Infrared detection inspection system

Abstract

To provide an inspection system using infrared detection that can more easily grasp abnormalities on the surface of a building even when there are pixels in the surface to be inspected that represent singular values ​​based on, for example, a heat source or reflected light, and that allows anyone to easily grasp abnormalities such as deterioration or defects on the surface to be inspected. [Solution] An inspection system for detecting abnormalities in a surface to be inspected based on the amount of infrared radiation emitted from the surface to be inspected. This inspection system includes an infrared image detection means for detecting the amount of infrared radiation emitted from the surface to be inspected for each pixel, a mode extraction means for extracting at least one mode of the amount of infrared radiation for each pixel detected by the infrared image detection means, first to Nth pixel identification means for identifying pixels having different numerical values ​​in a range of first to Nth (N is a natural number equal to or greater than 2) based on the at least one mode extracted by the mode extraction means, and a display means for displaying the pixels identified by the first to Nth pixel identification means so that they can be distinguished from one another.

Description

[Technical Field]

[0001] The present invention relates to an inspection system that detects infrared rays emitted from a surface to be inspected, such as the outer or inner surface of a structure or the surface of various structural elements, to detect abnormalities in that surface. [Background technology]

[0002] As a method for detecting abnormalities such as deterioration or defects on an inspected surface by infrared detection, Patent Document 1 discloses an inspection method in which an infrared detector is placed on a transport body, the transport body is moved so that infrared rays emitted from the inspected surface are received by the infrared detector, and abnormalities on the inspected surface are detected from differences in the amount of detected infrared rays. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Special Publication No. 2-20941 Summary of the Invention [Problem to be solved by the invention]

[0004] According to the inspection method disclosed in Patent Document 1, the presence and extent of defects, etc. can be determined from differences in the amount of detected infrared rays. Patent Document 1 also proposes coloring an image according to the amount of detected infrared rays in order to visualize the presence and extent of such defects, etc. However, when inspecting the exterior walls of buildings for abnormalities, it is difficult to visually identify the presence of an abnormality from the color state unless temperature zones that require attention are precisely color-coded.

[0005] When coloring infrared images, a gradation is typically applied within the range between the maximum temperature (maximum infrared intensity) and minimum temperature (minimum infrared intensity) within the image. This gradation-based coloring is extremely effective when visually identifying objects with a clear temperature difference from their surroundings, such as fever spots in animals or overheated areas in machinery. However, when inspecting the exterior walls of buildings, the temperature difference between the abnormal areas (such as lifting, peeling, or poor insulation) that are the target of detection and the healthy areas around them is very small. Furthermore, the contrast between the inspection area where the abnormal areas are located and areas exposed to sunlight or high outdoor temperatures is large. This results in a large temperature range between the maximum and minimum temperatures for the entire image, and the temperature range of the inspection area (target area) is buried within this temperature range for the entire image, making it extremely difficult to correctly identify the abnormal areas.

[0006] For example, consider a case where the infrared temperature of multiple human bodies is detected and colored. In this case, if a high-temperature flame or extremely low-temperature ice is captured in an infrared image together with the human bodies, the temperature range of the entire image becomes large, and the human body temperatures (approximately 36-37°C) are lost within this temperature range. As a result, even if there is a slight difference in body temperature between multiple people (e.g., 36.5°C and 37.2°C), a gradation based on the maximum and minimum temperatures cannot reflect this difference to a degree that is visible as a color difference. In other words, when the overall temperature range is large, the minute temperature difference of interest is lost within that temperature range, making it extremely difficult to identify the area of ​​interest.

[0007] The same is true for the inspection of building exterior walls, where detecting abnormalities requires flexible setting of the color range according to the object of investigation. In addition, because the level of infrared reflection varies depending on the material and structure of the building surface, it can sometimes be difficult to determine abnormalities simply by measuring the temperature value.

[0008] The present invention solves the above-mentioned problems of the prior art, and its purpose is to provide an inspection system using infrared detection that can more easily identify abnormalities on the surface of a building, even when there are pixels within the surface being inspected that represent singular values ​​based on, for example, heat sources or reflected light.

