WO2022153852A1 - 画像解析装置、画像解析方法及びプログラム - Google Patents
画像解析装置、画像解析方法及びプログラム Download PDFInfo
- Publication number
- WO2022153852A1 WO2022153852A1 PCT/JP2021/048467 JP2021048467W WO2022153852A1 WO 2022153852 A1 WO2022153852 A1 WO 2022153852A1 JP 2021048467 W JP2021048467 W JP 2021048467W WO 2022153852 A1 WO2022153852 A1 WO 2022153852A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- temperature
- deformation
- image
- temperature change
- distribution
- Prior art date
Links
- 238000010191 image analysis Methods 0.000 title claims abstract description 53
- 238000003703 image analysis method Methods 0.000 title claims abstract description 11
- 230000008859 change Effects 0.000 claims abstract description 162
- 238000009826 distribution Methods 0.000 claims description 249
- 230000008439 repair process Effects 0.000 claims description 22
- 235000008733 Citrus aurantifolia Nutrition 0.000 claims description 10
- 235000011941 Tilia x europaea Nutrition 0.000 claims description 10
- 239000004571 lime Substances 0.000 claims description 10
- 230000007547 defect Effects 0.000 abstract description 24
- 238000001514 detection method Methods 0.000 abstract description 7
- 238000007689 inspection Methods 0.000 abstract 2
- 206010040844 Skin exfoliation Diseases 0.000 description 98
- 238000009795 derivation Methods 0.000 description 68
- 238000000034 method Methods 0.000 description 46
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 38
- 235000011389 fruit/vegetable juice Nutrition 0.000 description 26
- 238000003860 storage Methods 0.000 description 18
- 238000012545 processing Methods 0.000 description 16
- 238000013527 convolutional neural network Methods 0.000 description 15
- 238000010801 machine learning Methods 0.000 description 11
- 230000002950 deficient Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 238000011156 evaluation Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 230000002159 abnormal effect Effects 0.000 description 7
- 238000003708 edge detection Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000010521 absorption reaction Methods 0.000 description 6
- 238000001228 spectrum Methods 0.000 description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 239000000463 material Substances 0.000 description 4
- 230000015654 memory Effects 0.000 description 4
- 235000010627 Phaseolus vulgaris Nutrition 0.000 description 3
- 244000046052 Phaseolus vulgaris Species 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 239000004576 sand Substances 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000000491 multivariate analysis Methods 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000002547 anomalous effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000012850 discrimination method Methods 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000002893 slag Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N25/00—Investigating or analyzing materials by the use of thermal means
- G01N25/72—Investigating presence of flaws
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30132—Masonry; Concrete
Definitions
- the present invention relates to an image analysis device, an image analysis method and a program.
- the surface temperature of the repair mark may differ from the surface temperature of the surrounding concrete because the thermal conductivity of the repair material is different from that of the surrounding concrete.
- the infrared emissivity of the repair mark is different from that of the surrounding concrete, the apparent surface temperature of the repair mark in the infrared thermal image may be different from that of the surrounding concrete. Even if foreign matter such as free lime is attached to the surface, the actual temperature and / or apparent temperature at that part may differ from that of the surrounding concrete.
- the actual and / or apparent surface temperature may be caused by uneven color (mold, moss, release agent, water effect, etc.), joints, steps, glue, sand streaks, rust juice, rust, water leakage, surface unevenness, bean board, etc. Can produce parts that are different from the surroundings.
- the infrared survey has a problem that the surface temperature is different from the surroundings, that is, there are many erroneous detections even though there are no defects such as floats inside the structure.
- Patent Document 1 discloses the following method. Identify the factors that affect the thermal image of the structure and the relational expression of multivariate analysis that uses the information of this factor to determine the probability that the anomalous part extracted from the thermal image of the structure contains defects. do. Then, the structure is photographed with an infrared camera to acquire a thermal image, an abnormal part having a temperature different from that of the surroundings is extracted from the thermal image, and the factor information in the abnormal part is discriminated, and then the information of the discriminated factor is quantified. , Apply the numerical value to the relational expression of multivariate analysis, and find the probability that the extracted abnormal part contains a defect.
- the "surface condition" means the presence or absence of deformation on the surface of the concrete structure such as uneven color, non-land surface, and free lime. Since most of the false positives (abnormal parts where the surface temperature is different from the surroundings even though there are no defects inside the structure) are caused by the deformation of the structure surface, it is determined whether or not the structure surface is deformed. Things are considered to be effective in reducing false positives.
- the present invention has been made in view of such circumstances, and an object of the present invention is to provide an image analysis device, an image analysis method, and a program capable of reducing false detection of defective parts.
- the image analysis device is an image analysis device including a processor, and the processor acquires an infrared thermal image of a structure to be inspected and obtains a visible image of the structure to be inspected.
- the temperature change is determined from the acquired infrared thermal image, and the temperature change is determined based on at least the temperature change information obtained from the infrared thermal image and the surface deformation information obtained from the visible image. Estimate the cause.
- the temperature deformation information includes information obtained from the temperature distribution and / or the temperature distribution of the infrared thermal image regarding the temperature change.
- the temperature change information includes information on the shape and / or size of the temperature change.
- the surface deformation information includes information obtained from the luminance distribution and / or the luminance distribution of the visible image.
- the surface deformation information includes at least one information of the type, shape and position of the surface deformation.
- the processor estimates the cause of the temperature deformation based on the similarity between the temperature deformation information and the surface deformation information.
- the processor estimates that the cause of the temperature deformation is the surface deformation
- the processor estimates the temperature distribution due to the surface deformation and reduces it from the infrared thermal image.
- the similarity includes a partial similarity.
- the processor determines that the surface deformation corresponding to the temperature deformation is a crack or peeling, evaluates the similarity between the temperature deformation information and the surface deformation information, and at least a part thereof. If they are similar, it is presumed that the cause of the temperature change is floating with cracks or peeling.
- the processor determines that the surface deformation corresponding to the temperature change is a crack or peeling, and is within the magnitude of the temperature change and / or near the boundary of the temperature change. Evaluate the presence or absence of surface deformation, and if there is, presume that the cause of the temperature deformation is floating with cracks or peeling.
- the processor estimates the cause of the temperature change based on the inclination of the temperature at the boundary of the temperature change.
- the surface deformation includes at least one of repair marks, free lime, joints, steps, cracks and peeling.
- the visible image is an image obtained by imaging the reflection intensity distribution in two or more different wavelength ranges in the wavelength range of visible light.
- the image analysis device further includes a display device, and the processor displays the estimation result of the cause of the temperature change on the display device.
- the image analysis method includes a step of acquiring an infrared thermal image of the structure to be inspected, a step of acquiring a visible image of the structure to be inspected, and a temperature change from the infrared thermal image.
- the program to be executed by the computer according to the sixteenth aspect includes a step of acquiring an infrared thermal image of the structure to be inspected, a step of acquiring a visible image of the structure to be inspected, and infrared heat.
- the cause of the temperature change is determined based on the step of determining the temperature change from the image and at least the temperature change information obtained from the infrared thermal image and the surface change information obtained from the visible image.
- FIG. 1 is a block diagram showing an example of the hardware configuration of the image analysis device.
- FIG. 2 is a block diagram showing a processing function realized by the CPU.
- FIG. 3 is a diagram showing information stored in the storage unit.
- FIG. 4 is a flow chart showing an analysis method using an image analysis device.
- FIG. 5 is an infrared thermal image and a visible image of a structure to which rust juice is attached.
- FIG. 6 is an image showing the deformed shape with respect to the image of FIG. 5 in binary.
- FIG. 7 is an infrared thermal image and a visible image of the structure including the repaired portion.
- FIG. 8 is an image showing the deformed shape of the image of FIG. 7 in binary values.
- FIG. 5 is an infrared thermal image and a visible image of a structure to which rust juice is attached.
- FIG. 6 is an image showing the deformed shape with respect to the image of FIG. 5 in binary.
- FIG. 7 is an infrared thermal image and a
- FIG. 9 is an infrared thermal image and a visible image of a structure including a peeled portion.
- FIG. 10 is an infrared thermal image and a visible image of a structure including two peeled parts.
- FIG. 11 is an image showing the deformed shape of the image of FIG. 9 in binary values.
- FIG. 12 is an image showing the deformed shape with respect to the image of FIG. 10 in binary.
- FIG. 13 is an infrared thermal image and a visible image of a structure including a cracked portion.
- FIG. 14 is an infrared thermal image and a visible image of a structure including another cracked portion.
- FIG. 15 is an image showing the deformed shape with respect to the image of FIG. 13 in binary.
- FIG. 16 is an image showing the deformed shape with respect to the image of FIG. 14 in binary.
- FIG. 17 is a diagram showing an example of a display result in which the estimation result is displayed on the display device.
- the image analysis device is an image analysis device including a processor, and the processor acquires an infrared thermal image of a structure to be inspected and a visible image of the structure to be inspected. , Determine the temperature change based on the infrared thermal image, derive the temperature change information, derive the surface change information from the surface change based on the visible image of the part corresponding to the temperature change, and derive the temperature change information. And the cause of the temperature deformation is estimated based on the surface deformation information.
- the present inventor has found the following as a result of diligent studies on reduction of erroneous detection of defective portions, and has reached the present invention.
- the present inventor compared and investigated the infrared thermal image obtained by photographing the concrete structure and the visible image, and discriminated the surface deformation (color unevenness, joints, steps, rust juice, etc.) identified in the visible image and the infrared thermal image.
- temperature deformation The relationship of temperature deformation (hereinafter, the part where the surface temperature is locally different from the surroundings in the infrared thermal image is called temperature deformation) differs depending on the thermal environment such as the position in the structure and the shooting time, that is, the concrete surface. It was found that the effect of surface deformation on the surface temperature differs depending on the thermal environment such as the position in the structure and the shooting time.
- the present inventor thinks that this difference is due to the fact that there was solar radiation on the wall balustrade in the sunny daytime, and the amount of absorption was different between the part of color unevenness and rust juice and the other part.
- the effect of surface deformation on the surface temperature differs depending on the thermal environment, so it is necessary to determine the effect (determine the cause of the temperature deformation), and therefore the positions of the surface deformation and temperature deformation.
- the present inventor has found that it is indispensable to analyze relationships such as shape and shape.
