JP2020159793A - Painted surface inspection device, learning device, and program - Google Patents

Abstract

To provide a painted surface inspection device which can automatically inspect a painted surface of an object.SOLUTION: A painted surface inspection device 1 includes: a reception unit 11 for receiving a painted surface image as an image of a painted surface of an object with a painted metal plate; and a region specification unit 13 for specifying at least a degraded region other than a rusty region of the painted surface in the image. In that way, it becomes possible to specify a degraded region other than a rusty region of a paint surface, such as a region with a crack of painting, a region with a peeling of painting, or a region with a swelling and to conduct a higher-level inspection on a painted surface.SELECTED DRAWING: Figure 1

Description

The present invention relates to a painted surface inspection device and the like for inspecting a painted surface.

Conventionally, a diagnosis regarding rust color has been performed on a pole transformer or the like using captured images (see, for example, Patent Documents 1 and 2).

Japanese Unexamined Patent Publication No. 2012-189523 Japanese Unexamined Patent Publication No. 2013-053958

Deterioration of the painted surface of an object such as a pole transformer includes peeling, cracking, and swelling in addition to rust. Usually, these deteriorations have a large effect as well as rust, so the presence or absence of such deterioration other than rust is important in determining whether to continue using the object.

Conventionally, such peeling, cracking, and swelling have not been inspected using captured images, and have been determined by visually comparing them with limit samples and the like. Therefore, for example, when the object is a pole transformer, it is necessary to prepare a bucket car or perform traffic regulation to perform a visual inspection, which is a problem that the burden for the inspection is heavy. was there. In addition, when the inspection is performed visually, there is a problem that the accuracy of the inspection result differs depending on the sense and experience of the observer.

The present invention has been made to solve the above problems, and provides a painted surface inspection device capable of automatically inspecting a painted surface of an object, and a learning device used for the inspection. The purpose.

In order to achieve the above object, the painted surface inspection apparatus according to the present invention has a reception unit that receives a painted surface image which is a photographed image of a painted surface of an object having a painted metal plate, and a painted surface image other than rust color. It is provided with an area specifying portion for at least specifying the deteriorated area of the above.
With such a configuration, it is possible to identify the deteriorated region of the coating other than the rust color, which could not be identified only by the detection of the rust color of the conventional example, by using the photographed image of the painted surface of the object. become. Therefore, inspections such as confirmation of the presence or absence of such a deteriorated region and identification of the position of the deteriorated region can be automatically performed using a photographed image of the object to be inspected, and compared with visual inspection. Therefore, the burden of inspection will be reduced. In addition, it becomes possible to realize an examination that does not depend on the sense or experience of the observer.

Further, in the painted surface inspection apparatus according to the present invention, the deterioration region of the coating other than the rust color includes the cracking and peeling region of the coating, and the region specifying portion is the rust color region and the coating in the painted surface image. Areas may be specified, and areas other than the specified areas may be specified as paint cracking and peeling areas.
With such a configuration, it is possible to automatically identify the cracked or peeled area of the paint, which could not be specified only by the conventional rust color detection, by using the photographed image of the painted surface of the object. Will be.

Further, in the painted surface inspection apparatus according to the present invention, the deteriorated region of the coating other than the rust color includes the swelling region of the coating, and the region specifying portion identifies and identifies the coating region in the painted surface image. In the painted area, a swelling area of the paint, which is a region where the brightness dispersion is larger than the threshold value, may be specified.
With such a configuration, it is possible to automatically identify the swelling region of the coating, which could not be identified only by the conventional rust color detection, by using the photographed image of the painted surface of the object. Become.

