JP2017003473A - Defect detection device, defect detection method and computer program - Google Patents

Abstract

A defect detection apparatus, a defect detection method, and a computer program capable of realizing defect inspection of a wide range of subjects with a simpler configuration. According to one embodiment, a defect detection apparatus includes an information acquisition unit, a defect detection unit, and a defect depth determination unit. The information acquisition unit acquires surface temperature information indicating the surface temperature of the subject and temperature difference information indicating the temperature difference in the vicinity of the subject for a predetermined period. The defect detection unit detects the suspected part of the defect in the subject and the size of the suspected part based on the distribution of the surface temperature of the subject indicated by the surface temperature information. The defect depth determination unit is configured to detect the suspect based on the size of the suspected part detected by the defect detecting part and the temperature difference information according to the timing at which the surface temperature information from which the suspected part was detected was acquired. Determine the depth of the site. [Selection] Figure 2

Description

  Embodiments described herein relate generally to a defect detection apparatus, a defect detection method, and a computer program.

  Conventionally, a defect inspection method for determining the presence or absence of defects such as floating or peeling in structures such as tunnels and bridges based on changes in the surface temperature of a subject has been put into practical use. In such a defect inspection method, a test body generated so as to have a simulated defect is used in order to estimate the depth of the defect that the subject has. For example, the test body is produced by injecting a foaming agent into concrete, and a cavity in the concrete formed by the foaming agent is regarded as a defect. Such a test body is placed beside the examination site of the subject, and the temperature change is measured together with the subject. If the temperature change is measured in this way, the presence or absence of a defect can be estimated from the temperature change of the subject, and the depth of the defect can be estimated by combining with the temperature change of the specimen.

  Since such a defect inspection method is based on the temperature change of the subject and the test body, it is greatly affected by the temperature of the environment in which the subject and the test body are placed. Therefore, although it is good if the examination location is local, in order to examine a subject over a wider range as the temperature varies depending on the examination location, it is necessary to install test specimens at many locations of the subject. For example, tunnels have different temperatures at the entrance and the center. For this reason, the conventional defect inspection method may be difficult to implement in terms of labor and cost when inspecting a wide range of subjects such as tunnels.

JP 09-189670 A JP 2012-173280 A JP 2012-177675 A Japanese Patent No. 4113100 Japanese Patent No. 4081479 Japanese Patent No. 5070135 Japanese Patent No. 3969647

  The problem to be solved by the present invention is to provide a defect detection apparatus, a defect detection method, and a computer program capable of realizing defect inspection of a wide range of subjects with a simpler configuration.

  The defect detection apparatus according to the embodiment includes an information acquisition unit, a defect detection unit, and a defect depth determination unit. The information acquisition unit acquires surface temperature information indicating the surface temperature of the subject and temperature difference information indicating the temperature difference in the vicinity of the subject for a predetermined period. The defect detection unit detects the suspected part of the defect in the subject and the size of the suspected part based on the distribution of the surface temperature of the subject indicated by the surface temperature information. The defect depth determination unit is configured to detect the suspect based on the size of the suspected part detected by the defect detecting part and the temperature difference information according to the timing at which the surface temperature information from which the suspected part was detected was acquired. Determine the depth of the site.

Schematic which shows the outline | summary of the defect detection method in embodiment. The functional block diagram which shows the function structure of the defect detection apparatus 100 of 1st Embodiment. The figure which shows the specific example of the thermal image classified and memorize | stored for every acquisition period. The figure which shows the specific example of a thermal image. The figure which shows the specific example of temperature difference information. The figure which shows the specific example of dictionary information. The figure which shows the specific example of defect information. The flowchart which shows the flow of a defect detection process. The figure which shows the specific example of the thermal image classified into the temperature difference group. The figure which shows the specific example of an averaged image. The flowchart which shows the flow of a defect depth determination process. The flowchart which shows the flow of a defect determination process. The figure which shows the specific example of the change of an averaged image. The figure which shows the specific example of the change of a temperature difference image. The figure which shows the specific example of the change of the magnitude | size of a suspected site | part. The figure which shows an example of the screen which displays a defect detection result. The functional block diagram which shows the function structure of the defect detection apparatus 100a of 2nd Embodiment. The flowchart which shows the flow of a position alignment process. The figure which shows the specific example of the thermal image of a tunnel entrance. The figure which shows the specific example of a binarized image. The figure which shows the specific example of the common area | region extracted from a thermal image.

(Outline)
FIG. 1 is a schematic diagram showing an outline of a defect detection method in the following embodiment.
The inspection vehicle 10 (information acquisition vehicle) is a vehicle including means for measuring the surface temperature of the subject and means for measuring the air temperature in the vicinity of the subject. For example, the inspection vehicle 10 includes a thermal imaging device 11 (imaging unit) as a means for measuring the surface temperature of the subject, and a thermocouple 12 as a means for measuring the temperature near the subject. For example, the thermal image pickup device 11 is an infrared camera having sensitivity to infrared light in the vicinity of 10 μm. Further, the thermocouple 12 may be used in combination with a data logger in order to increase the accuracy of temperature measurement. Tunnel 20 is an example of a wide range of subjects.