- FIG. 1 is a block diagram showing an example of the hardware configuration of the image analysis device according to the embodiment.
- the image analysis device 10 shown in FIG. 1 a computer or a workstation can be used.
- the image analysis device 10 of this example mainly includes an input / output interface 12, a storage unit 16, an operation unit 18, a CPU (Central Processing Unit) 20, a RAM (Random Access Memory) 22, a ROM (Read Only Memory) 24, and the like. It is composed of a display control unit 26.
- a display device 30 is connected to the image analysis device 10, and a display device 30 is displayed under the command of the CPU 20 under the control of the display control unit 26.
- the display device 30 is composed of, for example, a monitor.
- the input / output interface 12 (input / output I / F in the figure) can input various data (information) to the image analysis device 10.
- the data stored in the storage unit 16 is input via the input / output interface 12.
- the CPU (processor) 20 collectively controls each unit by reading various programs stored in the storage unit 16 or the ROM 24 or the like, expanding them into the RAM 22, and performing calculations. Further, the CPU 20 reads a program stored in the storage unit 16 or the ROM 24, performs calculations using the RAM 22, and performs various processes of the image analysis device 10.
- the infrared camera 32 shown in FIG. 1 photographs the structure 36 to be inspected and acquires an infrared thermal image of the surface of the structure.
- the visible camera 34 photographs the structure 36 to be inspected and acquires a visible image of the structure 36.
- the image analysis device 10 can acquire an infrared thermal image from the infrared camera 32 via the input / output interface 12. Further, the image analysis device 10 can acquire a visible image from the visible camera 34 via the input / output interface 12. The acquired infrared thermal image and visible image can be stored in the storage unit 16, for example.
- FIG. 2 is a block diagram showing a processing function realized by the CPU 20.
- the CPU 20 includes an infrared thermal image acquisition unit 51, a visible image acquisition unit 53, a temperature deformation information derivation unit 55, a surface deformation information derivation unit 57, a cause estimation unit 59, and an information display unit 61.
- the specific processing functions of each part will be described later. Since the infrared thermal image acquisition unit 51, the visible image acquisition unit 53, the temperature deformation information derivation unit 55, the surface deformation information derivation unit 57, the cause estimation unit 59, and the information display unit 61 are a part of the CPU 20, the CPU 20 is a part of each unit. It can also be referred to as executing the processing of.
- the storage unit 16 stores data and programs for operating the image analysis device 10, such as an operating system and a program for executing an image analysis method. Further, the storage unit 16 stores information and the like used in the embodiments described below.
- FIG. 3 is a diagram showing information and the like stored in the storage unit 16.
- the storage unit 16 is a memory composed of various semiconductor memories such as a CD (Compact Disk), a DVD (Digital Versatile Disk), a hard disk (Hard Disk), and a flash memory.
- the storage unit 16 mainly stores the infrared thermal image 101, the temperature deformation information 102, the visible image 103, and the surface deformation information 104.
- the infrared thermal image 101 is an image taken by the infrared camera 32, detects infrared radiant energy radiated from the structure 36, converts the infrared radiant energy into temperature, and heats the surface of the structure. It is an image showing the distribution.
- the temperature change information 102 is information obtained from the temperature distribution and / or the temperature distribution of the infrared thermal image 101 regarding the temperature change.
- the visible image 103 is an image taken by the visible camera 34 and shows the distribution of the reflection intensity of visible light from the surface of the structure 36.
- a visible image is composed of an RGB image in which reflection intensity distributions in three different wavelength ranges are imaged in the wavelength range of visible light, that is, each pixel has color information (RGB signal value).
- RGB signal value RGB signal value
- the brightness of the visible image 103 which will be described later, indicates the signal value of the visible image 103, and the brightness of each pixel of the visible image 103 is at a position on the surface of the structure 36 to which each pixel corresponds. It reflects the reflection intensity of visible light.
- the surface deformation information 104 is information on the surface deformation of the portion corresponding to the temperature change in the visible image 103, and is information obtained from the luminance distribution and / or the luminance distribution of the visible image 103.
- the operation unit 18 shown in FIG. 1 includes a keyboard and a mouse, and the user can cause the image analysis device 10 to perform necessary processing via these devices.
- the display device 30 can function as an operation unit.
- the display device 30 is a device such as a liquid crystal display, and can display the result obtained by the image analysis device 10.
- FIG. 4 is a flow chart showing an image analysis method using the image analysis device 10.
- the image analysis method includes an infrared thermal image acquisition step (step S1), a visible image acquisition step (step S2), a temperature deformation information derivation step (step S3), and surface deformation information derivation. It includes a step (step S4), a cause estimation step (step S5), and an estimation result display step (step S6).
- the infrared thermal image acquisition unit 51 acquires an infrared thermal image obtained by photographing the structure 36 to be inspected (infrared thermal image acquisition step: step S1).
- the infrared thermal image is an infrared thermal image 101 stored in the storage unit 16.
- the infrared thermal image 101 is acquired from the storage unit 16 by the infrared thermal image acquisition unit 51.
- the infrared thermal image acquisition unit 51 acquires the infrared thermal image 101 from the outside.
- the infrared thermal image acquisition unit 51 can acquire the infrared thermal image 101 through the network via the input / output interface 12, and the infrared thermal image acquisition unit 51 can acquire infrared rays from the infrared camera 32 via the input / output interface 12.
- the thermal image 101 can be acquired.
- the visible image acquisition unit 53 acquires a visible image obtained by photographing the structure to be inspected (visible image acquisition step: step S2).
- the visible image is a visible image 103 stored in the storage unit 16.
- the visible image 103 is acquired from the storage unit 16 by the visible image acquisition unit 53.
- the visible image acquisition unit 53 acquires the visible image 103 from the outside.
- the visible image acquisition unit 53 can acquire the visible image 103 through the network via the input / output interface 12, and the visible image acquisition unit 53 can acquire the visible image 103 from the visible camera 34 via the input / output interface 12. You can get it.
- the temperature change information derivation unit 55 determines the temperature change based on the infrared thermal image 101 and derives the temperature change information 102 (temperature change information derivation step: step S3).
- the temperature change information derivation unit 55 derives the temperature change information 102 from the infrared thermal image 101 by locally cohesively determining and extracting a portion having a difference in surface temperature from the surroundings as a temperature change.
- a portion of the surface of the structure of the infrared thermal image 101 that exceeds a predetermined temperature difference from the average temperature (a portion having a higher temperature than the surroundings when the temperature rises such as in the daytime, and a temperature higher than the surroundings when the temperature drops such as at night).
- the lower part can be determined as a temperature change and extracted.
- a portion where the surface temperature is different from that of the surroundings and is spatially connected in a cohesive manner, or a portion where the surface temperature is distributed at a distance closer than a predetermined value even if they are not connected can be determined as one temperature change.
- the temperature change is not necessarily the part where the actual surface temperature is different from the surroundings. That is, even if the actual surface temperature is the same as the surroundings, the infrared emissivity is different, so that the surface temperature is different from the surroundings in the infrared thermal image 101, and it may be determined that the temperature is deformed.
- the infrared thermal image 101 may be the infrared thermal image itself obtained by photographing the target concrete structure 36 with the infrared camera 32. Further, in order to facilitate the determination of the temperature change and / or the derivation of the temperature change information, the original infrared thermal image 101 may be processed. For example, the surface temperature of the structure 36 often has an inclination (temperature gradient) due to a partial difference in the amount of heat received on the surface of the structure 36 or the amount of heat radiated from the surface of the structure 36. .. Therefore, the original infrared thermal image 101 may be processed so as to reduce the temperature gradient, and the temperature change may be determined and extracted from the processed image to derive the temperature change information 102.
- the temperature change information 102 is information representing the temperature distribution (spatial distribution of temperature) of the temperature change, and extends to at least the range including the entire temperature change (the temperature change extends to the edge of the infrared thermal image 101). In the case, it is information representing the temperature distribution (range including the end). In the cause estimation step (step S5) described later, it is preferable that the information represents the temperature distribution in as wide a range as possible for the cause estimation.
- the temperature deformation information 102 may be the temperature distribution itself in the original infrared thermal image 101, or may be a coarsely quantized distribution of the original temperature distribution, for example, a binarized, ternary, or quaternized distribution.
- the temperature change information 102 may be information indicating the shape of the temperature change.
- the binarized temperature distribution can be said to be information representing the shape of the temperature deformation.
- the temperature change information may be information indicating the size when the temperature change is approximated by a rectangle or an ellipse.
- the temperature change information 102 is information obtained from the temperature distribution and / or the temperature distribution of the infrared thermal image 101 regarding the temperature change.
- the surface deformation information deriving unit 57 derives the surface deformation information 104 based on the visible image 103 (surface deformation information deriving step: step S4).
- Deformation has the meaning of "a state changed from the initial state” or "a state different from the normal state", but in the embodiment, particularly, repair marks, foreign matter adhesion such as free lime, color unevenness (mold, rust, etc.) It affects the temperature of the concrete surface in infrared thermal images such as release agent, water effect, etc.), joints, steps, slag, sand streaks, rust juice or rust, water leakage, surface unevenness, bean plate, etc., causing temperature deformation.
- the obtained surface condition is called “surface deformation”. Cracks and peeling are also called “surface deformation”.
- the surface deformation information 104 is information indicating the presence / absence, type, shape, position, etc. of the surface deformation, and is obtained from the brightness distribution (spatial distribution of brightness) of the visible image 103. Further, in the case of surface deformation such as color unevenness, rust juice or rust in the visible image 103, the brightness distribution is effective in estimating the cause of the temperature deformation.
- the brightness of each pixel of the visible image 103 reflects the reflection intensity of visible light at the position on the surface of the structure 36 to which each pixel corresponds.
- the difference in brightness reflects the difference in reflection intensity with respect to visible light that uniformly illuminates the surface of the structure 36, that is, the difference in reflectance. That is, it reflects the difference in the absorption rate of visible light. Therefore, the brightness distribution on the surface of the structure 36 in the visible image 103 reflects the distribution of the absorption rate of visible light such as sunlight that uniformly illuminates the surface of the structure 36, that is, the distribution of the absorption amount, and similarly.