Further, in the painted surface inspection device according to the present invention, the area specifying portion determines the training input information which is a photographed image of the painted surface of the object having the painted metal plate and the deteriorated area of the coating other than the rust color in the photographed image. Even if the deteriorated area of the paint other than the rust color contained in the painted surface image is specified by applying the painted surface image to the learner learned by using a plurality of sets with the training output information including the teacher figure shown. Good.
With such a configuration, it becomes possible to inspect the deteriorated region of the coating other than the rust color by using the learning device. Therefore, it is not necessary to set rules and conditions for identifying the deteriorated area of the coating other than the rust color, and to adjust the set value, and the burden is reduced.

Further, the painted surface inspection device according to the present invention may further include a region output unit that outputs an output for each region specified by the region identification unit.
With such a configuration, the user can know information about a deteriorated area of the coating such as peeling, cracking, and swelling of the coating.

Further, in the painted surface inspection device according to the present invention, the area specifying portion also specifies a rust-colored region in the painted surface image, and the object is an object according to the information regarding each region specified by the region specifying portion. A determination unit for determining whether or not the painted surface is abnormal, and a determination result output unit for outputting the determination result by the determination unit may be further provided.
With such a configuration, it is possible to automatically inspect whether the painted surface of the object is abnormal, and it is possible to realize an inspection that does not depend on the sense or experience of the observer. Further, when it is determined that there is an abnormality, for example, by repairing the object, it is possible to prevent the object from malfunctioning in advance.

Further, the painted surface inspection device according to the present invention has a reception unit that receives information on a rust-colored region and a paint deterioration region other than the rust-colored region on the painted surface of an object having a painted metal plate, and each received by the reception unit. It is provided with a determination unit that determines whether or not the painted surface of the object is abnormal according to information about the region, and a determination result output unit that outputs the determination result by the determination unit.
With such a configuration, by using the information on the rust-colored region and the deteriorated region of the coating other than the rust-colored on the painted surface of the object, it is possible to automatically inspect whether the painted surface is abnormal. It will be possible to realize tests that do not depend on the observer's senses and experience.

Further, in the painted surface inspection apparatus according to the present invention, the determination unit includes training input information which is information on a rust-colored region and a non-rust-colored coating deterioration region on the painted surface of an object having a painted metal plate, and a target. By applying the information about each area on the painted surface of the object to be judged to the learner learned by using a plurality of sets with the training output information which is the judgment result of whether the painted surface of the object is abnormal. It may be determined whether or not the painted surface of the object to be determined is abnormal.
With such a configuration, it becomes possible to inspect whether or not the painted surface is abnormal by using a learning device. Therefore, the work of setting rules and conditions for determining an abnormality on the painted surface and adjusting the set values becomes unnecessary, and the burden is reduced.

Further, the painted surface inspection device according to the present invention is a reception unit that receives a painted surface image which is a photographed image of the painted surface of an object having a painted metal plate, and a photographed image of the painted surface of the object having a painted metal plate. By applying the painted surface image to the learner learned by using a plurality of sets of the training input information and the training output information which is the judgment result of whether or not the painted surface of the object is abnormal, the object to be judged It is provided with a determination unit for determining whether or not the painted surface of an object is abnormal, and a determination result output unit for outputting a determination result by the determination unit.
With such a configuration, by using a photographed image of the painted surface of the object, it is possible to automatically inspect whether the painted surface is abnormal, and realize an inspection that does not depend on the sense or experience of the observer. You will be able to do it. Further, since the inspection is performed using a learning device, it is not necessary to set rules and conditions for determining an abnormality on the painted surface and to adjust the set values, which reduces the burden.

Further, in the painted surface inspection device according to the present invention, the reception is the reception of the reception target, the output is the transmission of the output target, and a charging unit that charges according to the output may be further provided.
With such a configuration, for example, it is possible to automatically inspect the painted surface of the object in response to a request from the user, transmit the inspection result, and charge the user for the inspection result. Therefore, the user can request the inspection by the painted surface inspection device only when necessary without introducing the painted surface inspection device.