  The inspection vehicle 10 travels through the tunnel 20 while photographing inside the tunnel 20 and measuring the temperature. By performing such traveling, the inspection vehicle 10 can measure the surface temperature of the subject over a wide range and the temperature near the subject at each point. Hereinafter, traveling in which such temperature measurement is performed is referred to as inspection traveling. The purpose of measuring the temperature in the vicinity of the subject at each point is to acquire different daily temperature differences at each point in the tunnel 20. Therefore, it is desirable that the inspection traveling performed by the inspection vehicle 10 be performed at least twice a day and at a timing at which the temperature difference is maximized. Further, in order to facilitate comparison of thermal images acquired in the inspection traveling at different timings, it is desirable that the inspection traveling is performed at the same traveling route and at the same speed.

  In addition, when the inspection run can be performed only once a day, for example, the measurement at the lowest temperature and the measurement at the highest temperature may be alternately performed every other day. In this case, you may make it select the measurement data used for acquisition of a temperature difference according to the weather. In addition, in the case where the inspection traveling is performed at a timing at which the temperature difference is not maximized, the temperature difference obtained from the measurement data may be corrected to the daily temperature difference. In this case, the coefficient used for correction may be changed according to the season and weather.

  The defect detection apparatus of the embodiment detects a defect of the subject based on the surface temperature of the subject acquired in this way, and detects the detected defect based on the temperature difference of the day near the subject. Depth (hereinafter referred to as “defect depth”) is determined. Such defect inspection eliminates the need for a specimen and eliminates the need for temperature measurement for each part of a wide range of subjects. Therefore, defect inspection of a wide range of subjects can be realized with a simpler configuration, and labor and cost required for inspection can be reduced.

  Hereinafter, a defect detection apparatus, a defect detection method, and a computer program according to embodiments will be described with reference to the drawings.

(First embodiment)
FIG. 2 is a functional block diagram illustrating a functional configuration of the defect detection apparatus 100 according to the first embodiment.
The defect detection apparatus 100 includes a CPU (Central Processing Unit), a memory, an auxiliary storage device, and the like connected by a bus, and executes a defect detection program. The defect detection apparatus 100 includes a thermal image acquisition unit 101, a thermal image storage unit 102, a temperature difference information acquisition unit 103, a temperature difference information storage unit 104, a dictionary information storage unit 105, a defect detection unit 106, a defect by executing a defect detection program. It functions as an apparatus including a depth determination unit 107, a defect information storage unit 108, and a defect determination unit 109. All or some of the functions of the defect detection apparatus 100 may be realized using hardware such as an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA). . The defect detection program may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The defect detection program may be transmitted via a telecommunication line.

  The thermal image acquisition unit 101 acquires surface temperature information indicating the surface temperature of the subject. The surface temperature information may be acquired by any method as long as it is information indicating the surface temperature of the subject, but in the present embodiment, a wide range of subjects are assumed, and the thermal image acquisition unit 101 captures thermal images. A thermal image acquired by the apparatus 11 is acquired as surface temperature information. The thermal image acquisition unit 101 outputs the acquired thermal image of the subject to the thermal image storage unit 102.

  The thermal image storage unit 102 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The thermal image storage unit 102 stores the thermal image output from the thermal image acquisition unit 101. The thermal image storage unit 102 stores thermal images classified for each predetermined acquisition period corresponding to the thermal image capturing timing. For example, as illustrated in FIG. 3, the thermal image storage unit 102 classifies the imaging timing into acquisition periods every three months and stores thermal images.

FIG. 3 is a diagram illustrating a specific example of thermal images classified and stored for each acquisition period.
FIG. 3 shows that thermal images acquired between 2013 and 2015 are classified and stored in 12 acquisition periods every three months. K in the figure represents a number for identifying each acquisition period. In the example of FIG. 3, k takes a value of 0 to 11. k = 0, 1, 2, 3 corresponds to each acquisition period of 2013, k = 4, 5, 6, 7 corresponds to each acquisition period of 2014, k = 8, 9, 10, 11 It corresponds to each acquisition period in 2015. The identification numbers of the acquisition periods in each year correspond to the acquisition periods from January to February, April to May, July to August, and October to November, in ascending order of value. Here, March, June, September, and December are excluded from the acquisition period. This is for providing a separation period between the acquisition periods to emphasize the characteristics of the daily temperature difference according to the acquisition period (for example, the season). The reason for emphasizing the characteristics of the temperature difference in this way is that the influence of defects appearing on the surface temperature differs depending on the temperature difference of the day. In general, the larger the daily temperature difference, the more deeply the effects of defects appear on the surface temperature.

FIG. 4 is a diagram illustrating a specific example of a thermal image.
The thermal image 200 in FIG. 4 is an example of a thermal image P including a total of i _max × j _max pixels, i _max in the vertical I direction and j _{max in the} horizontal J direction. P (i, j) represents a pixel or a pixel value at the position of the i th vertical and j th horizontal starting from the upper left of the thermal image 200. The pixel value P (i, j) in the thermal image P represents the surface temperature of the subject imaged by the pixel. In the following description, similarly to the thermal image 200, an arbitrary pixel or pixel value in the image Z is described as Z (x, y) using the vertical pixel number x and the horizontal pixel number y. .