- the surface deformation information 104 may be information representing the brightness distribution of the surface deformation, the brightness distribution of the visible image 103 itself, or a coarsely quantized distribution of the brightness distribution, for example, a binary value. It may be a quantized or quaternized distribution.
- the surface deformation information 104 is information on the surface deformation of the portion corresponding to the temperature change in the visible image 103, and is information obtained from the luminance distribution and / or the luminance distribution of the visible image 103.
- the luminance distribution of the location corresponding to the temperature deformation is the presence / absence, type, shape, position and luminance distribution of the surface deformation of the location. (Information to be included), and the luminance distribution may be the surface deformation information 104.
- the peeled part has a distinctly different brightness than the original concrete surface, and the texture, contrast, and frequency spectrum of the brightness distribution are also different.
- the portion exceeding the difference between the above is judged to be peeled and extracted. Since there is a step between the peeled part and the original concrete surface and the step part is dark (low brightness), the darkness of the boundary part (low brightness) can also be an effective feature for determining peeling. ..
- the visible image 103 is locally cohesive, and the part where the brightness is different from the surroundings and / or the part where the texture of the brightness distribution is different and / or the part where the frequency spectrum of the brightness distribution is different and / Alternatively, a portion having a different contrast in the luminance distribution is determined and extracted.
- a portion exceeding a predetermined brightness difference from the average brightness is determined to be another surface deformation and extracted.
- One surface deformation is a part where one or more of the brightness, the texture of the brightness distribution, the frequency spectrum, and the contrast are different from the surroundings, and the part which is spatially connected in a cohesive manner or the part which is distributed at a distance closer than a predetermined even if it is not connected. And extract.
- step S5 it is not always necessary to finely determine the type of surface deformations in the cause estimation step (step S5) described later (repair marks, free lime, color unevenness, joints, steps, rust, sand). Streaks, rust juice or rust, water leaks, surface irregularities, bean boards, etc.). However, if the type of surface deformation is determined, the cause can be estimated more appropriately in the cause estimation step (step S5).
- the surface deformation in which the difference in the amount of visible light absorbed such as color unevenness, rust juice or rust from the surrounding concrete is the main cause of the difference in surface temperature from the surroundings in the infrared thermal image 101 is described above.
- the luminance distribution reflects the distribution of the amount of visible light absorbed, the luminance distribution is effective for the cause estimation in the cause estimation step (step S5) as the surface deformation information 104.
- the case of surface deformation in which the luminance distribution is not effective for estimating the cause is as follows.
- the difference in thermal conductivity and infrared emissivity from the surrounding concrete such as repair marks and free lime is the main cause of the difference in surface temperature from the surroundings in the infrared thermal image 101.
- the shape is effective as the surface deformation information 104 in the cause estimation step (step S5). Therefore, it is preferable to finely discriminate the types of other surface deformations.
- the type can be finely discriminated based on the characteristics such as the average brightness, contrast, brightness dispersion, texture, frequency spectrum, and shape of the extracted surface deformation.
- the visible image 103 usually consists of an RGB image in which reflection intensity distributions in three different wavelength ranges are imaged in the wavelength range of visible light. Cracks, peeling, and other surface deformations may be determined and extracted from any of the RGB brightness distributions, but surface deformations such as rust juice or rust have a difference in brightness (difference in reflection intensity) from concrete. Since it differs greatly depending on RGB, it is preferable to determine and extract from the luminance distribution of the channel having the largest difference. For example, in the case of rust juice or rust, the brightness of B is particularly low as compared with concrete, that is, the difference in absorption amount, which is the difference in reflection intensity in the wavelength range of B, is large, so it is preferable to judge from the brightness distribution of B.
- the channels having a large difference in brightness from concrete differ depending on the type of surface deformation.
- the difference in brightness of the R channel is particularly large, contrary to rust juice or rust. Therefore, it is preferable to evaluate the magnitude of the variation in brightness in the brightness distribution of each of the RGB channels, and determine and extract cracks, peeling, and other surface deformations from the channel with the largest variation.
- the value obtained by dividing the standard deviation of the brightness at the location corresponding to the temperature change by the average brightness of the concrete, that is, the coefficient of variation, is obtained, and cracks, peeling, and peeling are performed from the channel having the largest coefficient of variation. And other surface deformations may be determined and extracted.
- the average brightness of concrete the average value of the brightness in the portion corresponding to the temperature change may be adopted, or the average value of the brightness in a wider range including the portion corresponding to the temperature change may be adopted. ..
- the type of other surface deformation When the type of other surface deformation is finely discriminated, it can be discriminated based on the characteristics such as the average brightness, contrast, luminance dispersion, texture, and frequency spectrum of each of the extracted RGB channels of the surface deformation. Similarly, when there are two or four or more types of visible image 103, cracks, peelings and other surface deformations are determined and extracted in the channels with the largest variation, and two or four or more types of each are used. Other types of surface deformation can be finely discriminated based on characteristics such as average brightness and contrast of channels.
- the portion corresponding to the temperature change is larger than the spatial range corresponding to the temperature change when the temperature change is determined from the infrared thermal image 101 in the temperature change information derivation step (step S3). It is a wide range.
- the cause estimation step (step S5) surface deformation information in a wider range than the range corresponding to the temperature deformation is required in the analysis of the cause estimation of the temperature deformation. In particular, for peeling, it is necessary to analyze the relationship between the temperature change and the peeling at a different position.
- step S1 the order of the infrared thermal image acquisition step (step S1), the visible image acquisition step (step S2), the temperature deformation information derivation step (step S3), and the surface deformation information derivation step (step S4).
- these orders can be changed as appropriate.
- the cause estimation unit 59 estimates the cause of the temperature change based on the temperature change information 102 and the surface change information 104 (cause estimation step: step S5).
- step S4 Since the process of estimating the cause is slightly different depending on the type of the surface deformation information 104 derived in the surface deformation information derivation step (step S4), the first aspect, the second aspect, and the third aspect will be described respectively.
- ⁇ First aspect> As a first aspect, a case where the luminance distribution of the portion corresponding to the temperature deformation is used as the surface deformation information in the surface deformation information derivation step (step S4) will be described. The method of estimating the cause of cracks and peeling and other surface deformations will be described.
- a range corresponding to the temperature deformation derived in the temperature deformation information derivation step (step S3) or a slightly wide range including the temperature deformation is extracted. ..
- the similarity between the luminance distribution and the temperature distribution in this range is evaluated, and if they are similar, the cause of the temperature deformation is presumed to be this surface deformation. On the other hand, if they are not similar, it is presumed that the surface deformation is other than this, and it is presumed that it is an internal defect such as floating.
- the similarity between the two distributions There are many methods for evaluating the similarity between the two distributions. For example, after standardizing each distribution so that the minimum and maximum values of the luminance distribution and the temperature distribution are the same, the following equation (1) Similarity can be evaluated by calculating the Euclidean distance given in. For example, the equation (1) may be calculated and determined to be similar when the distance is equal to or less than a predetermined value (the closer it is to 0, the more similar it can be determined).
- the relationship between the brightness level and the temperature level may be the same or vice versa depending on the type of surface deformation and the timing of shooting, so it is necessary to calculate both cases.
- the Euclidean distance is calculated for both the case where the brightness distribution is the same and the distribution where the brightness is reversed (for example, the distribution obtained by subtracting the original brightness value from 255), and when either distance is less than or equal to the predetermined value. Judge as similar.
- v (x, y) is the pixel value of the coordinates (x, y) in the normalized luminance distribution
- t (x, y) is the pixel value of the coordinates (x, y) in the normalized temperature distribution. Represents a value.
- the product-moment correlation coefficient of Pearson given by the following equation (2) is calculated, and the absolute value of the correlation coefficient is equal to or more than a predetermined value (the closer it is to 1, the more similar it is). Then, it may be determined that it is similar to the case of).
- the similarity can be evaluated regardless of whether the relationship between the brightness level and the temperature level is the same or vice versa.
- v (x, y) is the pixel value of the coordinates (x, y) in the brightness distribution
- v_ave is the average value of the brightness distribution
- t (x, y) is the coordinate (x, y) in the temperature distribution.
- the pixel value is represented by t_ave, which represents the average value of the temperature distribution.
- FIG. 5 is an image of a concrete structure photographed in the daytime
- FIG. 5 (A) is a visible image
- FIG. 5 (B) is an infrared thermal image.
- rust juice adheres to the surface of the photographed concrete structure to be inspected.
- step S3 it is determined and extracted from the infrared thermal image (FIG. 5 (B)) as a temperature change having a higher surface temperature than the surroundings. Rust juice is reflected in the visible image (Fig. 5 (A)) of the part corresponding to the temperature change, and the brightness distribution and the temperature distribution of the temperature change are similar (relationship between the high and low brightness and the high and low temperature). The opposite), it is presumed that the cause of the temperature deformation is this surface deformation (rust juice).
- FIG. 6 shows the deformed shape with respect to the image of FIG. 5 as binary values, and the pixel with the deformity is represented by 1 (white) and the pixel without the deformity is represented by 0 (black).
- FIG. 6 (A) shows the shape of the surface deformation (rust juice) derived from the visible image (FIG. 5 (A)), and
- FIG. 6 (B) shows the temperature derived from the infrared thermal image (FIG. 5 (B)). It represents a deformed shape. Comparing FIG. 6 (A) and FIG. 6 (B), it is determined that they are similar.
- FIG. 7 is an image of a concrete structure photographed in the daytime
- FIG. 7 (A) is a visible image
- FIG. 7 (B) is an infrared thermal image.
- the surface of the concrete structure to be inspected includes the repaired part.
- step S3 it is determined and extracted from the infrared thermal image (FIG. 7 (B)) as a temperature change having a higher surface temperature than the surroundings.
- the repair mark is shown in the visible image (Fig. 7 (A)) of the part corresponding to the temperature change, and the brightness distribution and the temperature distribution of the temperature change are similar (relationship between the high and low brightness and the high and low temperature). The same), it is presumed that the cause of the temperature deformation is this surface deformation (repair mark).
- FIG. 8 shows the deformed shape with respect to the image of FIG. 7 as binary values, and similarly, the pixel with the deformity is represented by 1 (white) and the pixel without the deformity is represented by 0 (black).