Further, the learning device according to the present invention includes training input information which is a photographed image of a painted surface of an object having a painted metal plate, and a teacher figure showing a deteriorated area of painting other than rust color in the photographed image. When a painted surface image, which is a photographed image of a painted surface of an object having a painted metal plate, is applied, which is a learning device that is a learning result of a plurality of sets with output information, the rust color included in the painted surface image is applied. It is possible to acquire an image showing a deteriorated region of coating other than the above.

Further, in the learning device according to the present invention, the input information for training, which is information on the rust-colored region and the deteriorated region of the coating other than the rust-colored on the painted surface of the object having the painted metal plate, and the painted surface of the object are abnormal. It is a learning device that is a learning result of a plurality of sets with the training output information that is the judgment result of whether or not, and relates to a rust-colored region and a paint deterioration region other than the rust-colored on the painted surface of the object to be judged. When the information is applied, it is possible to obtain a determination result of whether or not the painted surface of the object to be determined is abnormal.

Further, the learning device according to the present invention has training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information which is a determination result of whether or not the painted surface of the object is abnormal. When a painted surface image, which is a photographed image of a painted surface of an object having a painted metal plate, is applied to a learning device that is a learning result of a plurality of sets, a judgment result of whether or not the painted surface of the object is abnormal is applied. Can be obtained.

According to the painted surface inspection device or the like according to the present invention, it is possible to automatically inspect the painted surface of an object, for example, to identify a deteriorated region of paint other than rust color, and to check whether the painted surface is abnormal. It can be judged.

A block diagram showing a configuration of a painted surface inspection device according to the first embodiment of the present invention. A flowchart showing the operation of the painted surface inspection device according to the same embodiment. A flowchart showing the operation of the painted surface inspection device according to the same embodiment. The figure which shows an example of the painted surface image in the same embodiment The figure which shows an example of the specific result such as a rust color region in the same embodiment. The figure which shows an example of the output in the same embodiment A block diagram showing another example of the configuration of the painted surface inspection device according to the same embodiment. Schematic diagram showing an example of a communication network in the same embodiment The figure which shows an example of the training input information and the training output information in the same embodiment. A block diagram showing another example of the configuration of the painted surface inspection device according to the same embodiment. A flowchart showing another example of the operation of the painted surface inspection device according to the same embodiment. A block diagram showing a configuration of a painted surface inspection device according to a second embodiment of the present invention. A flowchart showing the operation of the painted surface inspection device according to the same embodiment.

Hereinafter, the painted surface inspection apparatus according to the present invention will be described with reference to embodiments. In the following embodiments, the components and steps having the same reference numerals are the same or correspond to each other, and the description thereof may be omitted again.

(Embodiment 1)
The painted surface inspection apparatus according to the first embodiment of the present invention will be described with reference to the drawings. The painted surface inspection device according to the present embodiment identifies a deteriorated region of paint other than rust color in a painted surface image of an object having a painted metal plate, and determines whether or not the painted surface is abnormal. is there.

FIG. 1 is a block diagram showing a configuration of a painted surface inspection device 1 according to the present embodiment. The painted surface inspection device 1 according to the present embodiment includes a reception unit 11, a storage unit 12, an area identification unit 13, an area output unit 14, a determination unit 15, and a determination result output unit 16. The inspection of the painted surface performed by the painted surface inspection device 1 may be, for example, identification of a deteriorated region of the coating on the painted surface, or determination of whether or not the painted surface is abnormal. In the present embodiment, the case where the inspection of the painted surface is both will be mainly described. Further, the painted surface inspection device 1 is provided with, for example, an imaging means or the like, and even if it is an information processing terminal (for example, a tablet terminal, a smartphone, etc.) capable of executing all processes from imaging to output of inspection results. Often, it may be an information processing device (for example, a stand-alone device, a server device in a server client system, etc.) that receives a captured image, performs an inspection using the image, and outputs the inspection result.