Returning to the description of FIG.
The temperature difference information acquisition unit 103 acquires temperature difference information indicating a daily temperature difference near the subject. The temperature difference information may be acquired by any method as long as it is information indicating the daily temperature difference in the vicinity of the subject. In this embodiment, the temperature difference information acquisition unit 103 assumes a wide range of subjects. Acquires temperature difference information based on the temperature measured by the thermocouple 11. Further, in the present embodiment, it is assumed that the time when the lowest temperature of the day is 5 o'clock and the time when the highest temperature is 15 o'clock is 15:00, and the inspection run is performed twice a day at 5 o'clock and 15 o'clock. To do. The temperature difference information acquisition unit 103 calculates the temperature difference from 5 o'clock to 15 o'clock (a predetermined period of the day) obtained by the daily inspection run, and the temperature difference information indicating the calculated temperature difference and the date of the day Is generated. The temperature difference information acquisition unit 103 outputs the acquired temperature difference information to the temperature difference information storage unit 104.

  The temperature difference information storage unit 104 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The temperature difference information storage unit 104 stores the temperature difference information output from the temperature difference information acquisition unit 103.

FIG. 5 is a diagram illustrating a specific example of temperature difference information.
The temperature difference information table 1041 is an example of a mode in which the temperature difference information is stored as a table. The temperature difference information table 1041 has a temperature difference information record for each date when the temperature difference indicated by the temperature difference information is measured. The temperature difference information record has items of date and temperature difference. The item of date represents the date when the temperature difference indicated by the temperature difference information is measured. The item of temperature difference represents the temperature difference corresponding to the date. The temperature difference item is provided for each year in which the temperature difference is measured, and the temperature difference corresponding to each date is registered in the item corresponding to the measured year. In addition, each area | region shown with a broken line in FIG. 5 represents the response | compatibility with the temperature difference information memorize | stored in the temperature difference information table 1041, and the acquisition period k.

Returning to the description of FIG.
The dictionary information storage unit 105 stores dictionary information indicating the relationship between the daily temperature difference near the subject, the size of the defect, and the defect depth. The dictionary information storage unit 105 stores in advance dictionary information (correlation information) corresponding to the nature of the subject and the environment in which the subject is placed.

FIG. 6 is a diagram illustrating a specific example of dictionary information.
The dictionary information tables 1051 and 1052 are an example of a mode in which dictionary information is stored as a table. The dictionary information table 1051 is a specific example in which the defect depth is classified into a group of consecutive temperature differences (hereinafter referred to as “temperature difference group”), and the dictionary information table 1052 indicates the defect depth. It is a specific example in the case of classifying into discrete temperature difference groups. C in the figure is an identification number of the temperature difference group.

  The dictionary information tables 1051 and 1052 both have dictionary information records for each defect size. Each dictionary information record has items of defect size and defect depth. The item of defect size represents the size of a defect that can be detected by the defect detection apparatus 100. The item of defect depth represents the defect depth corresponding to the size of the defect. The defect depth item is provided for each temperature difference group, and a preset defect depth is registered for each temperature difference range indicated by the temperature difference group.

  The dictionary information may be stored in either form of the dictionary information table 1051 or 1052. However, if the defect depth is to be estimated more accurately, the dictionary information table 1052 having discrete temperature difference groups is used. Good. However, when the dictionary information table 1052 is used, the number of thermal images that can be used for defect detection decreases, so that the defect detection accuracy may decrease. Therefore, when the dictionary information is discretized, it is necessary to adjust the interval and granularity (resolution) of the temperature difference groups so that the accuracy of both is maintained in a well-balanced manner.

Returning to the description of FIG.
The defect detection unit 106 performs a defect detection process for detecting a suspected portion of a defect in the subject based on the distribution of the surface temperature of the subject indicated by the thermal image. The defect detection unit 106 outputs information indicating the suspected part detected by the defect detection process to the defect depth determination unit 107.

  The defect depth determination unit 107 determines the defect depth of the suspected part detected by the defect detection unit 106. Specifically, the defect depth determination unit 107 detects the defect of the suspected part based on the size of the suspected part, the temperature difference on the acquisition date of the thermal image in which the suspected part is detected, and the dictionary information. A defect depth determination process for determining the depth is performed. The defect depth determination unit 107 generates defect information regarding the suspected portion determined in the defect depth determination process, and outputs the defect information to the defect information storage unit 108.

  The defect information storage unit 108 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The defect information storage unit 108 stores the defect information output from the defect depth determination unit 107.

FIG. 7 is a diagram illustrating a specific example of defect information.
The defect information table 1081 is an example of a mode in which defect information is stored as a table. The defect information table 1081 has a defect information record for each detected suspected part. The defect information record includes items of an acquisition period, a defect position, a temperature difference group, and a defect size. The item “acquisition period” represents an acquisition period in which a thermal image in which the suspected site is detected is captured. The item of defect position represents position information indicating the position of the suspected part. For example, the positional information includes information such as the time or number of frames until the thermal image in which the suspected part is detected from the start of the test run, the imaging time of the thermal image, and the suspected in the thermal image. It is represented by the combination with the detection position of the part. The item of the temperature difference group represents a temperature difference group into which the thermal image in which the suspected part is detected is classified. The item of the defect size represents the size of the detected suspicious site.

  The defect determination unit 109 determines whether or not the suspected site detected by the defect detection unit 106 is a defect. Specifically, when the defect determination unit 109 compares the size of the suspected part detected based on the thermal images acquired on the same day with the temperature difference, and the size of the suspected part shows a tendency to increase A defect determination process for determining the suspected part as a defect is performed.