- FIG. 8 (A) shows the shape of the surface deformation (repair mark) derived from the visible image (FIG. 7 (A)), and
- FIG. 8 (B) shows the temperature derived from the infrared thermal image (FIG. 7 (B)). It is a deformed shape. Comparing FIG. 8 (A) and FIG. 8 (B), it is determined that they are similar.
- the temperature gradient at the boundary of the temperature change is steep.
- the temperature deformation caused by the surface deformation is characterized by a steep temperature gradient at the boundary (however, it depends on the type of surface deformation).
- internal defects such as floating
- heat diffuses between the internal defects and the surface so the temperature gradient at the boundary of temperature deformation due to the internal defects is gentle, and the deeper the internal defects, that is, the internal defects. The wider the space between the surface and the surface, the more heat is diffused, and the gentler the temperature gradient becomes.
- the cause of the temperature change may be estimated based on this feature (a feature in which the temperature slope is steep at the boundary of the temperature change).
- the temperature distribution and the luminance distribution are similar in a range corresponding to the temperature deformation derived in the temperature deformation information derivation step (step S3) or a slightly wider range, and / or the temperature at the boundary of the temperature deformation.
- a predetermined value a preset threshold value
- the cause of the temperature deformation may be presumed to be the surface deformation.
- the average value of the maximum slopes of each point on the boundary may be obtained. For example, at each point (x, y) on the boundary, the equation (3) may be calculated and the average value thereof may be obtained.
- the slope of temperature can be indirectly evaluated by using the spatial second derivative of temperature. Any method may be used to evaluate the slope of temperature.
- This feature is particularly noticeable in "repair marks,” “joints,” and “steps,” and is effective in estimating the cause.
- FIGS. 9 and 10 are images of a concrete structure taken in the daytime.
- FIG. 9A is a visible image and
- FIG. 9B is an infrared thermal image.
- FIG. 10A is a visible image
- FIG. 10B is an infrared thermal image.
- the photographed concrete structure to be inspected includes a peeled portion on the surface.
- a portion having a high surface temperature (a portion having a light color in the infrared thermal image) can be seen adjacent to the peeled portion. It is considered that this part is in a floating state due to air entering from the peeled part. In this way, in the part where the surface is peeled off, there are many parts where air enters and floats from there.
- step S3 From the temperature distribution derived in the temperature deformation information derivation step (step S3) and the brightness distribution derived in the surface deformation information derivation step (step S4), floating (or peeling with floating) with peeling is performed in this way. Can be judged. Specifically, the similarity between the temperature distribution and the luminance distribution is evaluated for each of the whole and the part of the temperature deformation, and when the whole is not similar but partially similar, it is determined that the float is accompanied by peeling.
- the evaluation of the similarity of the entire temperature variation is performed by extracting the range corresponding to the temperature variation from the luminance distribution and evaluating the similarity between the luminance distribution and the temperature distribution in this range.
- a predetermined range centered on each point on the boundary of the temperature-deformed is extracted from the brightness distribution based on the shape of the temperature-deformed, and the brightness distribution of each range is performed. It is done by evaluating the similarity between the temperature distribution and the temperature distribution. If any part is similar, it is judged to be a float with peeling. Since the shape of the boundary (step) of the peeled portion in the luminance distribution and the shape of the boundary of the temperature deformation in the temperature distribution match, it can be determined by such an evaluation. As described above, the evaluation of similarity is performed so as to cover the case where the relationship between the high and low brightness of the luminance distribution and the high and low temperature of the temperature distribution is the same or vice versa.
- the surface temperature is lower than the surroundings at the locations separated from the infrared thermal images of FIGS. 9 and 10.
- the surface temperature at the peeled spot is higher than that of the surroundings. It is considered that the reason why the surface temperature of the peeled portion is different from that of the surroundings is that the peeled portion is deeper than the surroundings. Since the surface temperature of the peeled portion is different from that of the surroundings, it is determined to be a temperature change and extracted. Temperature deformation caused by peeling can be distinguished from internal defects such as floating and other surface deformations because the temperature relationship with the surroundings is different.
- the peeled part of the luminance distribution and the part different from the surroundings in the temperature distribution coincide with each other, the overall temperature distribution and the luminance distribution of the temperature deformation are clearly similar. Therefore, since peeling and other surface deformations can be distinguished from the characteristic brightness distribution accompanied by a dark portion due to a step, that is, a portion having low brightness, it can be easily determined that the cause of this temperature deformation is peeling. Note that in the case of a float with peeling, the temperature change and the peeling are adjacent and the positions are different.
- the temperature gradient at the boundary of the temperature deformation is steep, so this feature is presumed to be the cause. Can be used for.
- the temperature inclination is steep only in the portion adjacent to the peeling in the boundary of the temperature deformation, and the temperature inclination is gentle in the other portions as in the case of normal floating. Therefore, in the case of a float accompanied by peeling, the temperature slope at the entire boundary of the temperature deformation is gentle, but when the temperature slope at a part of the boundary is steep, it may be determined that the float is accompanied by peeling.
- the temperature is determined for the whole and part of the temperature deformation. Evaluate the similarity between the distribution and the brightness distribution, and evaluate the slope of the temperature distribution when the whole is not similar but partially similar, and the whole and the part of the boundary of the temperature deformation are evaluated, and the total slope is When it is equal to or less than a predetermined value and the inclination is partially equal to or more than a predetermined value, the cause of the temperature change may be presumed to be floating accompanied by peeling.
- the maximum slope that is, the magnitude of the temperature distribution gradient vector
- the average value (the average value of the maximum slopes of all the points on the boundary of the temperature deformation) may be obtained.
- the slope of the temperature at the boundary of the temperature deformation can be obtained by extracting a predetermined range centered on each point on the boundary and finding the average value of the maximum slopes of all the points on the boundary included in the extracted predetermined range. good.
- the part similar to the luminance distribution and the part where the temperature slope of the boundary is steep are the same. That is, in the temperature distribution of the temperature deformation, the temperature distribution and the brightness distribution are similar in the portion adjacent to the peeling, and the inclination of the temperature at the boundary is steep. Therefore, for each part of the temperature deformation, the similarity with the luminance distribution and the evaluation of the temperature slope are performed at the same time, and when there is a part similar to the luminance distribution and the temperature slope is steep, the temperature deformation A method of presuming that the cause of the above is floating accompanied by peeling is preferable.
- the temperature slope may be standardized by an absolute value or the like.
- the similarity with the luminance distribution is evaluated including the feature that the temperature slope at the boundary is steep, but as described above.
- the temperature gradient may be evaluated.
- FIG. 11 shows the deformed shape with respect to the image of FIG. 9 in binary
- FIG. 12 shows the deformed shape with respect to the image of FIG. 10 in binary
- Pixels with deformation are represented by 1 (white)
- pixels without deformation are represented by 0 (black).
- FIG. 11 (A) shows the shape of the surface deformation (peeling) derived from the visible image (FIG. 9 (A))
- FIG. 11 (B) shows the temperature variation derived from the infrared thermal image (FIG. 9 (B)). It represents the shape of the shape.
- FIG. 11 when FIG. 11 (A) and FIG. 11 (B) are compared, they are not similar as a whole, but they are adjacent to the peeling at the upper right part of the temperature change, and are adjacent to the peeling. Since the shape of the boundary between the temperature change and the peeling is similar in the part where the temperature changes, it can be presumed that the floating is accompanied by the peeling.
- FIG. 12 (A) shows the shape of the surface deformation (peeling) derived from the visible image (FIG. 10 (A)), and FIG. 12 (B) shows the temperature variation derived from the infrared thermal image (FIG. 10 (B)). It represents the shape of the shape.
- FIG. 12 (A) and FIG. 12 (B) are compared, there is a small peeling on the left side in addition to the peeling on the right side, both of which are adjacent to each other due to the temperature change and are adjacent to each other. Since the shape is similar to the temperature deformation in the part, it can be estimated that both of the two temperature deformations are floating with peeling.
- FIG. 13 and 14 are images of concrete structures taken in the daytime.
- FIG. 13 (A) is a visible image
- FIG. 13 (B) is an infrared thermal image
- FIG. 14 (A) is a visible image
- FIG. 14 (B) is an infrared thermal image.
- 13 and 14 include cracks on the surface of the photographed concrete structure to be inspected.
- the floating with cracks (or cracks with floating) can be determined.
- a known edge is detected in the temperature distribution, and the boundary of the temperature deformation is extracted.
- edge detection such as the Sobel method, the Laplacian method, and the Canny method.
- the luminance distribution in a predetermined range centered on each point on the boundary of the temperature deformation and the distribution of the boundary extracted of the temperature deformation are extracted and the similarity is evaluated. .. If any part of the boundary is similar, it is determined that there is a crack along the boundary at that part, that is, a float with a crack.
- the luminance distribution in a predetermined range centered on each point inside the temperature deformation and the edge-extracted distribution of the temperature deformation are extracted and the similarity is evaluated. do.
- the evaluation of similarity is performed so as to cover the case where the values of the two distributions to be evaluated are the same or the opposite.
- the similarity may be evaluated in consideration of the fact that the brightness of the crack is lower than that of the surroundings in the brightness distribution.
- the temperature gradient at some boundaries is steeper than the temperature deformation caused by internal defects such as floating. That is, the temperature slope is steep at the cracked portion along the boundary of the temperature deformation. In addition, the temperature inclination is steep even in the cracked part inside the temperature deformation. Therefore, this feature may also be used to determine cracked floats. For example, after performing edge detection on the temperature distribution and extracting the boundary of the temperature change and the part where the temperature change is abrupt inside, the edge size is equal to or more than the predetermined value, that is, the temperature slope is equal to or more than the predetermined value. Extract the steep part.
- the cause of the temperature deformation is other than the floating with cracks, and it is presumed that it is an internal defect such as a floating.
- the similarity between each part and the brightness distribution is evaluated, and when it is similar in any part (there is a linear part with a steep temperature slope, and the linear part). If the shape of the portion matches the shape of the crack in the luminance distribution), the cause of the temperature change may be presumed to be a floating with a crack.
- the temperature slope may be standardized by the absolute value of the temperature difference.
- the visible image usually consists of an RGB image.
- the cause may be estimated from any of the brightness distributions of RGB, but it is preferable to evaluate the magnitude of the variation in brightness in the brightness distribution of each channel of RGB and estimate the cause from the channel having the largest variation. The same applies when there are two types or four or more types of visible images.
