ) (2)

The image analysis unit 5208 may round the calculated instruction value to a predetermined digit precision. For example, the image analysis unit 5208 rounds the calculated instruction value to an integer.

The image analysis unit 5208 may apply an offset value to the acquired indication value. Due to the angle of view of the imaging device and the effects of processing, a difference may occur between the indication value calculated by the image analysis unit 5208 and the indication value recognized by the user by visually observing the meter 3.

Therefore, when a difference equal to or exceeds a predetermined threshold occurs between the instruction value calculated by the image analysis unit 5208 and the instruction value visually recognized by the user, the image analysis unit 5208 adds or subtracts an offset value to the calculated instruction value.

This allows the image analysis unit 5208 to obtain an instruction value that is closer to the instruction value visually recognized by the user. Note that the offset value to be added to or subtracted from the instruction value and the predetermined threshold value are not particularly limited.

(Flow of analysis process)
Next, an example of the flow of analysis processing for a float meter will be described with reference to Fig. 18. Fig. 18 is a flowchart showing an example of the flow of analysis processing for a float meter according to this embodiment.

As shown in FIG. 18, first, the correction unit 5202 of the image analysis terminal 5 performs distortion correction on the meter image of the float meter (S502).

Next, the correction unit 5202 performs bilinear interpolation on the meter image after distortion correction (S504).

Next, the image processing unit 5206 of the image analysis terminal 5 performs binarization based on the luminance value on the meter image after bilinear interpolation (S506). As a result, the image processing unit 5206 generates a binary image of the meter image of the float meter.

Next, the image processing unit 5206 generates a feature value image (S508).

Next, the image analysis unit 5208 analyzes the generated feature value image and identifies the position of the float (S510).

Finally, the image analysis unit 5208 obtains the indication value based on the identified float position (S512).

As described above, the image analysis terminal 5 according to the present embodiment binarizes a meter image in which a pointer meter is captured, and generates a binary image of the meter image. The image analysis terminal 5 also binarizes a histogram relating to the distribution of pixels indicating the pointer in the binary image of the meter image, and generates a binary image of the histogram.
The image analysis terminal 5 identifies the position of the needle where the needle is pointing to the indicated value based on the binary image of the histogram, and obtains the indicated value at the needle position.

With this configuration, the image analysis terminal 5 obtains the indicated value of the pointer meter by image analysis specific to the pointer meter. Therefore, the image analysis terminal 5 can obtain the indicated value of the pointer meter with high accuracy.

The image analysis terminal 5 according to this embodiment also binarizes the meter image obtained by capturing an image of the float meter, and generates a binary image of the meter image. The image analysis terminal 5 also generates a feature value image based on the pixel values of pixels indicating an axis in the binary image of the meter image and the pixel values of pixels on a predetermined line parallel to the axis in the binary image of the meter image.
The image analysis terminal 5 identifies the position of the float based on the feature value image, and obtains an indication value at the position of the float.

With this configuration, the image analysis terminal 5 obtains the indicated value of the float meter through image analysis specific to the float meter. Therefore, the image analysis terminal 5 can obtain the indicated value of the float meter with high accuracy.

Moreover, the image analysis terminal 5 according to the present embodiment generates a meter image by cutting out the area of each meter from a captured image in which a plurality of meters having an indication portion that indicates an indication value are captured.
The image analysis terminal 5 determines the type of the meter for each meter image based on the meter image and the meter information indicating the type of the meter.
Based on the determination result, the image analysis terminal 5 performs an analysis process on the meter image according to the type of the meter, and obtains the indicated values of each of the multiple meters.

With this configuration, the image analysis terminal 5 determines the types of multiple meters, and then obtains the indicated value of each meter through image analysis according to the type of meter. Therefore, the image analysis terminal 5 can obtain the indicated value of each meter with high accuracy according to the type of meter.

As a result, the image analysis terminal 5 according to this embodiment can obtain the indicated values of various meters with high accuracy by performing image analysis according to the type of meter.