  Hereinafter, specific examples of the defect detection process, the defect depth determination process, and the defect determination process described above will be described.

FIG. 8 is a flowchart showing the flow of the defect detection process.
The defect detection process shown in the flowchart of FIG. 8 is a process for detecting a suspected part from a thermal image acquired in each acquisition period k. First, the defect detection unit 106 acquires a thermal image acquired in the processing target acquisition period k from the thermal image storage unit 102 (step S101). The defect detection unit 106 classifies the acquired thermal image into each temperature difference group based on the temperature difference on the day when the thermal image is acquired (step S102). For example, when the acquisition period of the processing target is k = 5, the defect detection unit 106 classifies the thermal image with reference to the temperature difference corresponding to the acquisition period of k = 5 in the temperature difference information table 1041 of FIG. .

FIG. 9 is a diagram illustrating a specific example of thermal images classified into temperature difference groups.
The thermal images 300-1 to 300-m (m is an integer of 1 or more) and the thermal images 301-1 to 301-n (n is an integer of 1 or more) shown in FIG. It is a thermal image acquired in the acquisition period (from May 1 to May 31, 2014). A square in each thermal image represents an assumed suspected part to be detected. In FIG. 9, the thermal images 300-1 to 300-m are classified into C = 1 temperature difference groups (temperature difference 3 to 6 ° C.), and the thermal images 301-1 to 301-n are C = 2 temperature difference groups. The case where it was classified into (temperature difference 6-9 degreeC) is shown. As described above, generally, the larger the daily temperature difference is, the more deeply the influence of defects appears on the surface temperature. Therefore, by classifying the thermal image into the temperature difference group, it is possible to estimate the defect depth, the size of the defect in the depth direction, and the like. For example, in the case of the example in FIG. 9, assuming that the suspected part in the thermal image is 5 cm square, the suspected part is a defect with a depth ranging from 20 mm to 30 mm based on the dictionary information in the example of FIG. It can be estimated that there is.

Returning to the description of FIG.
Subsequently, the defect detection unit 106 generates an averaged image obtained by averaging the thermal images classified in step S102 for each temperature difference group (step S103). For example, the defect detection unit 106 generates an averaged image based on a thermal image acquired by a predetermined inspection traveling (for example, an inspection traveling at 15:00) of the daily inspection traveling. Specifically, the defect detection unit 106 calculates an average pixel value between the thermal images for each pixel of the thermal image classified into a certain temperature difference group. The average pixel value is obtained by the following equation (1), for example.

In the formula (1), d (d = 1, 2,..., D _max ) is an identification number of the thermal image in each temperature difference group. P _ck is acquired in the acquisition period k (in this embodiment, k = 0, 1,..., 11) and is classified into the temperature difference group c (in this embodiment, c = 0, 1, 2, 3). _Represents an averaged image of the image P _ckd . P _ck (i, j) represents each pixel value of the averaged image.

FIG. 10 is a diagram illustrating a specific example of the averaged image.
The averaged image 400 of FIG. 10 is based on the thermal images 301-1 to 301-n classified into the C = 2 temperature difference group (temperature difference 6 to 9 ° C.) among the thermal images classified in FIG. Represents an averaged image generated in the above manner. By generating an averaged image of a plurality of thermal images in this way, it is possible to remove (or reduce) the influence of noise included in the plurality of thermal images. For example, the thermal image 301-2 in FIG. 10 shows a suspected part larger than the suspected parts in the other thermal images 301-1 and 301-3 to 301-n. Here, it is assumed that this difference in size is caused by noise. In this case, in the averaged image 400 generated based on the thermal images 301-1 to 301-n, the influence of noise on the thermal image 301-2 is caused by the other thermal images 301-1 and 301-3 to 301-n. It is diluted. Therefore, the defect detection apparatus 100 can suppress erroneous detection of a suspected site due to noise by generating an averaged image.

Returning to the description of FIG.
Subsequently, the defect detection unit 106 detects the suspected part of the defect from the averaged image generated in step S103. Specifically, the defect detection unit 106 detects the suspected portion by performing the following process on the averaged image.

  First, the defect detection unit 106 calculates an average pixel value (hereinafter referred to as “image average value”) for the entire image for each averaged image. The image average value is obtained by the following equation (2), for example.

P _{ck_avg} in Equation (2) represents an image average value in the averaged image P _ck of the thermal images P _ckd acquired in the acquisition period k and classified into the temperature difference group c. i _max represents the number of pixels in the vertical direction of the averaged image, and j _max represents the number of pixels in the horizontal direction.

Next, the defect detection unit 106 generates a temperature difference image in which each pixel value of the averaged image is represented by a temperature difference from the image average value (step S104). The temperature difference image is obtained by the following equation (3), for example. R _ck (i, j) in Expression (3) represents each pixel value in the temperature difference image R _ck .

Next, the defect detection unit 106 detects a pixel having a pixel value equal to or greater than a predetermined threshold among the pixels of the generated temperature difference image as a suspected part (step S105). This predetermined threshold is preset in the defect detection unit 106. For example, when the determination threshold is represented as _Tc , the defect detection unit 106 detects the suspected region by performing the determination of the following equation (4) for all the pixels of the temperature difference image.