- FIG. 15 shows the deformed shape with respect to the image of FIG. 13 in binary
- FIG. 16 shows the deformed shape with respect to the image of FIG. 14 in binary. Pixels with deformation are represented by 1 (white), and pixels without deformation are represented by 0 (black).
- FIG. 15 (A) shows the shape of the surface deformation (crack) derived from the visible image (FIG. 13 (A)), and FIG. 15 (B) shows the temperature change derived from the infrared thermal image (FIG. 13 (B)). It represents the shape of the boundary of the shape.
- FIG. 15 since the shape of the surface deformation (crack) in FIG. 15 (A) and the shape of the boundary of the temperature deformation in FIG. 15 (B) are at least partially similar, the floating with cracks. Can be estimated.
- FIG. 16 (A) shows the shape of the surface deformation (crack) derived from the visible image (FIG. 14 (A)), and FIG. 16 (B) shows the temperature derived from the infrared thermal image (FIG. 14 (B)). It represents the shape of the edge (the part where the temperature changes suddenly) inside the deformation. In FIG. 16, the boundary of the temperature change is also extracted as an edge, but the boundary is omitted. Since the shape of the surface deformation (crack) in FIG. 16A and the shape of the edge inside the temperature deformation in FIG. 16B are at least partially similar, it can be presumed to be a floating with cracks.
- step S4 From the brightness distribution derived in the surface deformation information derivation step (step S4), the range corresponding to the temperature deformation derived in the temperature deformation information derivation step (step S3) is extracted, and the brightness distribution in this range is used. Evaluate the similarity of temperature distribution. At that time, the temperature slope at the boundary of the temperature deformation may be evaluated.
- the cause of the temperature deformation is presumed to be surface deformation. Further, it may be determined whether the surface deformation is peeling or another surface deformation based on the relationship of the temperature difference with the surroundings of the temperature deformation.
- a predetermined range centered on each point on the boundary of the temperature deformation derived in the temperature deformation information derivation step (step S3) is set. Extract and evaluate the similarity between the brightness distribution and the temperature distribution in each range. At that time, the temperature slope of the boundary of the temperature change in each range may be evaluated.
- edge detection is performed on the temperature distribution of the temperature deformation derived in the temperature deformation information derivation step (step S3).
- a predetermined range centered on each point on and inside the temperature deformation boundary derived in the temperature deformation information derivation step (step S3) is extracted.
- the similarity between the brightness distribution in each range and the temperature distribution detected at the edge is evaluated. At that time, only a portion having an edge size of a predetermined value or more may be extracted from the temperature distribution detected at the edge.
- ⁇ Second aspect> As a second aspect, in the surface deformation information derivation step (step S4), the presence / absence, type, shape, position, and brightness distribution of the surface deformation are explicitly derived from the brightness distribution of the portion corresponding to the temperature change. The case where the type of other surface deformation is not discriminated will be described. The method of estimating the cause of cracks and peeling and other surface deformations will be described.
- the range includes both the range of this surface deformation and the range corresponding to the temperature deformation derived in the temperature deformation information derivation step (step S3). Similarity may be evaluated. Also, based on the shape rather than the temperature distribution and brightness distribution, for example, a distribution in which pixels with temperature deformation are binarized to 1 and pixels without temperature deformation are binarized to 0, pixels with other surface deformation are set to 1, and pixels without other surface deformation are set to 0. The similarity of the binarized distribution may be evaluated.
- the temperature deformation caused by other surface deformations is characterized by a steeper temperature inclination at the boundary than the temperature deformations caused by internal defects.
- the cause of the temperature change may be estimated based on the above. Since the method has been described in the first aspect, the description thereof will be omitted.
- the cause of the temperature change is presumed to be a cause without peeling, for example, an internal defect such as floating without peeling.
- the position and shape of the temperature deformation derived in the temperature deformation information derivation step (step S3) are compared with the position and shape of the peeling, and the temperature deformation and the peeling position are different and adjacent to each other.
- the cause of the temperature change is presumed to be floating with peeling.
- a distribution in which pixels with temperature deformation are binarized to 1 and pixels without temperature deformation is binarized to 0, and a distribution in which pixels with peeling are binarized to 1 and pixels without peeling are binarized to 0, based on the shape of the temperature deformation.
- a predetermined range centered on each point on the boundary of the deformation is extracted, and the temperature deformation shape distribution in each range (pixels with temperature deformation are binarized to 1 and pixels without temperature deformation are binarized to 0).
- the similarity of the peeled shape distribution (the distribution in which the pixels with peeling are binarized to 1 and the pixels without peeling are binarized to 0) is evaluated, and if they are similar in any part, it is estimated that the floating is accompanied by peeling.
- the temperature distribution and brightness distribution in a predetermined range centered on each point on the boundary of the temperature deformation are extracted, the similarity between the temperature distribution and brightness distribution in each range is evaluated, and they are similar in any part. It may be presumed that the floating is accompanied by peeling.
- the peeling is outside the magnitude of the temperature deformation derived in the temperature deformation information derivation step (step S3) and the boundary of the peeling is located near the boundary of the magnitude of the temperature deformation, the peeling is accompanied. It may be presumed to be floating.
- the temperature change caused by the floating accompanied by the peeling has a steep temperature slope in the portion adjacent to the peeling in the boundary of the temperature change. Therefore, the similarity between the shape or temperature distribution of the temperature deformation and the shape or brightness distribution of the peeling is evaluated in a predetermined range centered on each point on the boundary of the temperature deformation, and the temperature deformation in the predetermined range is evaluated. Evaluate the temperature gradient of the boundary, and if there is a portion where the temperature deformation shape or temperature distribution is similar to the peeling shape or brightness distribution and / or there is a steep temperature gradient, peeling is accompanied. It may be presumed to be floating.
- the boundary of the peeling is located near the boundary of the magnitude of the temperature deformation, and the temperature gradient of the temperature deformation is steep near the boundary, the peeling is performed. It may be presumed to be an accompanying float.
- temperature slope may be standardized as described in the first aspect.
- the cause of the temperature change is presumed to be a cause without cracks, for example, an internal defect such as a float without cracks.
- the similarity between the shape of the temperature deformation derived in the temperature deformation information derivation step (step S3), that is, the shape of the boundary and the shape of the crack is evaluated. Then, if any part of the temperature deformation boundary is similar, it is presumed that there is a crack along the boundary at that part, that is, a floating with a crack.
- the evaluation of similarity is performed as follows, for example.
- step S3 evaluate the similarity between the distribution obtained by edge detection on the temperature distribution of the temperature deformation and the brightness distribution of the cracks, and if they are similar at any part on the boundary of the temperature deformation, the boundary at that part It may be presumed that there is a crack along the line, that is, a floating with a crack. Further, if there is a crack inside the temperature deformation derived in the temperature deformation information derivation step (step S3), it is presumed that the float is accompanied by a crack. Alternatively, if the crack is near or inside the boundary of the magnitude of the temperature deformation derived in the temperature deformation information derivation step (step S3), it may be presumed to be a floating with a crack.
- the temperature change caused by the floating with a crack has a steep temperature slope in the cracked part along the boundary of the temperature change and in the cracked part inside the temperature change. Is. Therefore, the similarity between the shape of the temperature deformation or the distribution obtained by edge detection of the temperature distribution in a predetermined range centered on each point on the boundary of the temperature deformation and the shape of the crack or the brightness distribution is evaluated. , The temperature inclination of the boundary of the temperature deformation in a predetermined range, for example, the size of the detected edge is evaluated, and the edge is detected in the shape or temperature distribution of the temperature deformation in any part, and the shape or brightness of the crack.
- the distribution is similar and / or there is a portion with a steep temperature gradient, it may be presumed to be a floating with cracks. Further, if there is a crack inside the temperature deformation, the temperature gradient of the temperature distribution in the cracked portion may be evaluated, and if the temperature gradient is steep in the cracked portion, it may be presumed to be a floating with a crack. At that time, in a predetermined range including the cracked part, the similarity between the distribution obtained by edge detection in the temperature distribution and the shape or brightness distribution of the crack is also evaluated, and if they are similar and / or the temperature slope is steep. In some cases, it may be presumed to be a floating with cracks.
- temperature slope may be standardized as described in the first aspect.
- step S4 the presence or absence of other surface deformations derived in the surface deformation information derivation step (step S4), and if any, the surface deformations and the temperature deformations derived in the temperature deformation information derivation step (step S3). Evaluate similarity. At this time, the temperature slope at the boundary of the temperature deformation may be evaluated.
- step S4 the presence or absence of peeling derived in the surface deformation information derivation step (step S4), and if there is, the partial similarity between the peeling and the temperature deformation derived in the temperature deformation information derivation step (step S3). Evaluate the positional relationship. At this time, the temperature slope of the boundary of the temperature change in each portion may be evaluated.
- step S4 the presence or absence of cracks and the positional relationship (evaluating the presence or absence of cracks near the boundary of the temperature deformation or inside) derived in the surface deformation information derivation step (step S4) are evaluated, and if there are cracks, the cracks and temperature are evaluated.
- the boundary and internal similarity of the temperature deformation derived in the deformation information derivation step (step S3) are evaluated. At this time, the temperature slope of the temperature change in each portion may be evaluated.
- the cause of the temperature change is the floating with the crack. Presumed to be.
- ⁇ Third aspect> As a third aspect, in the surface deformation information derivation step (step S4), the presence / absence, type, shape, position, and brightness distribution of the surface deformation are explicitly derived from the brightness distribution of the portion corresponding to the temperature change. Further, a case where the type of other surface deformation is discriminated will be described.
- the cause estimation method for cracks and peeling is the same as the second aspect, and only the cause estimation method for other surface deformations is different from the second aspect.
- the information effective for estimating the cause differs depending on the type of surface deformation as follows.
- the brightness distribution is effective for estimating the cause as surface deformation information.
- the brightness distribution is effective for estimating the cause, but for the repair mark in FIG. 7, the brightness distribution on the surface of the repair mark becomes noise, and it can be understood that only the shape is more effective. Therefore, in the third aspect, the cause is estimated for other surface deformations basically by the same method as in the second aspect, but the surface deformation information used for the evaluation of similarity is used properly depending on the type of surface deformation. Specifically, the brightness distribution or the shape should be used properly.