[0009] Another object of the present invention is to provide an inspection system that uses infrared detection, which allows anyone to easily detect abnormalities such as deterioration or defects on the surface to be inspected. [Means for solving the problem]

[0010] According to the present invention, there is provided an inspection system for detecting abnormalities in a surface to be inspected based on the amount of infrared radiation emitted from the surface to be inspected. The inspection system of the present invention comprises infrared image detection means for detecting the amount of infrared radiation emitted from the surface to be inspected for each pixel, mode extraction means for extracting at least one mode (MODE) of the amount of infrared radiation for each pixel detected by the infrared image detection means, first to Nth pixel identification means for identifying pixels having different numerical values ​​in a range of first to Nth (N is a natural number of 2 or more) based on the at least one mode extracted by the mode extraction means, and display means for displaying the pixels identified by the first to Nth pixel identification means so that they can be distinguished from one another.

[0011] In this invention, the most frequent value of the amount of infrared light for each pixel is extracted, and pixels having different numerical values ​​in a range of 1 to N based on this frequent value are identified, and these identified pixels are displayed so that they can be distinguished from one another. In the exterior walls of buildings, etc., the difference in the amount of infrared light emitted from areas with and without abnormalities is very small, and with conventional technology it has been very difficult to identify abnormal areas. However, according to this invention, by using the most frequent value (MODE) of the amount of infrared light for each pixel in the image as a reference and focusing on the temperature difference in the surrounding area, singular values ​​(outliers) are eliminated, making it possible to identify minute abnormalities efficiently and visually.

[0012] It is preferable that the display means is means for displaying the pixels identified by the first to Nth pixel identifying means in different colors, respectively. By displaying the first to Nth pixels in different colors, abnormal areas can be more easily and clearly recognized.

[0013] It is also preferable that the mode extraction means is a means for obtaining an infrared amount score by scoring the amount of infrared light for each pixel detected by the infrared image detection means, rounding the obtained infrared amount score to obtain a mosaic value, and obtaining at least one mode of the obtained mosaic value. By scoring and further obtaining the mosaic value, it becomes possible to quantitatively handle the temperature shading (amount of infrared light), and the mode can be reliably obtained.

[0014] It is also preferable that the mode extraction means is means for extracting a plurality of modes of the amount of infrared light for each pixel detected by the infrared image detection means, and the first to Nth pixel identification means is means for identifying pixels having different numerical values ​​in the ranges of 1 to N based on each of the plurality of modes extracted by the mode extraction means. Since a plurality of modes, which are the most characteristic score values, are extracted and pixels having numerical values ​​in the ranges of 1 to N are identified for each of them, the visual enhancement effect regarding the amount of infrared light can be enhanced, and abnormal locations can be grasped easily and more clearly, thereby enabling accurate detection of abnormalities.

[0015] It is also preferable that the infrared image detection means is configured to detect infrared images by overlapping the imaging range of the surface to be inspected. By overlapping the imaging range, the amount of infrared light detected at different angles for the same location can be obtained, and the difference between the reflected and radiated components can be relatively understood.

[0016] It is also preferable that the infrared image detection means be configured to detect the amount of infrared light while facing the surface to be inspected approximately perpendicularly (directly). If the infrared image detection means faces the surface to be inspected at a large inclination angle other than approximately perpendicular, the infrared light may be totally reflected, resulting in information being obtained that differs from the actual temperature. If the infrared image detection means faces the surface to be inspected at an angle close to directly facing the surface, the correct amount of infrared light can be obtained, allowing for accurate detection of abnormalities.

[0017] It is also preferable that the infrared image detection means is configured to detect the amount of infrared light in close proximity to the surface to be inspected. By detecting the amount of infrared light in close proximity to the surface to be inspected, unnecessary image portions other than the surface to be inspected, such as an image of the background at infinity, are reduced in the image, preventing extreme bias in the infrared score distribution of the entire image and stabilizing the mode. Furthermore, close-up photography increases the change in the reflection angle, making it easier to distinguish between the effects of reflected light and radiant components, thereby improving processing accuracy.