The reception unit 11 receives a painted surface image which is a photographed image of the painted surface of an object having a painted metal plate. The object having the painted metal plate is not particularly limited, but for example, an object having a housing of the painted metal plate, for example, a transformer such as a pillar transformer, a switch, or the like may be used and painted. It may be a structure made of a metal plate, for example, a structure such as a bridge, or another object having a painted metal plate. The painted surface is usually a surface that has been painted, but may be at least a surface that has been painted in the past. Therefore, as will be described later, at the time of photographing, it is possible that the paint does not remain on at least a part of the painted surface due to the deterioration of the paint.

The photographed image of the painted surface is usually a color photographed image. Strictly speaking, the captured image is the data of the captured image, but for convenience of explanation, it is simply referred to as the captured image. Further, the predetermined image processing may be performed before or after the painted surface image is received by the reception unit 11. The image processing may be to modify the shape of the painted surface to a predetermined shape in the painted surface image relating to the painted surface having a predetermined shape. Specifically, the image processing may be a circular correction that corrects the bottom surface of the cylindrical pole transformer from an elliptical shape to a circular shape in the captured image, and is any rectangle of the rectangular parallelepiped transformer. It may be a trapezoidal correction that corrects a surface from a trapezoidal shape to a rectangular shape in a captured image. Further, the image processing may be trimming that cuts out only a predetermined painted surface. Such image processing may be performed automatically, for example, or may be performed manually. In the present embodiment, a case where the painted surface image subjected to such image processing is received by the reception unit 11 will be mainly described.

The reception unit 11 may receive, for example, a painted surface image from a photographing means such as a camera, or may receive a painted surface image transmitted via a wired or wireless communication line, and may receive a predetermined recording medium (for example, a predetermined recording medium (for example). , Magnetic disk, semiconductor memory, etc.) may be accepted. The reception unit 11 may or may not include a device for receiving (for example, a modem or a network card). Further, the reception unit 11 may be realized by hardware, or may be realized by software such as a driver that drives a predetermined device.

The storage unit 12 stores setting information (for example, a threshold value described later) used for specifying an area by the area specifying unit 13 and setting information (for example, rules and conditions described later) used for determination by the determination unit 15. You may be. When a learning device is used for identifying or determining an area, the learning device may be stored in the storage unit 12.

The process of storing information in the storage unit 12 does not matter. For example, the information may be stored in the storage unit 12 via the recording medium, the information transmitted via the communication line or the like may be stored in the storage unit 12, or The information input via the input device may be stored in the storage unit 12. The storage in the storage unit 12 may be temporary storage in RAM or the like, or long-term storage. The storage unit 12 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, or the like).

The area specifying unit 13 specifies a rust-colored area, a painted area, and a deteriorated area of the coating other than the rust-colored area in the painted surface image received by the receiving unit 11. In the present embodiment, the case where the deteriorated region of the coating other than the rust color includes the cracking and peeling region of the coating and the swelling region of the coating will be mainly described. The rust-colored area is an area of the painted surface that has changed to a rust-colored color such as brown or black. The area of painting is the area of paint color on the painted surface. The rust-colored region and the painted region may be specified by using, for example, color information in the painted surface image. The cracked and peeled area of the coating is an area where the coating is cracked or peeled on the painted surface. Strictly speaking, the area where the paint is cracked or peeled off is not the painted surface, but as mentioned above, it is also called the painted surface. It is assumed that a layer of paint remains on the painted surface in areas other than the cracked and peeled areas of the paint. The swelling area of the coating is an area on the painted surface in which the coating layer is raised from the object to be coated and becomes a layer. The rust-colored region may include, for example, a plurality of regions for each rust color. For example, at least one or more of a black rust region, a white rust region, and a brown rust region may be included in the rust-colored region. The specific method for specifying the area will be described later. Further, the area specifying unit 13 may calculate the ratio (ratio) of the area of each specified area. The ratio may be, for example, the ratio of the area of the specified area to the total area of the painted surface to be specified as the area.