- the temperature deformation caused by the surface deformation has a steeper slope of the temperature at the boundary than the temperature deformation caused by the internal defect.
- there are some surface deformations such as the rust juice shown in FIG. 5 in which the temperature inclination of the boundary is not necessarily steep. Therefore, depending on the type of surface deformation, the temperature slope of the boundary of the temperature deformation is evaluated or not, and the temperature slope is evaluated for the surface deformation such as rust juice in which the temperature slope of the boundary is not necessarily steep. A method in which the above is not carried out is preferable.
- Estimatiation result display step> The information display unit 61 displays the estimation result of the cause of the temperature change on the display device 30 via the display control unit 26 (estimation result display step: step S6).
- the cause estimated in the cause estimation step (step S5) is displayed. It may be displayed near each temperature change on the infrared thermal image, or may be displayed together with the position and shape of the temperature change on the visible image. The infrared thermal image and the visible image may be displayed on any processed image and / or any other image. In the case of the third aspect (when the type of other surface deformation is determined), if the cause estimated in the cause estimation step (step S5) is "other surface deformation", the type (color unevenness). , Joints, steps, rust juice, etc.).
- FIG. 17 is a diagram showing an example of a display result in which the estimation result is displayed on the display device 30.
- FIG. 17 displays an infrared thermal image corresponding to FIG. 10 (B). Further, the part determined to be the temperature change in the temperature change information derivation step (step S3) and extracted is surrounded by a white frame and displayed, and the cause estimated in the cause estimation step (step S5) "floating with peeling" is displayed. Is displayed in white near the white frame.
- FIG. 17 is just an example, and in order to make it easier to see the cause "floating with peeling", only the background of the character may be a white background, and there are innumerable variations in how to show it, and there is no particular limitation.
- the cause of the temperature deformation is estimated to be another surface deformation so that the temperature difference due to the cause inside the structure can be discriminated, the temperature distribution due to the surface deformation is changed to the temperature. It may be estimated based on the information and surface deformation information and reduced from the original temperature distribution. There are various methods.
- the brightness distribution of the surface deformation which is the surface deformation information
- a predetermined blurring process the temperature distribution due to the surface deformation is blurred due to the cause such as heat conduction compared to the brightness distribution
- the contrast may be adjusted to best match the original temperature distribution, and the contrast-optimized luminance distribution may be presumed to be the temperature distribution due to surface deformation and subtracted or removed from the original temperature distribution. ..
- the method of estimating the temperature distribution from the luminance distribution of the surface deformation like this method is not suitable for the surface deformation whose luminance distribution and temperature distribution are not similar. That is, as described above, this method is suitable for surface deformation such as color unevenness, rust juice or rust, in which the difference in the amount of visible light absorbed from concrete is the main cause of the difference in temperature.
- temperature differences due to surface deformations such as repair marks and free lime, which are the main causes of temperature differences due to differences in thermal conductivity and infrared emissivity from concrete, and structural causes such as joints and steps. Not suitable for surface deformations that occur.
- various thermal parameters related to surface deformation are set according to the type of surface deformation, and thermal simulation (simulation including heat conduction, heat radiation, and convection) is performed based on these parameters and the shape of the surface deformation. May be performed to simulate the temperature distribution, and the temperature distribution in the simulation that best matches the original temperature distribution may be estimated as the temperature distribution due to surface deformation and subtracted from or removed from the original temperature distribution.
- thermal conductivity and infrared emissivity are set as particularly important parameters, and structural joints, steps, etc. are set.
- structural parameters such as the depth and height of the unevenness of the joint and the height of the step are set as particularly important parameters.
- the original temperature distribution is represented by a temperature distribution that is approximated by a lower-order mathematical formula, for example, a low-order polynomial such as a linear equation (plane) or a quadratic equation (curved surface), and is the most similar to the original temperature distribution.
- the matching temperature distribution may be estimated as the temperature distribution due to surface deformation and subtracted from the original temperature distribution.
- a distribution obtained by subjecting the original temperature distribution to a predetermined blurring process may be estimated as a temperature distribution due to surface deformation and subtracted from the original temperature distribution.
- step S3 the temperature deformation determination and temperature deformation information derivation in the temperature deformation information derivation step (step S3), the surface deformation information derivation in the surface deformation information derivation step (step S4), and the cause estimation step (step).
- step S5 The method of performing each process of estimating the cause of S5) on a rule basis has been described.
- each step of the temperature deformation information derivation step (step S3), the surface deformation information derivation step (step S4), and the cause estimation step (step S5) may be carried out by various machine learning methods. can.
- FCN Frully Convolutional Network
- SegNet A Deep Convolutional Encoder-Decoder
- Architecture for Image Segmentation U-Net (Convolutional Networks for Biomedical Image Segmentation), and other machine learning methods that detect objects from images and extract their regions can be used.
- the same machine learning method can be used for the method of deriving the surface deformation information from the visible image in the surface deformation information derivation step (step S4).
- a visible image when there are a plurality of types of images such as an RGB image, all types of visible images may be input.
- R-CNN (Regions with CNN (Convolutional Neural Network) features), Fast R-CNN, Yolo (Regions with CNN (Convolutional Neural Network) features), Fast R-CNN, Yolo (Regions with CNN (Convolutional Neural Network) features), Fast R-CNN, Yolo (Regions with CNN (Convolutional Neural Network) features) Machine learning methods such as You Only Look Once) and SSD (Single Shot MultiBox Detector) can also be used.
- step S4 first, the surface deformation is determined, and the feature amount (feature amount such as average brightness, shape, texture of brightness distribution, contrast, frequency spectrum) is extracted from the brightness distribution.
- feature amount feature amount such as average brightness, shape, texture of brightness distribution, contrast, frequency spectrum
- discrimination method discriminating the type of surface deformation (other surface deformation, peeling, cracking, and other types of surface deformation such as uneven color, joints, steps, rust juice, etc.) from the feature amount, as the discrimination method.
- Machine learning methods that classify objects based on features, such as logistic regression, linear discriminant analysis, K-nearest method, decision tree (classification tree), random forest, and support vector machine (SVM), can be used.
- the method of estimating the cause of the temperature deformation from the temperature deformation information and the surface deformation information is also carried out by various machine learning methods according to the form of the temperature deformation information and the surface deformation information.
- various machine learning methods for example, when inputting the temperature distribution of the temperature change as the temperature change information and the brightness distribution of the portion corresponding to the temperature change as the surface change information, a machine learning method such as DNN (Deep Neural Network) or CNN is used.
- DNN Deep Neural Network
- the causes of temperature deformation are "other surface deformation (when classifying other types of surface deformation," color unevenness "," joint “,” step “,” rust juice ", etc.)” "floating with peeling” " It can be classified into “floating with cracks” and "floating”.
- the temperature deformation information derivation step (step S3), the surface deformation information derivation step (step S4), and the cause estimation step (step S5) can all be carried out by one machine learning method. That is, the infrared thermal image and the visible image (if there are multiple types of images such as RGB, all types of visible images) are input, and the FCN, SegNet, U-Net, R-CNN, and Fast mentioned above are input. It is possible to detect temperature changes by machine learning methods such as R-CNN, Yolo, and SSD, and to classify the causes. When executing machine learning, it is preferable that the learning data is abundant.
- the causes of other surface deformations, peeling, and cracks were estimated in order, and one cause was estimated as the cause of the temperature deformation.
- Multiple candidates may be estimated as the cause of the temperature change. That is, the cause may be estimated for each of the other surface deformations, peeling, and cracks, and a plurality of candidates such as other surface deformations, floating with peeling, and floating with cracks may be performed as the causes of the temperature deformation. May be estimated.
- the cause may be estimated probabilistically instead of being estimated by alternatives. For example, for other surface deformations, peeling, and cracks, if they are not located at the locations corresponding to the temperature deformations in the visible image, the probability of being the cause of the temperature deformations is 0%, and if there are, the temperature deformations. As the probability that is the cause of The value obtained by converting (such as the probability correlation coefficient of) into a probability may be adopted.
- a value converted into a probability may be adopted including the calculated value of the temperature gradient inside and the boundary of the temperature deformation.
- the program of the above embodiment may be implemented by a dedicated analysis program, and the device to be implemented does not matter. For example, it can be carried out on a personal computer. Also, the devices or programs that perform each step may be integrated or separated.
- the hardware structure of the processing unit that executes various processes is various processors as shown below.
- the circuit configuration can be changed after manufacturing the CPU (Central Processing Unit), FPGA (Field Programmable Gate Array), etc., which are general-purpose processors that execute software (programs) and function as various processing units.
- a dedicated electric circuit which is a processor having a circuit configuration specially designed to execute a specific process such as a programmable logic device (PLD) and an ASIC (Application Specific Integrated Circuit). Is done.
- One processing unit may be composed of one of these various processors, or may be composed of two or more processors of the same type or different types (for example, a plurality of FPGAs or a combination of a CPU and an FPGA). You may. Further, a plurality of processing units can be configured by one processor. As an example of configuring a plurality of processing units with one processor, first, one processor is configured by a combination of one or more CPUs and software, as represented by a computer such as a client or a server. There is a form in which the processor functions as a plurality of processing units.
- SoC System On Chip
- a processor that realizes the functions of the entire system including a plurality of processing units with one IC (Integrated Circuit) chip is used.