[0018] It is also preferable that the inspection apparatus further comprises a visible light image detection means for detecting a visible light image of the surface to be inspected, and that the display means is configured to display on the same screen an image based on the amount of infrared light for each pixel detected by the infrared image detection means and the visible light image detected by the visible light image detection means. By simultaneously acquiring an infrared image and a visible light image and displaying them on the same screen, it is possible to avoid misidentification due to differences in viewpoint or time, and to compare the correspondence between the temperature distribution and the structure.

[0019] It is also preferable that the infrared image detection means is a detection means mounted on a drone. By using a drone to detect the amount of infrared light, it is possible to ensure uniform close-range photography and a uniform photography angle (angle of view), thereby maintaining consistent evaluation standards during analysis. In addition, visible light images allow for inspections similar to visual inspections. Furthermore, it is possible to capture images of surfaces to be inspected that are not visible from a fixed position, such as on the ground. [Effects of the Invention]

[0020] According to the present invention, the most frequent value (MODE) in the amount of infrared light for each pixel in an image is used as a criterion, and by focusing on the temperature difference around it, singular values ​​(outliers) are eliminated, making it possible to efficiently and visually identify minute abnormalities. [Brief explanation of the drawings]

[0021] [Figure 1] 1 is a diagram showing a schematic diagram of a partial configuration of an inspection system that uses a drone to detect abnormalities in the exterior walls of a building as one embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram showing an outline of the electrical configuration of the inspection system of FIG. [Figure 3] 2 is a flowchart schematically showing a processing flow of an image processing device in the inspection system of FIG. 1. [Figure 4] 2 is a diagram for explaining part of the processing content of the image processing device in the inspection system of FIG. 1. FIG. [Figure 5] 2 is a diagram for explaining part of the processing content of the image processing device in the inspection system of FIG. 1. FIG. [Figure 6] 2 is a diagram for explaining part of the processing content of the image processing device in the inspection system of FIG. 1. FIG. [Figure 7] 1. FIG. 3 is a diagram showing an example of a display of a processed infrared image and a visible light image when the exterior wall of a building is inspected by the inspection system of FIG. [Figure 8] 1. FIG. 3 is a diagram showing an example of a display of a processed infrared image and a visible light image when the exterior wall of a building is inspected by the inspection system of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0022] FIG. 1 shows a schematic diagram of a partial configuration of an inspection system that uses a drone to detect abnormalities in the exterior walls of buildings as one embodiment of the present invention.

[0023] In FIG. 1, reference numeral 10 denotes a building having an exterior wall 10a (corresponding to the inspection surface of the present invention) to be inspected; 11 denotes a drone operated by a control system (not shown) equipped with an infrared and visible light camera 11a (corresponding to the infrared image detection means and visible light image detection means of the present invention) capable of detecting both infrared and visible light images; and 12 denotes an image processing device located in a position capable of wireless communication with the drone 11, receiving captured image data wirelessly transmitted from the drone 11, and performing image processing and displaying the image data. By mounting the infrared and visible light camera 11a on the drone 11, uniform close-range photography and a uniform shooting angle (angle of view) can be ensured, thereby maintaining consistent evaluation criteria during analysis. Furthermore, by using the drone 11, it is possible to capture images of the exterior wall 10a in locations that cannot be seen from a fixed position, such as on the ground. In this embodiment, the image processing device 12 is located in a position capable of wireless communication with the drone 11 and is composed of a programmable computer device. Note that the image processing portion of the image processing device 12 may be located anywhere, such as on a cloud.

[0024] The drone 11 is equipped with a three-axis camera gimbal-equipped infrared and visible light camera 11a and is configured to fly close to the exterior wall 10a of the building 10 to be inspected. The drone 11 may be of any type as long as it can be equipped with a camera gimbal-equipped infrared and visible light camera 11a. In this embodiment, however, a Matrice 300RTK from DJI Japan, Inc. is used. The Matrice 300RTK has flight performance features such as a maximum flight time of approximately 40 minutes, hovering accuracy of 10 cm, and six-directional obstacle detection, and the camera gimbal can be replaced. The drone's dimensions are 810 mm in length, 670 mm in width, and 430 mm in height. The drone weighs approximately 7.1 kg (when equipped with the infrared and visible light camera 11a and flight battery). The configuration and operation of the drone 11 itself, as well as the configuration and operation of the drone's control system, are publicly known and will not be described further herein.