FIG. 11 is a flowchart showing the flow of the defect depth determination process.
The defect depth determination process shown in the flowchart of FIG. 11 is a process of determining the defect depth of the suspected part detected by the defect detection process of FIG. First, the defect depth determination unit 107 acquires the size of the detected suspicious part (step S201). Specifically, the defect depth determination unit 107 obtains the size of the detected suspected part by counting the number of pixels detected as the suspected part.

Subsequently, the defect depth determination unit 107 determines the defect depth of the suspected part based on the acquired size of the suspected part and the temperature difference group corresponding to the temperature difference image in which the suspected part is detected (step) S202). Specifically, the defect depth determination unit 107 refers to the dictionary information table 1051 and acquires the value of the defect depth according to the acquired size of the suspected part and the corresponding temperature difference group, Determine the defect depth of the suspected part. The defect depth determination unit 107 records the size of the suspected region determined in this way (defect size) in the defect information table 1081 in association with the acquisition period, the defect position, and the temperature difference group.
When the aspect ratio of the suspected part is significantly different from a square such as 2: 1, for example, the size of the defect in the dictionary information of FIG. 6 may be set according to the shape such as 10 cm square and 5 cm × 20 cm. Thus, by setting dictionary information according to a shape, the defect depth determination part 107 can determine a defect depth more accurately according to the shape of a suspected part.

FIG. 12 is a flowchart showing the flow of the defect determination process.
First, the defect determination unit 109 acquires defect information regarding the suspected part to be determined (step S301). Specifically, the defect determination unit 109 refers to the defect information table 1081 and selects defect information records having the same value for the temperature difference group and defect position items, thereby acquiring the same suspected part at different acquisition periods. Acquired defect information.

  Subsequently, the defect determination unit 109 acquires the tendency of change in the size of the suspected region (hereinafter referred to as “growth tendency”) by comparing the acquired defect information in time series (step S302). Specifically, the defect determination unit 109 acquires the growth tendency of the suspected region by comparing in the order of the acquisition period k. The defect determination unit 109 determines whether or not an increase tendency is seen in the acquired growth tendency (step S303). When the acquired growth tendency shows an expansion tendency (step S303—YES), the defect determination unit 109 determines the suspected part as a defect (step S304). On the other hand, if the acquired growth tendency does not show an expansion tendency (NO in step S303), the defect determination unit 109 ends the process without determining the suspected part as a defect.

FIG. 13 is a diagram illustrating a specific example of the change in the averaged image.
FIG. 13 is a diagram comparing averaged images generated based on thermal images classified into C = 2 temperature difference groups (temperature difference 6-9 ° C.) in the acquisition periods of the same period of each year. . As shown in FIG. 13, in the averaged image state, the possible range of pixel values changes according to the timing of the acquisition period, so it is necessary to provide a threshold for determining the suspected site according to each timing. is there. Therefore, the defect determination unit 109 generates a temperature difference image from the averaged image in order to match the possible range of pixel values.

FIG. 14 is a diagram illustrating a specific example of a change in the temperature difference image.
FIG. 14 is a diagram showing a comparison when the averaged image in FIG. 13 is replaced with a temperature difference image. Thus, by generating the temperature difference image from the averaged image, the possible range of pixel values can be matched to the same range at each time. Thereby, it becomes possible to set the threshold value for determining the suspected part according to the temperature difference.

  In addition, the temperature difference image in FIG. 14 shows that the suspected site was detected for the first time in the acquisition period of k = 4, and the size of the suspected site was increased in the acquisition period of k = 11. The defect determination unit 109 determines that the suspected part is a defect when such an expansion tendency is seen in the growth tendency of the suspected part. In the defect determination process described with reference to FIG. 12, the defect determination unit 109 acquires the growth tendency of the suspected region by comparing the temperature difference images of the same temperature difference group in time series. However, since the growth tendency generally shows a gradual change, it is not always possible to capture the change in the size of the suspected region by comparison between previous and subsequent acquisition periods. In such a case, the defect determination unit 109 may determine a growth tendency based on a temperature difference image in an acquisition period that is separated to some extent. For example, the defect determination unit 109 may compare the temperature difference images between acquisition periods (for example, acquisition periods represented by k = 0, 4, and 8) corresponding to the same period of the year.

FIG. 15 is a diagram illustrating a specific example of a change in the size of the suspected region.
FIG. 15 shows an example of a change in the size of a suspected part detected based on thermal images of a temperature difference group with C = 1 and a temperature difference group with C = 2. The year-on-year difference is shown. In FIG. 15, two temperature difference groups indicate changes in different suspected parts. In this case, in the temperature difference group of C = 1, an obvious increasing tendency is seen in the growth tendency of the suspected part. In such a case, the defect determination unit 109 determines the suspected part as a defect.

  On the other hand, in the temperature difference group of C = 2, the growth tendency of the suspected part is unknown. For example, when compared in the acquisition periods represented by k = 4 and 11, a decreasing tendency is seen in the growth tendency of the suspected part. In general, it is rare in nature to reduce the size of defects. Therefore, such a decreasing tendency is considered to be an influence of noise during measurement. The influence of short-term noise is removed by the generation of an averaged image, but in some cases, such long-term noise may occur. Therefore, it is possible to detect a defect with higher accuracy by observing the growth tendency of the suspected portion over a long period of time.