The above describes the embodiment of the present invention. The image analysis terminal 5 in the above embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed to realize the function. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. The term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, and storage devices such as hard disks built into a computer system. The term "computer-readable recording medium" may also include a medium that dynamically holds a program for a short period of time, such as a communication line when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, and a medium that holds a program for a certain period of time, such as a volatile memory inside a computer system that is a server or client in that case. The above program may be a program for realizing part of the above-mentioned function, or may be a program that can realize the above-mentioned function in combination with a program already recorded in the computer system, or may be a program that is realized using a programmable logic device such as an FPGA (Field Programmable Gate Array).

The above describes in detail the embodiments of the present invention with reference to the drawings, but the specific configuration is not limited to the above, and various design changes can be made without departing from the spirit of the present invention.

REFERENCE SIGNS LIST 1 Image analysis system 2 Imaging device 3 Meter 4 Imaging control terminal 5 Image analysis terminal 20 Imaging area 30 Alignment marker 410 Communication unit 420 Control unit 430 Storage unit 440 Output unit 510 Communication unit 520 Control unit 530 Storage unit 540 Output unit 4202 Imaging control unit 4204 Output processing unit 5202 Correction unit 5204 Determination unit 5206 Image processing unit 5208 Image analysis unit 5210 Output processing unit NW Network

Claims (11)

an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a pointer, and generates a second binary image by binarizing a histogram relating to a distribution of pixels indicating the pointer in the first binary image;
an image analysis unit that identifies a pointer position where the pointer is pointing to the indicated value based on the second binarized image and acquires the indicated value at the pointer position;
An image analysis device comprising: When the meter is a circular meter in which the pointer is rotatably provided,
2. The image analyzing device according to claim 1, wherein the image analysis unit determines, based on the second binary image, the position of the tip side of the pointer pointing to the indicated value and the position of the end side of the pointer not pointing to the indicated value, and identifies the position of the tip side of the pointer as the pointer position. The image analysis device according to claim 2, wherein the image analysis unit calculates an angle for each pixel that indicates the pointer in the first binarized image and generates a histogram that indicates a distribution of the angles. When the meter is an edgewise meter,
The image analyzing device according to claim 1 , wherein the image analyzing section identifies a position of the needle of the edgewise meter based on the second binarized image. The image analysis device according to claim 4, wherein the image analysis unit obtains a position on the meter for each pixel that indicates the pointer in the first binarized image, and generates a histogram that indicates a distribution of the positions. an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a float movable along an axis, and generates a second binary image based on pixel values of pixels indicating the axis in the first binary image and pixel values of pixels on a predetermined line parallel to the axis in the first binary image;
an image analysis unit that identifies a position of the float based on the second binarized image and acquires the indication value at the position of the float;
An image analysis device comprising: The image analysis device according to claim 6, wherein the image analysis unit identifies, in the second binarized image, an area of a size corresponding to the float size in the direction of the predetermined axis among areas formed by pixels having the same pixel value as the pixel value of the pixel indicating the float in the first binarized image, as the area indicating the float. an image processing unit binarizes an image of a meter whose indication value is indicated by a pointer to generate a first binary image, and binarizes a histogram relating to a distribution of pixels indicating the pointer in the first binary image to generate a second binary image;
an image analysis unit specifying a pointer position where the pointer is pointing to the indicated value based on the second binarized image, and acquiring the indicated value at the pointer position;
An image analysis method comprising: Computer,
an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a pointer, and generates a second binary image by binarizing a histogram relating to a distribution of pixels indicating the pointer in the first binary image;
an image analysis unit that identifies a pointer position where the pointer is pointing to the indicated value based on the second binarized image and acquires the indicated value at the pointer position;
A program that functions as a an image processing unit generates a first binary image by binarizing a captured image of a meter whose indicated value is indicated by a float movable along an axis, and generates a second binary image based on pixel values of pixels indicating the axis in the first binary image and pixel values of pixels on a predetermined line parallel to the axis in the first binary image;
an image analysis unit identifying a position of the float based on the second binarized image and acquiring the indication value at the position of the float;
An image analysis method comprising: Computer,
an image processing unit that generates a first binary image by binarizing an image of a meter whose indicated value is indicated by a float movable along an axis, and generates a second binary image based on pixel values of pixels indicating the axis in the first binary image and pixel values of pixels on a predetermined line parallel to the axis in the first binary image;
an image analysis unit that identifies a position of the float based on the second binarized image and acquires the indication value at the position of the float;
A program that functions as a

Images