[0025] The infrared and visible light camera 11a can be any type of camera capable of detecting both infrared and visible light images. In this embodiment, however, a Zenmuse H20T from DJI Japan, Inc. is used. The Zenmuse H20T is a high-performance infrared camera with 640 x 512 pixels and a thermal resolution of 0.05°C, and has excellent heat dissipation capabilities. Furthermore, it can be remotely switched to a visible light camera with a maximum zoom of 200x.

[0026] As in this embodiment, by mounting the infrared and visible light camera 11a with a three-axis camera gimbal on the drone 11, it is possible to detect the amount of infrared light by facing the exterior wall 10a approximately perpendicularly (directly) with slight angle adjustments. If the camera faces the exterior wall 10a at a large inclination angle other than approximately perpendicular, the infrared light may be totally reflected, resulting in information that differs from the actual temperature being acquired. Therefore, it is desirable to face the exterior wall 10a at an angle close to directly facing it or at a slightly downward inclination angle (downward by about 5°).

[0027] Furthermore, by mounting the infrared and visible light camera 11a on the drone 11, the amount of infrared light can be detected close to the exterior wall 10a. Detecting the amount of infrared light close to the exterior wall 10a reduces the inclusion of unnecessary image parts other than the exterior wall 10a in the image, such as an image of the infinitely distant background, preventing extreme bias in the infrared distribution of the entire image and stabilizing the infrared image data. Furthermore, close-up photography increases the change in the reflection angle, making it easier to distinguish between the effects of reflected light and radiant components, improving processing accuracy.

[0028] Furthermore, by using an infrared and visible light camera 11a with a triaxial camera gimbal mounted on the drone 11, infrared images can be detected by overlapping the imaging range of the exterior wall 10a. Overlapping imaging ranges allow images of the same location to be acquired from different angles, making it possible to relatively grasp the differences between the reflected and emitted components, thereby enabling accurate anomaly detection. Furthermore, by remotely operating the infrared and visible light camera 11a, visible light images can be detected along with the detection of the amount of infrared light, and an image based on the detected amount of infrared light for each pixel and the detected visible light image can be displayed on the same screen in the image processing device 12. By simultaneously acquiring and displaying infrared images and visible light images, misidentification due to differences in viewpoint or time can be avoided, and it becomes possible to compare the correspondence between the temperature distribution and the structure.

[0029] FIG. 2 shows a schematic diagram of the electrical configuration of the inspection system in this embodiment. As shown in FIG. 2, an infrared and visible light camera 11a with a camera gimbal is attached to a drone 11. This infrared and visible light camera 11a is wirelessly connected to an image processing device 12 and is configured to be able to transmit captured infrared image data and / or visible light image data to this image processing device 12. As described above, the image processing device 12 is configured by a programmable computer device and includes an image data acquisition unit 12a that acquires the infrared image data and visible light image data sent from the infrared and visible light camera 11a, a memory unit 12b that stores the acquired image data, a mode extraction unit 12c (corresponding to the mode extraction means of the present invention) that extracts three modes (corresponding to at least one mode of the present invention) of the infrared image data stored in the memory unit 12b, and a third mode extraction unit 12c (corresponding to the mode extraction means of the present invention) that extracts three modes of the infrared image data stored in the memory unit 12b. The system includes a first-range, second-range, and third-range pixel specifying unit 12d (corresponding to the first-range, second-range, and third-range pixel specifying means of the present invention) that specifies pixels having values ​​in a first range, a second range, and a third range (corresponding to the first to Nth ranges of the present invention), a color map conversion unit 12e that colors the pixels having values ​​in the first range, the second range, and the third range, and a display unit 12f (corresponding to the display means of the present invention) that displays image data and visible light image data of the colored pixels having values ​​in the first range, the second range, and the third range. The memory unit 12b stores at least one mode extracted by the mode extracting unit 12c, pixel data having values ​​in the first range, the second range, and the third range based on each mode specified by the pixel specifying unit 12d, and color data of pixels having values ​​in the first range, the second range, and the third range that have been colored by the color map conversion unit 12e.