  The defect detection device 100 according to the first embodiment configured as described above detects a defect of the subject based on the surface temperature of the subject, and is detected based on the daily temperature difference near the subject. Defect depth is determined. If defect inspection is performed by such a method, a specimen is not necessary, and it is not necessary to perform temperature measurement for each part of a wide range of subjects. Therefore, defect inspection of a wide range of subjects can be realized with a simpler configuration, and labor and cost required for inspection can be reduced.

  Hereinafter, modifications of the defect detection apparatus 100 according to the first embodiment will be described.

  The defect detection apparatus 100 may be configured not to include the defect determination unit 109. In this case, the defect detection apparatus 100 may be configured to determine the detected suspicious site as a defect.

  The defect detection apparatus 100 detects a suspected site using a thermal image acquired by any one of the thermal images acquired a plurality of times per day. The suspected part may be detected based on the difference image. For example, the defect detection unit 106 may generate an averaged image based on a difference image between a thermal image acquired at 5 o'clock and a thermal image acquired at 15:00. Since the temperature difference between the normal part and the defective part is reversed in the early morning (nighttime) and daytime, when such a difference image is used, it is considered that the temperature difference is amplified by the difference, and the defective part can be detected more easily (for example, , See "Infrared Thermography Deformation Investigation Manual, Civil Engineering Research Center").

  The defect detection apparatus 100 may be configured to visualize information related to defect inspection and provide it to the user. For example, in this case, the defect detection apparatus 100 has a defect in a display unit (not shown) configured using a display device such as a CRT (Cathode Ray Tube) display, a liquid crystal display, or an organic EL (Electro-Luminescence) display. A display control unit for displaying information related to the examination is provided. The display control unit may display any information on the display unit as long as the information is stored in the defect detection apparatus 100. Further, the display control unit may display information on the defect inspection on the display unit in any manner. Next, FIG. 16 shows an example in which information regarding defect inspection is displayed.

FIG. 16 is a diagram illustrating an example of a screen that displays the defect detection result.
A screen example 500 shown in FIG. 16 displays a temperature difference image 501 indicating the detection result of the defect and a detection condition display 502 indicating the detection condition. The temperature difference image 501 may be the temperature difference image itself generated by the above processing, or may be a temperature difference image processed by the processing result or the determination result. For example, the temperature difference image 501 may be an image that displays only a suspected part determined as a defect, or may be an image that displays all suspected parts. In that case, you may add and display the identification information whether it is a defect in each suspected part. Further, as shown in the figure, each suspected part may be expressed by a color or pattern corresponding to the defect depth.

  On the other hand, as shown in the figure, the detection condition display 502 may display a detectable defect depth, a defect size, and the like, or other conditions related to defect detection. For example, the timing at which the thermal image is acquired, the traveling speed of the inspection vehicle 10, the imaging conditions (for example, the frame rate) at the time of acquiring the thermal image, and the like may be displayed.

(Second Embodiment)
FIG. 17 is a functional block diagram illustrating a functional configuration of the defect detection apparatus 100a according to the second embodiment.
The defect detection device 100a is different from the defect detection device 100 of the first embodiment in that it further includes an alignment unit 110.
The alignment unit 110 performs an alignment process of extracting a common region in which the same part of the subject is imaged from thermal images acquired at different examination travel timings. The alignment unit 110 stores the extracted image of the common area in the thermal image storage unit 102 in association with the acquisition period.

FIG. 18 is a flowchart showing the flow of the alignment process.
Here, the alignment process with respect to the thermal image acquired by the inspection vehicle 10 which images the inside of the tunnel 20 will be described as an example of the alignment process. First, the alignment unit 110 acquires a series of thermal images in which the inside of the tunnel 20 is captured from the thermal image storage unit 102 (step S401). A series of thermal images acquired here are thermal images acquired by one inspection run. The alignment unit 110 selects a thermal image obtained by capturing the entrance of the tunnel 20 from the acquired series of thermal images (step S402).

FIG. 19 is a diagram illustrating a specific example of a thermal image of a tunnel entrance.
In general, since the temperature outside the tunnel and the temperature inside the tunnel are greatly different, the thermal image obtained by imaging the entrance or exit of the tunnel has a large temperature difference. Therefore, first, the alignment unit 110 selects a thermal image having a large temperature difference from the acquired series of thermal images. As a result, an image in which the entrance or exit of the tunnel is imaged is selected. FIG. 19A is an example of a thermal image obtained by imaging the tunnel entrance, and FIG. 19B is an example of a thermal image obtained by imaging the tunnel exit. As described above, due to the difference in temperature described above, the thermal image obtained by imaging the entrance or exit of the tunnel is often an image that can clearly identify the outline of the entrance or exit of the tunnel. The alignment unit 110 performs image processing on the selected thermal image to identify the contour of the entrance or exit of the tunnel. The alignment unit 110 selects the thermal image having the lower temperature inside the identified contour as the thermal image in which the tunnel entrance is captured. The alignment unit 110 may simply determine that the thermal image captured first is the thermal image captured of the tunnel entrance.

Returning to the description of FIG.
Subsequently, the alignment unit 110 acquires the coordinates of the boundary portion of the tunnel 20 from the selected thermal image (step S403). For example, the alignment unit 110 acquires the coordinates of the boundary using a binarized image obtained by binarizing the thermal image between the outside and the inside of the tunnel as shown in FIG.