- the various processing units are configured by using one or more of the above-mentioned various processors as a hardware-like structure.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
Description
図1は、実施形態に係る画像解析装置のハードウェア構成の一例を示すブロック図である。
赤外線熱画像取得部51は、点検対象の構造物36を撮影した赤外線熱画像を取得する(赤外線熱画像取得ステップ:ステップS1)。赤外線熱画像は、記憶部16に記憶された赤外線熱画像101である。赤外線熱画像101が、赤外線熱画像取得部51により記憶部16から取得される。なお、記憶部16に赤外線熱画像101が記憶されていない場合には、赤外線熱画像取得部51は外部から赤外線熱画像101を取得する。例えば、赤外線熱画像取得部51は、入出力インターフェイス12を介してネットワークを通して、赤外線熱画像101を取得でき、また、赤外線熱画像取得部51は、入出力インターフェイス12を介して赤外線カメラ32から赤外線熱画像101を取得できる。
可視画像取得部53は、点検対象の構造物を撮影した可視画像を取得する(可視画像取得ステップ:ステップS2)。可視画像は、記憶部16に記憶された可視画像103である。可視画像103が、可視画像取得部53により記憶部16から取得される。なお、記憶部16に可視画像103が記憶されていない場合には、可視画像取得部53は外部から可視画像103を取得する。例えば、可視画像取得部53は、入出力インターフェイス12を介してネットワークを通して、可視画像103を取得でき、また、可視画像取得部53は、入出力インターフェイス12を介して可視カメラ34から可視画像103を取得できる。
次に、温度変状情報導出部55は、赤外線熱画像101に基づいて温度変状を判定し、温度変状情報102を導出する(温度変状情報導出ステップ:ステップS3)。
表面変状情報導出部57は、可視画像103に基づいて表面変状情報104を導出する(表面変状情報導出ステップ:ステップS4)。
原因推定部59は、温度変状情報102と表面変状情報104とに基づき、温度変状の原因を推定する(原因推定ステップ:ステップS5)。
第一態様として、表面変状情報導出ステップ(ステップS4)で温度変状に対応する箇所の輝度分布を表面変状情報とした場合を説明する。ひび及び剥離と、他の表面変状の夫々について原因推定の方法を説明する。
表面変状情報導出ステップ(ステップS4)で導出した輝度分布から、温度変状情報導出ステップ(ステップS3)で導出した温度変状に該当する範囲、または温度変状を含むやや広い範囲を抽出する。この範囲の輝度分布と温度分布の類似性を評価し、類似する場合、温度変状の原因をこの表面変状であると推定する。一方で、類似しない場合、この表面変状以外であると推定し、例えば浮きなどの内部欠陥であると推定する。
ここで、v(x,y)は規格化した輝度分布における座標(x,y)の画素の値を、t(x,y)は規格化した温度分布における座標(x,y)の画素の値を表す。
ここでv(x,y)は輝度分布における座標(x,y)の画素の値を、v_aveは輝度分布の平均値を、t(x,y)は温度分布における座標(x,y)の画素の値を、t_aveは温度分布の平均値を表す。
なお、温度変状の境界における温度の傾きは、日射や外気温などの熱環境によって大きく変化する。そこで温度変状と周囲の温度の差分の絶対値や、温度変状と温度変状を含む広い所定範囲における構造物表面の平均温度の差分の絶対値などを計算し、その差分絶対値の最大値や平均値などによって境界における温度の傾きを規格化し(例えば除する)、規格化した温度の傾きが所定値以上の場合に、温度変状の原因を表面変状であると推定してもよい。
図9と図10とは、昼間にコンクリート構造物を撮影した画像である。図9(A)は可視画像であり、図9(B)は赤外線熱画像である。また、図10(A)は可視画像であり、図10(B)は赤外線熱画像である。図9及び図10では、撮影された点検対象のコンクリート構造物の表面に剥離した箇所を含んでいる。
図13と図14とは、昼間にコンクリート構造物を撮影した画像である。図13(A)は可視画像であり、図13(B)は赤外線熱画像である。また、図14(A)は可視画像であり、図14(B)は赤外線熱画像である。図13及び図14では、撮影された点検対象のコンクリート構造物の表面にひびが有る箇所を含んでいる。
第二態様として、表面変状情報導出ステップ(ステップS4)で、温度変状に対応する箇所の輝度分布から、表面変状の有無、種類、形状、位置、輝度分布を明示的に導出し、他の表面変状について種類判別しなかった場合を説明する。ひび及び剥離と、他の表面変状の夫々について原因推定の方法を説明する。
他の表面変状が無い場合、温度変状の原因を、この表面変状以外であると推定し、例えば浮きなどの内部欠陥であると推定する。他の表面変状が有る場合、第一態様で説明した同様の方法で表面変状の輝度分布と温度分布の類似性を評価し、類似する場合、温度変状の原因をこの表面変状であると推定し、類似しない場合、この表面変状以外、例えば浮きなどの内部欠陥であると推定する。但し、他の表面変状の形状を既に導出してあるので、この表面変状の範囲と温度変状情報導出ステップ(ステップS3)で導出した温度変状に該当する範囲の両方を含む範囲で類似性を評価してもよい。また温度分布と輝度分布ではなく形状に基づき、例えば温度変状の有る画素を1、無い画素を0に2値化した分布と、他の表面変状の有る画素を1、無い画素を0に2値化した分布の類似性を評価してもよい。
剥離が無い場合、温度変状の原因を、剥離を伴わない原因、例えば剥離の伴わない浮きなどの内部欠陥であると推定する。剥離が有る場合、温度変状情報導出ステップ(ステップS3)で導出した温度変状の位置及び形状と、剥離の位置及び形状とを比較し、温度変状と剥離の位置が異なり、且つ、隣接して境界の形状が部分的に類似する場合に、温度変状の原因を、剥離を伴う浮きと推定する。
ひびが無い場合、温度変状の原因を、ひびを伴わない原因、例えばひびの伴わない浮きなどの内部欠陥であると推定する。ひびが有る場合、温度変状情報導出ステップ(ステップS3)で導出した温度変状の形状すなわち、境界の形状と、ひびの形状との類似性を評価する。そして、温度変状の境界上の何れかの部分において類似する場合、その部分で境界に沿ってひびが有る、つまりひびを伴う浮きと推定する。類似性の評価は、例えばいかの通り行う。温度変状の境界の画素を1、それ以外の画素を0とした分布と、ひびの有る画素を1、無い画素を0とした分布について、温度変状の境界上の夫々の点を中心とした所定範囲を抽出する。そして、夫々の範囲の温度変状境界形状分布、すなわち温度変状の境界の画素を1、それ以外の画素を0とした分布と、ひび形状分布、すなわちひびの有る画素を1、無い画素を0とした分布との類似性を評価することにより行う。又は、温度変状の温度分布にエッジ検出を行った分布と、ひびの輝度分布との類似性を評価し、温度変状の境界上の何れかの部分において類似する場合、その部分において境界に沿ってひびが有る、つまりひびを伴う浮きと推定してもよい。また、温度変状情報導出ステップ(ステップS3)で導出した温度変状の内部にひびが有る場合、ひびを伴う浮きと推定する。又は、ひびが、温度変状情報導出ステップ(ステップS3)で導出した温度変状の大きさの境界付近又は内部に有る場合にひびを伴う浮きと推定してもよい。
第三態様として、表面変状情報導出ステップ(ステップS4)で、温度変状に対応する箇所の輝度分布から、表面変状の有無、種類、形状、位置、輝度分布を明示的に導出し、更に他の表面変状について種類判別した場合を説明する。
情報表示部61は、温度変状の原因の推定結果を、表示制御部26を介して表示装置30に表示する(推定結果表示ステップ:ステップS6)。
浮きなどのコンクリート構造物内部の原因による構造物表面温度の差異が、表面変状に起因する構造物表面温度の差異に埋もれ、構造物内部の原因による温度の差異を判別できないことが生じうる。例えば、コンクリート表面において剥離した箇所は補修材が充てんされ補修されるが、その補修材が数年後に元のコンクリート構造物から浮いてくる場合がある。その浮きによる温度の差異は判別できることが望ましいが、補修材自体の熱伝導率や赤外線放射率が周囲のコンクリートと異なることで発生する温度の差異に埋もれて判別できないことが生じる。
上記実施形態において、各種の処理を実行する処理部(processing unit)のハードウェア的な構造は、次に示すような各種のプロセッサ(processor)である。各種のプロセッサには、ソフトウェア(プログラム)を実行して各種の処理部として機能する汎用的なプロセッサであるCPU(Central Processing Unit)、FPGA(Field Programmable Gate Array)などの製造後に回路構成を変更可能なプロセッサであるプログラマブルロジックデバイス(Programmable Logic Device:PLD)、ASIC(Application Specific Integrated Circuit)などの特定の処理を実行させるために専用に設計された回路構成を有するプロセッサである専用電気回路などが含まれる。
12 入出力インターフェイス
16 記憶部
18 操作部
20 CPU
22 RAM
24 ROM
26 表示制御部
30 表示装置
32 赤外線カメラ
34 可視カメラ
36 構造物
51 赤外線熱画像取得部
53 可視画像取得部
55 温度変状情報導出部
57 表面変状情報導出部
59 原因推定部
61 情報表示部
101 赤外線熱画像
102 温度変状情報
103 可視画像
104 表面変状情報
S1 ステップ
S2 ステップ
S3 ステップ
S4 ステップ
S5 ステップ
S6 ステップ
Claims (16)
- プロセッサを備える画像解析装置であって、
前記プロセッサは、
点検対象の構造物を撮影した赤外線熱画像を取得し、
点検対象の前記構造物を撮影した可視画像を取得し、
前記赤外線熱画像から温度変状を判定し、
前記温度変状について、
少なくとも、前記赤外線熱画像から得られる温度変状情報と、
前記可視画像から得られる表面変状情報と、に基づき、
前記温度変状の原因を推定する、画像解析装置。 - 前記温度変状情報は、前記温度変状について、前記赤外線熱画像の温度分布及び/又は前記温度分布から得られる情報を含む、請求項1に記載の画像解析装置。
- 前記温度変状情報は、前記温度変状の形状及び/又は大きさの情報を含む、請求項2に記載の画像解析装置。
- 前記表面変状情報は、前記可視画像の輝度分布及び/又は輝度分布から得られる情報を含む、請求項1から3のいずれか一項に記載の画像解析装置。
- 前記表面変状情報は、前記表面変状の種類、形状及び位置の少なくとも一つの情報を含む、請求項4に記載の画像解析装置。
- 前記プロセッサは、
前記温度変状情報と前記表面変状情報との類似性に基づき前記温度変状の原因を推定する、請求項1から5のいずれか一項に記載の画像解析装置。 - 前記プロセッサは、
前記温度変状の原因を前記表面変状であると推定した場合、
前記表面変状による温度分布を推定し、前記赤外線熱画像から低減する、請求項6に記載の画像解析装置。 - 前記類似性は部分的な類似性を含む、請求項6に記載の画像解析装置。
- 前記プロセッサは、
前記温度変状に対応する前記表面変状をひび又は剥離と判定し、
前記温度変状情報と前記表面変状情報との類似性を評価し、
少なくとも一部分が類似する場合に、前記温度変状の原因が、前記ひび又は剥離を伴う浮きであると推定する、
請求項1から5のいずれか一項に記載の画像解析装置。 - 前記プロセッサは、
前記温度変状に対応する前記表面変状をひび又は剥離と判定し、
前記温度変状の大きさの中に、及び/又は前記温度変状の境界付近に前記表面変状の有無を評価し、
有りの場合に前記温度変状の原因が、前記ひび又は剥離を伴う浮きであると推定する、
請求項1から5のいずれか一項に記載の画像解析装置。 - 前記プロセッサは、
前記温度変状の境界の温度の傾きに基づき前記温度変状の原因を推定する、
請求項1から10のいずれか一項に記載の画像解析装置。 - 前記表面変状は補修跡、遊離石灰、目地、段差、ひび及び剥離の少なくとも一つを含む、請求項11に記載の画像解析装置。
- 前記可視画像は可視光の波長域において2種類以上の異なる波長域での反射強度分布を画像化した画像である、請求項1から12のいずれか一項に記載の画像解析装置。
- さらに、表示装置を備え、
前記プロセッサは、前記表示装置に前記温度変状の原因の推定結果を表示する、請求項1から13のいずれか一項に記載の画像解析装置。 - 点検対象の構造物を撮影した赤外線熱画像を取得するステップと、
点検対象の前記構造物を撮影した可視画像を取得するステップと、
前記赤外線熱画像から温度変状を判定するステップと、
前記温度変状について、
少なくとも、前記赤外線熱画像から得られる温度変状情報と、
前記可視画像から得られる表面変状情報と、に基づき、
前記温度変状の原因を推定するステップと、
を含む画像解析方法。 - 点検対象の構造物を撮影した赤外線熱画像を取得するステップと、
点検対象の前記構造物を撮影した可視画像を取得するステップと、
前記赤外線熱画像から温度変状を判定するステップと、
前記温度変状について、
少なくとも、前記赤外線熱画像から得られる温度変状情報と、
前記可視画像から得られる表面変状情報と、に基づき、
前記温度変状の原因を推定するステップと、
をコンピュータに実行させるためのプログラム。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202180090589.7A CN116710763A (zh) | 2021-01-14 | 2021-12-27 | 图像分析装置、图像分析方法及程序 |
JP2022575515A JPWO2022153852A1 (ja) | 2021-01-14 | 2021-12-27 | |
EP21919738.1A EP4279908A4 (en) | 2021-01-14 | 2021-12-27 | IMAGE ANALYSIS DEVICE, IMAGE ANALYSIS METHOD AND PROGRAM |
US18/344,823 US20230342907A1 (en) | 2021-01-14 | 2023-06-29 | Image analysis apparatus, image analysis method, and program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021-004419 | 2021-01-14 | ||