[0030] In this embodiment, three modes are extracted, but one, two, or four or more modes may be extracted. In this embodiment, pixels having values ​​in a first range, a second range, and a third range are identified based on each mode. However, pixels having values ​​in a first to second range or a first to fourth or more ranges may be identified based on each mode. Furthermore, in this embodiment, an infrared and visible light camera 11a capable of detecting both infrared images and visible light images is used. However, an infrared camera capable of detecting infrared images and a visible light camera capable of detecting both visible light images may be used separately. Furthermore, the exterior wall of a building may be photographed using an infrared and visible light camera, or an infrared camera and a visible light camera, installed at a fixed position, without using a drone.

[0031] FIG. 3 shows a schematic flow of the processing operation of the image processing device 12 in the inspection system of this embodiment.

[0032] First, the image data acquisition unit 12a of the image processing device 12 acquires infrared image data and visible light image data sent from the infrared and visible light camera 11a mounted on the drone 11, and stores the acquired data in a specified file in the memory unit 12b (step S1).

[0033] Next, the mode extractor 12c reads one image data (e.g., data of 640 × 512 pixels) from the infrared image data stored in the memory unit 12b and extracts the mode (step S2). To explain the process of step S2 in more detail, (1) first, the grayscale value of each pixel of the selected infrared image data is obtained, and each pixel value is converted into a score (normalized to a value between 1 and 100) to extract a score value. (2) Next, the score value is converted into a mosaic value by rounding off the last digit. For example, 63 is rounded to 60, and 48 is rounded to 50. (3) Next, the occurrence frequency of the obtained mosaic value is tallied and the mode is extracted. In this embodiment, three modes (the value with the highest, second highest, and third highest occurrence frequency) are extracted. FIGS. 4A, 5A, and 6A show examples of scores for each pixel value of the infrared image data. However, these figures only show score values ​​for 10 × 10 pixels out of the total 640 × 512 pixel data. Figures 4(B), 5(B), and 6(B) show the mosaic values ​​of each pixel converted from these scores. Figure 4(B) shows the pixel with the highest frequency of occurrence, the first mode (60), Figure 5(B) shows the pixel with the second highest frequency of occurrence, the second mode (50), and Figure 6(B) shows the pixel with the third highest frequency of occurrence, the third mode (40).

[0034] Then, pixels having score values ​​within a predetermined range above and below each extracted mode are identified (step S3 in FIG. 3). In this embodiment, pixels having score values ​​in the range of ±5, ±8, and ±10 are identified. Note that these score value ranges can be fine-tuned manually by an operator. Furthermore, the score value ranges can be set from the beginning to numerical ranges other than ±5, ±8, and ±10, and the number of ranges can be set to a number other than three. FIG. 4(C) shows pixels having score values ​​in the range of ±5 based on the first mode (60), pixels having score values ​​in the range of ±8 based on the first mode (60), and pixels having score values ​​in the range of ±10 based on the first mode (60). Figure 5(C) shows pixels having score values ​​in the range of ±5 from the second most frequent value (50), pixels having score values ​​in the range of ±8 from the second most frequent value (50), and pixels having score values ​​in the range of ±10 from the second most frequent value (50). Figure 6(C) shows pixels having score values ​​in the range of ±5 from the third most frequent value (40), pixels having score values ​​in the range of ±8 from the third most frequent value (40), and pixels having score values ​​in the range of ±10 from the third most frequent value (40).

[0035] Next, display colors are selected for the identified pixels based on color mapping, and the display colors for all selected pixels are stored in the storage unit 12b (step S4). In the color mapping, the correspondence between the pixels identified by the mode and score values ​​and the display colors can be set arbitrarily, and the definition of RGB can also be set arbitrarily. Furthermore, the display colors of unidentified pixels can also be set arbitrarily.

[0036] Thereafter, it is determined whether or not pixel coloring has been performed for all score value ranges (step S5), and if it is determined that coloring has not been performed for all score value ranges (NO), the process returns to step S3 and repeats the processing of steps S3 to S5.

[0037] If it is determined that this process has been performed for all score value ranges (YES in step S5), it is determined whether this process has been performed for all three most frequent values ​​(step S6).If it is determined that this process has not been performed for all most frequent values ​​(NO), the process returns to step S3 and repeats the processes of steps S3 to S6.