FIG. 20 is a diagram illustrating a specific example of a binarized image.
Here, the position in the binarized image is represented by xy coordinates with the upper left as the origin O. In the binarized image of FIG. 20, the vertical axis represents the y coordinate and the horizontal axis represents the x coordinate. In this case, for example, the alignment unit 110 acquires the coordinates of the boundary point P having the smallest y-coordinate value on the boundary line of the binarized image as the coordinates of the boundary part.

Returning to the description of FIG.
The alignment unit 110 extracts the image of the common area from the series of thermal images based on the coordinates of the boundary points acquired in this way (step S404).

FIG. 21 is a diagram illustrating a specific example of the common area extracted from the thermal image.
The thermal images 600-1 to 600-t (t is an integer of 1 or more) shown in FIG. 21A and the thermal images 601-1 to 601-t shown in FIG. A series of thermal images acquired by traveling is shown. As can be seen from the figure, there is a slight difference in imaging timing between the thermal images 600-1 to 600-t and the thermal images 601-1 to 601-t. Therefore, the defect detected based on the thermal images 600-1 to 600-t and the defect detected based on the thermal images 601-1 to 601-t may be determined as different defects. Therefore, the alignment unit 110 corrects the imaging timing shift by extracting a common region from each thermal image based on the coordinates of the boundary portion of the tunnel entrance.

The y coordinate of the boundary point P ₁ in FIG. 21A is d ₁ , and the y coordinate of the boundary point P ₂ in FIG. 21B is d ₂ . The alignment unit 110 determines the length of the common region in the y-axis direction with reference to an image having a boundary point P ₁ with a larger y coordinate (here, FIG. 21A). In this case, the alignment unit 110 converts the region where the y coordinate is d ₁ ≦ y ≦ d ₁ + d ₃ (the region indicated by the broken line in the thermal image in FIG. 21A) to the thermal images 600-1 to 600-t. The region where the y coordinate is d ₂ ≦ y ≦ d ₂ + d ₃ (the region indicated by the broken line in the thermal image of FIG. 21B) is extracted from the thermal images 601-1 to 601-t. The region extracted in this manner is a common region in which the same part of the tunnel 20 is captured in the thermal images captured at the same timing in FIGS. 21A and 21B.

  If a thermal image is captured at 30 fps and the inspection vehicle 10 travels at a speed of 50 km / h, it is assumed that the imaging timing has shifted by 0.73 m. In this case, in order to correct the deviation of the imaging timing by the above method, the visual field range of the thermal imaging device 11 in the traveling direction may be secured about 1.5 m or more. In this case, since the visual field range of the thermal image pickup device 11 is about twice or more the image pickup timing, the images in the tunnel can be connected without missing.

  The defect detection apparatus 100a of the second embodiment is acquired at different timings by having an alignment unit that extracts a common region in which the same part of the subject is imaged from thermal images acquired at different inspection travel timings. It is possible to correct the deviation of the imaging timing of the thermal image. By performing such correction, the defect detection apparatus 100a can detect a defect with higher accuracy.

  According to at least one embodiment described above, a defect detection unit that detects a suspected part of a defect in the subject and the size of the suspected part based on the distribution of the surface temperature of the subject indicated by the surface temperature information, and the suspected part By having a defect depth determination unit for determining the depth of the detected suspected part based on the size of the temperature difference information on the day when the surface temperature information from which the suspected part was detected was acquired, A wide range of subject defect inspections can be realized with a simpler configuration.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 10 ... Inspection vehicle, 11 ... Thermal imaging device, 12 ... Thermocouple, 20 ... Tunnel, 100, 100a ... Defect detection device, 101 ... Thermal image acquisition part, 102 ... Thermal image storage part, 103 ... Temperature difference information acquisition part 104 ... Temperature difference information storage unit, 1041 ... Temperature difference information table, 105 ... Dictionary information storage unit, 1051 ... Dictionary information table, 1052 ... Dictionary information table, 106 ... Defect detection unit, 107 ... Determination unit, 108 ... Defect information Storage unit, 1081 ... Defect information table, 109 ... Defect determination unit, 110 ... Positioning unit, 200, 300-1 to 300-m, 301-1 to 301-n ... Thermal image, 400 ... Averaged image, 500 ... Screen example displaying defect detection results, 501... Temperature difference image, 502... Detection condition display, 600-1 to 600-t, 601-1 to 601-t.

Claims (18)