JP2021004419 | 2021-01-14 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US18/344,823 Continuation US20230342907A1 (en) | 2021-01-14 | 2023-06-29 | Image analysis apparatus, image analysis method, and program |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022153852A1 true WO2022153852A1 (ja) | 2022-07-21 |
Family
ID=82447283
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/048467 WO2022153852A1 (ja) | 2021-01-14 | 2021-12-27 | 画像解析装置、画像解析方法及びプログラム |
Country Status (5)
Country | Link |
---|---|
US (1) | US20230342907A1 (ja) |
EP (1) | EP4279908A4 (ja) |
JP (1) | JPWO2022153852A1 (ja) |
CN (1) | CN116710763A (ja) |
WO (1) | WO2022153852A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20220099588A1 (en) * | 2020-09-29 | 2022-03-31 | Canon Kabushiki Kaisha | Information processing apparatus, information processing method, and non-transitory computer-readable storage medium |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113167742B (zh) * | 2018-11-29 | 2024-02-27 | 富士胶片株式会社 | 混凝土构造物的点检辅助装置、点检辅助方法及记录介质 |
US20240255932A1 (en) * | 2022-08-03 | 2024-08-01 | Industrial Video Solutions Inc. | Systems and methods for monitoring and controlling industrial processes |
US20240212356A1 (en) * | 2022-08-03 | 2024-06-27 | Industrial Video Solutions Inc. | Systems and methods for monitoring and controlling industrial processes |
CN117462086B (zh) * | 2023-10-30 | 2024-07-26 | 北京鹰之眼智能健康科技有限公司 | 红外图像处理系统、电子设备及存储介质 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05307013A (ja) * | 1992-04-28 | 1993-11-19 | Jeol Ltd | サーモグラフィ装置を用いた検査方法 |
JP2011099687A (ja) * | 2009-11-04 | 2011-05-19 | Daiwa Odakyu Construction Co Ltd | 赤外線法によるコンクリート表層部の変状部検出方法 |
JP2013096741A (ja) | 2011-10-28 | 2013-05-20 | West Nippon Expressway Engineering Shikoku Co Ltd | 構造物の赤外線調査方法及び赤外線調査用演算装置 |
JP2013224849A (ja) * | 2012-04-20 | 2013-10-31 | West Nippon Expressway Engineering Shikoku Co Ltd | 構造物の不具合確率演算方法および不具合確率演算装置、構造物の不具合範囲判定方法および不具合範囲判定装置 |
JP2017203761A (ja) * | 2016-05-11 | 2017-11-16 | 直人 今西 | 物体内変状部分検出装置 |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10497109B2 (en) * | 2015-10-07 | 2019-12-03 | Fuchs Consulting, Inc. | Time-lapse infrared thermography system and method for damage detection in large-scale objects |
US10191014B2 (en) * | 2016-08-23 | 2019-01-29 | The Boeing Company | System and method for nondestructive evaluation of a test object |
-
2021
- 2021-12-27 EP EP21919738.1A patent/EP4279908A4/en active Pending
- 2021-12-27 JP JP2022575515A patent/JPWO2022153852A1/ja active Pending
- 2021-12-27 CN CN202180090589.7A patent/CN116710763A/zh active Pending
- 2021-12-27 WO PCT/JP2021/048467 patent/WO2022153852A1/ja active Application Filing
-
2023
- 2023-06-29 US US18/344,823 patent/US20230342907A1/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05307013A (ja) * | 1992-04-28 | 1993-11-19 | Jeol Ltd | サーモグラフィ装置を用いた検査方法 |
JP2011099687A (ja) * | 2009-11-04 | 2011-05-19 | Daiwa Odakyu Construction Co Ltd | 赤外線法によるコンクリート表層部の変状部検出方法 |
JP2013096741A (ja) | 2011-10-28 | 2013-05-20 | West Nippon Expressway Engineering Shikoku Co Ltd | 構造物の赤外線調査方法及び赤外線調査用演算装置 |
JP2013224849A (ja) * | 2012-04-20 | 2013-10-31 | West Nippon Expressway Engineering Shikoku Co Ltd | 構造物の不具合確率演算方法および不具合確率演算装置、構造物の不具合範囲判定方法および不具合範囲判定装置 |
JP2017203761A (ja) * | 2016-05-11 | 2017-11-16 | 直人 今西 | 物体内変状部分検出装置 |
Non-Patent Citations (1)
Title |
---|
See also references of EP4279908A4 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20220099588A1 (en) * | 2020-09-29 | 2022-03-31 | Canon Kabushiki Kaisha | Information processing apparatus, information processing method, and non-transitory computer-readable storage medium |
US11898966B2 (en) * | 2020-09-29 | 2024-02-13 | Canon Kabushiki Kaisha | Information processing apparatus, information processing method, and non-transitory computer-readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
US20230342907A1 (en) | 2023-10-26 |
EP4279908A1 (en) | 2023-11-22 |
EP4279908A4 (en) | 2024-07-10 |
CN116710763A (zh) | 2023-09-05 |
JPWO2022153852A1 (ja) | 2022-07-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2022153852A1 (ja) | 画像解析装置、画像解析方法及びプログラム | |
US11221107B2 (en) | Method for leakage detection of underground pipeline corridor based on dynamic infrared thermal image processing | |
Li et al. | A visual detection system for rail surface defects | |
KR102058427B1 (ko) | 검사 장치 및 방법 | |
He et al. | Deep regression neural network for industrial surface defect detection | |
CN106228546B (zh) | 一种板卡的检测方法及装置 | |
Ichi et al. | Effectiveness of infrared thermography for delamination detection in reinforced concrete bridge decks | |
CN111862064A (zh) | 一种基于深度学习的银质丝材表面瑕疵识别方法 | |
Miao et al. | An image processing-based crack detection technique for pressed panel products | |
US20130034305A1 (en) | Image-based crack quantification | |
US11747284B2 (en) | Apparatus for optimizing inspection of exterior of target object and method thereof | |
Yang et al. | An accurate mura defect vision inspection method using outlier-prejudging-based image background construction and region-gradient-based level set | |
US20170278235A1 (en) | Fast density estimation method for defect inspection application | |
TW201512649A (zh) | 偵測晶片影像瑕疵方法及其系統與電腦程式產品 | |
KR101677070B1 (ko) | 형태학적 영상 처리와 레이블링을 이용한 얼룩 결함 자동 검출 시스템 및 방법 | |
Xia et al. | An automatic machine vision-based algorithm for inspection of hardwood flooring defects during manufacturing | |
CN109900363A (zh) | 一种基于轮廓提取的目标物红外测温方法和装置 | |
KR102242996B1 (ko) | 자동차 사출제품의 비정형 불량 검출 방법 | |
CN116188379A (zh) | 边缘缺陷检测方法、装置、电子设备及存储介质 | |
CN110426395B (zh) | 一种太阳能el电池硅片表面检测方法及装置 | |
Chiu et al. | An innovative blemish detection system for curved LED lenses | |
JP2017062181A (ja) | 表面疵検査装置及び表面疵検査方法 | |
Chen et al. | An effective image segmentation method for noisy low-contrast unbalanced background in Mura defects using balanced discrete-cosine-transfer (BDCT) | |
TW201510517A (zh) | 用於改善晶圓表面檢查靈敏度之方法及系統 | |
CN109785290A (zh) | 基于局部光照归一化的钢板缺陷检测方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21919738 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2022575515 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202180090589.7 Country of ref document: CN |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2021919738 Country of ref document: EP Effective date: 20230814 |