[0038] If it is determined that the process has been performed for all the most frequent values ​​(YES in step S6), the colored image data and visible light image data are read out from the memory unit 12b and displayed on the display unit 12f (step S7), and the process of Figure 3 is terminated.

[0039] The infrared image and visible light image displayed in this way make it easy to detect abnormalities (deterioration) in the exterior wall 10a of the building 10. The location of the wall surface can be identified using the visible light image, and the heat distribution on that wall surface can be easily detected using the colored infrared image. For example, if a defect or flaking occurs in the tiles or painted areas that make up the finished part of the exterior wall 10a, these areas tend to be hotter during the day than normal areas that are in close contact, so the location of the defect or flaking can be easily estimated using the colored heat distribution image. In addition, visual inspection can also be performed using the visible light image.

[0040] 7 and 8 show examples of a processed infrared image and a visible light image, respectively, when the outer wall 10a of the building 10 is inspected using this inspection system.

[0041] 7 and 8, the top row shows nine images after image processing of the infrared images, and the bottom row shows one visible light image. Of the nine images after image processing shown on the top row, the top three images are based on the pixel with the first most frequent value, which is the value with the highest frequency of occurrence, the middle three images are based on the pixel with the second most frequent value, which is the value with the next highest frequency of occurrence, and the bottom three images are based on the pixel with the third most frequent value, which is the value with the next highest frequency of occurrence. Of these top, middle, and bottom images, the left column of images is an image obtained by color-mapping pixels having score values ​​in a range of ±5 based on the mode, the middle column of images is an image obtained by color-mapping pixels having score values ​​in a range of ±8 based on the mode, and the right column of images is an image obtained by color-mapping pixels having score values ​​in a range of ±10 based on the mode.

[0042] 7 and 8, in this embodiment, pixels having scores in the ranges of ±5, ±8, and ±10 based on each of the three most frequent values ​​are color-mapped and displayed in parallel, which enhances the visual effect of the infrared amount and makes it possible to easily and clearly identify abnormal areas (floating areas and skirt steak areas). Furthermore, visible light images are displayed in parallel with these infrared images, which avoids misidentification due to differences in viewpoint or time, and makes it possible to compare the correspondence between the temperature distribution and the structure.

[0043] As explained in detail above, according to this embodiment, an infrared image of the outer wall 10a of the building 10, which is the surface to be inspected, is detected, and the detected image data is processed in a predetermined manner by a computer-based image processing device 12 to make a judgment. Therefore, the inspection of the surface to be inspected can be repeated in accordance with certain judgment criteria, and can be carried out quickly, clearly, and easily by anyone.

[0044] More specifically, according to this embodiment, the mode of infrared radiation for each pixel is extracted, and pixels having different values ​​in a range of 1 to N based on this mode are identified, and these identified pixels are displayed in different colors. In the exterior walls of buildings, etc., the difference in the amount of infrared radiation emitted from areas with and without abnormalities is very small. However, according to this embodiment, the mode (MODE) of the infrared radiation for each pixel in the image is used as the reference, and colors are displayed based on the temperature difference around it. This eliminates singular values ​​(outliers), allowing for efficient and visual identification of minute abnormalities and making it easier and more clearly visible to visually identify and grasp abnormal areas. Furthermore, according to this embodiment, the detected amount of infrared radiation for each pixel is scored to obtain an infrared radiation score, and the obtained infrared radiation score is rounded to obtain a mosaic value. At least one mode of the obtained mosaic value is obtained. This allows for quantitative handling of temperature shading (infrared radiation amount) and ensures reliable acquisition of the mode. Furthermore, according to this embodiment, multiple modes of the amount of infrared light for each detected pixel are extracted, and pixels having different numerical values ​​in the range of 1 to N are identified based on each of the multiple extracted modes.Therefore, multiple modes, which are the most characteristic score values, are extracted, and pixels having numerical values ​​in the range of 1 to N are identified for each of them, which increases the visual enhancement effect regarding the amount of infrared light and makes it possible to easily and clearly identify abnormal areas.