An information acquisition unit that acquires surface temperature information indicating a surface temperature of a subject, and temperature difference information indicating a temperature difference of a predetermined period in the vicinity of the subject,
A defect detection unit for detecting a suspected part of a defect in the subject and a size of the suspected part based on a distribution of the surface temperature of the subject indicated by the surface temperature information;
The depth of the suspected part is determined based on the size of the suspected part detected by the defect detector and the temperature difference information corresponding to the timing at which the surface temperature information from which the suspected part was detected was acquired. A defect depth determination unit;
A defect detection apparatus comprising: A storage unit that stores in advance correlation information indicating the relationship between the size of the suspected part and the temperature difference and the depth of the suspected part;
The defect depth determination unit stores temperature difference information according to the timing at which the size of the suspected part detected by the defect detection part and the surface temperature information from which the suspected part was detected are acquired, and the storage part Determining the depth of the suspected site based on the correlation information
The defect detection apparatus according to claim 1. The suspected part is compared when the size of the suspected part detected based on the surface temperature information acquired at the timing when the temperature difference is within a predetermined range and tends to increase to the suspected part. A defect determination unit for determining the defect as a defect,
The defect detection apparatus according to claim 1. The surface temperature information is a thermal image indicating the surface temperature of the subject,
The defect detection unit uses, as the suspected region, an area indicated by a pixel in which a difference between an average value of the temperature indicated by each pixel of the thermal image and a temperature indicated by each pixel of the thermal image is equal to or greater than a predetermined threshold. To detect,
The defect detection apparatus according to claim 1. The defect detection unit generates an averaged image obtained by averaging the thermal images acquired within a predetermined acquisition period, and is included in the thermal image by detecting the suspected site based on the averaged image. Suppress false detection of the suspected site due to noise,
The defect determination unit determines an increasing tendency of the size of the suspected site by comparing the size of the suspected site detected based on the averaged image for each acquisition period.
The defect detection apparatus according to claim 4. An alignment unit that extracts a common area in which the same part of the subject is imaged from the thermal images acquired at different timings;
The defect detection unit generates the averaged image based on the image of the common region extracted by the alignment unit;
The defect detection apparatus according to claim 5. An information acquisition step of acquiring surface temperature information indicating a surface temperature of a subject and temperature difference information indicating a temperature difference in a predetermined period in the vicinity of the subject;
A defect detection step for detecting a suspected part in the subject and a size of the suspected part based on a distribution of the surface temperature of the subject indicated by the surface temperature information;
The depth of the suspected part is determined based on the size of the suspected part detected in the defect detection step and the temperature difference information corresponding to the timing at which the surface temperature information from which the suspected part was detected was acquired. A defect depth determination step;
A defect detection method comprising: In the defect depth determination step, the correlation information indicating the size of the suspected part and the relationship between the temperature difference and the depth of the suspected part, the size of the suspected part detected in the defect detection step, and the suspected part are detected. Determining the depth of the suspected part based on the temperature difference information corresponding to the timing at which the surface temperature information was acquired,
The defect detection method according to claim 7. The suspected part is compared when the size of the suspected part detected based on the surface temperature information acquired at the timing when the temperature difference is within a predetermined range and tends to increase to the suspected part. A defect determination step for determining the defect as a defect,
The defect detection method according to claim 7 or 8. The surface temperature information is a thermal image indicating the surface temperature of the subject,
In the defect detection step, an area indicated by a pixel in which a difference between an average value of the temperature indicated by each pixel of the thermal image and a temperature indicated by each pixel of the thermal image is equal to or greater than a predetermined threshold is set as the suspected part. To detect,
The defect detection method according to any one of claims 7 to 9. In the defect detection step, an averaged image obtained by averaging the thermal images acquired within a predetermined acquisition period is generated, and the suspected portion is detected based on the averaged image, and is included in the thermal image. Suppress false detection of the suspected site due to noise,
In the defect determination step, the increase tendency of the size of the suspected portion is determined by comparing the size of the suspected portion detected based on the averaged image for each acquisition period.
The defect detection method according to claim 10. An alignment step of extracting a common region in which the same part of the subject is imaged from the thermal images acquired at different timings;
In the defect detection step, the averaged image is generated based on the image of the common area extracted by the alignment unit.
The defect detection method according to claim 11. An information acquisition step of acquiring surface temperature information indicating a surface temperature of a subject and temperature difference information indicating a temperature difference in a predetermined period in the vicinity of the subject;
A defect detection step for detecting a suspected part in the subject and a size of the suspected part based on a distribution of the surface temperature of the subject indicated by the surface temperature information;
The depth of the suspected part is determined based on the size of the suspected part detected in the defect detection step and the temperature difference information corresponding to the timing at which the surface temperature information from which the suspected part was detected was acquired. A defect depth determination step;
A computer program for causing a computer to execute. In the defect depth determination step, the correlation information indicating the size of the suspected part and the relationship between the temperature difference and the depth of the suspected part, the size of the suspected part detected in the defect detection step, and the suspected part are detected. Determining the depth of the suspected part based on the temperature difference information corresponding to the timing at which the surface temperature information was acquired,
The computer program according to claim 13. The suspected part is compared when the size of the suspected part detected based on the surface temperature information acquired at the timing when the temperature difference is within a predetermined range and tends to increase to the suspected part. A defect determination step for determining the defect as a defect,
The computer program according to claim 13 or 14. The surface temperature information is a thermal image indicating the surface temperature of the subject,
In the defect detection step, an area indicated by a pixel in which a difference between an average value of the temperature indicated by each pixel of the thermal image and a temperature indicated by each pixel of the thermal image is equal to or greater than a predetermined threshold is set as the suspected part. To detect,
The computer program according to any one of claims 13 to 15. In the defect detection step, an averaged image obtained by averaging the thermal images acquired within a predetermined acquisition period is generated, and the suspected portion is detected based on the averaged image, and is included in the thermal image. Suppress false detection of the suspected site due to noise,
In the defect determination step, the increase tendency of the size of the suspected portion is determined by comparing the size of the suspected portion detected based on the averaged image for each acquisition period.
The computer program according to claim 16. An alignment step of extracting a common region in which the same part of the subject is imaged from the thermal images acquired at different timings;
In the defect detection step, the averaged image is generated based on the image of the common area extracted by the alignment unit.
The computer program according to claim 17.