[0045] In this embodiment, infrared images are captured by overlapping the imaging ranges of the outer wall 10a of the building 10, which is the surface to be inspected. Overlapping imaging ranges allows infrared amounts detected at different angles for the same location, making it possible to relatively grasp the difference between the reflected and radiated components, thereby enabling accurate detection of abnormalities. Furthermore, in this embodiment, the infrared amount is detected by facing the outer wall 10a of the building 10, which is the surface to be inspected, approximately perpendicularly (directly). Facing the surface at a large inclination angle other than approximately perpendicular can result in total reflection of the infrared light, potentially resulting in information that differs from the actual temperature. However, facing the surface at an angle close to the surface to be inspected enables accurate temperature detection. Furthermore, this embodiment is configured to detect infrared amounts close to the surface to be inspected. This reduces unnecessary image areas other than the surface to be inspected, such as the background at infinity, in the image. This prevents extreme bias in the infrared score distribution throughout the image and stabilizes the mode. Furthermore, close-up imaging increases the change in the reflection angle, making it easier to distinguish between the effects of reflected light and the radiated component, thereby improving processing accuracy.

[0046] According to this embodiment, the infrared and visible light camera 11a detects both the infrared image and the visible light image. The image based on the detected infrared amount for each pixel and the detected visible light image are displayed on the same screen of the display unit 12f of the image processing device 12. This avoids misidentification due to differences in viewpoint or time, and enables comparison of the temperature distribution and the conformity of the structure. Furthermore, the visible light image allows for inspections similar to visual inspections. Furthermore, according to this embodiment, the drone 11 is used to detect the infrared amount, ensuring uniform close-range photography and a uniform photography angle (angle of view), thereby maintaining consistent evaluation criteria during analysis. Furthermore, it is possible to capture images of the surface to be inspected at a location that cannot be seen from a fixed position, such as on the ground.

[0047] The above-described embodiments are merely illustrative of the present invention and are not limiting, and the present invention can be embodied in various other modified and altered forms. Therefore, the scope of the present invention is defined only by the claims and their equivalents. [Explanation of symbols]

[0048] 10 Buildings 10a Exterior wall 11. Drone 11a Infrared and visible light camera 12 Image processing device 12a Image data acquisition unit 12b Storage section 12c Mode extractor 12d Pixel identification section 12e Colormap conversion section 12f Display section

Claims

1. An inspection system for detecting an abnormality in a surface to be inspected based on an amount of infrared radiation emitted from the surface to be inspected, an infrared image detection system mounted on a drone, the infrared image detection means detecting the amount of infrared radiation emitted from the surface to be inspected for each pixel; an extraction means extracting multiple values ​​of infrared radiation from the infrared radiation amount for each pixel detected by the infrared image detection means in order of frequency of occurrence; first to Nth (N is a natural number of 2 or greater) pixel identification means determining upper and lower multiple ranges of infrared radiation amount in advance, and identifying pixels having numerical values ​​within the multiple ranges above and below the multiple values ​​of infrared radiation amount extracted by the extraction means; and a display means displaying the pixels identified by the first to Nth pixel identification means in different colors.

2. The infrared detection inspection system according to claim 1, characterized in that the extraction means is a means for scoring the amount of infrared light for each pixel detected by the infrared image detection means to obtain an infrared light amount score, rounding the obtained infrared light amount score to obtain a mosaic value, and obtaining the infrared light amounts of the multiple values ​​of the obtained mosaic value.

3. 2. The inspection system using infrared detection according to claim 1, wherein said infrared image detection means is configured to detect infrared images in a range overlapping with the inspection surface.

4. 2. The inspection system using infrared detection according to claim 1, wherein the infrared image detection means is configured to detect the amount of infrared light while facing the surface to be inspected substantially perpendicularly.

5. 2. The inspection system using infrared detection according to claim 1, wherein said infrared image detection means is configured to detect an amount of infrared light in proximity to said surface to be inspected.

6. 2. An inspection system using infrared detection as described in claim 1, further comprising a visible light image detection means for detecting a visible light image of the surface to be inspected, and wherein the display means is configured to display on the same screen an image based on the amount of infrared light for each pixel detected by the infrared image detection means and the visible light image detected by the visible light image detection means.

Images