The area output unit 14 outputs each area specified by the area identification unit 13. This output may be, for example, the output of an image that visually indicates each specified region in the painted surface image, and information indicating whether or not each region is specified (for example, a cracked and peeled region of the coating). It may be information such as whether or not it has been specified and whether or not a swelling area of the paint has been specified, and is an output of data regarding each specified area (for example, the ratio of the area of each area to the painted surface). You may.

Here, this output may be displayed on a display device (for example, a liquid crystal display, an organic EL display, etc.), may be transmitted via a communication line, may be printed by a printer, or may be stored in a recording medium. , May be handed over to other components. The area output unit 14 may or may not include a device that outputs (for example, a display device, a communication device, a printer, etc.). Further, the area output unit 14 may be realized by hardware, or may be realized by software such as a driver that drives those devices.

The determination unit 15 determines whether or not the painted surface of the object is abnormal according to the information regarding each area specified by the area identification unit 13. Here, an abnormality in the painted surface of the object may mean, for example, a situation in which the painted surface needs to be repaired, repair is impossible, and the object must be discarded. There may be. The fact that the painted surface of the object is not abnormal means that, for example, the painted surface of the object does not need to be repaired and can be used continuously as it is. That is, if the painted surface is abnormal, it may be considered that the object having the painted surface cannot be continuously used as it is (at least, a situation requiring repair or the like). When the object is a transformer such as a pole transformer or a switch, for example, when it is determined to be abnormal, it will be collected and repaired or discarded as appropriate. When it is determined that there is no abnormality, it will continue to be used as it is.

For example, the determination unit 15 may make a determination according to a predetermined rule, or may make a determination using a learning device. The determination using the learner will be described later. When making a determination according to the rule, for example, a condition for determining an abnormality and / or a condition for determining a non-abnormal (normal) condition are stored in the storage unit 12, and the determination unit 15 stores the condition. , It may be determined whether or not the identification result by the area specifying unit 13 is abnormal by determining whether or not the condition is satisfied. Specifically, if there is a condition that the total ratio of the cracked and peeled area and the swelling area of the paint exceeds a predetermined ratio, it is abnormal, and when that condition is satisfied, It may be determined to be abnormal. In addition, when there are a plurality of conditions determined to be abnormal, it is determined to be abnormal when any of the conditions is satisfied, and it is determined not to be abnormal when all the conditions are not satisfied. May be done. In addition, when there are a plurality of conditions to be determined to be normal, it is determined to be normal when any of the conditions is satisfied, and it is determined to be abnormal when all the conditions are not satisfied. May be good.

The determination result output unit 16 outputs the determination result by the determination unit 15. This determination result is a determination result indicating whether or not the painted surface of the painted surface image received by the reception unit 11 is abnormal.

Here, this output may be displayed on a display device (for example, a liquid crystal display, an organic EL display, etc.), may be transmitted via a communication line, may be printed by a printer, or may be an audio output by a speaker. It may be stored in a recording medium or handed over to another component. The determination result output unit 16 may or may not include a device that outputs (for example, a display device, a communication device, a printer, etc.). Further, the determination result output unit 16 may be realized by hardware, or may be realized by software such as a driver that drives those devices.

Further, the area output unit 14 and the determination result output unit 16 may be configured as an integrated output unit. In that case, for example, the output unit functions as the area output unit 14 when outputting the specified area, and functions as the determination result output unit 16 when outputting the determination result. become.

Next, the operation of the painted surface inspection device 1 will be described with reference to the flowcharts of FIGS. 2 and 3.
(Step S101) The reception unit 11 determines whether or not the painted surface image has been received. Then, if the painted surface image is received, the process proceeds to step S102, and if not, the process of step S101 is repeated until the painted surface image is received.

(Step S102) The region specifying unit 13 specifies a rust-colored region, a deteriorated region of coating other than the rust-colored region, and the like in the painted surface image. The details of this process will be described later with reference to the flowchart of FIG.

(Step S103) The determination unit 15 determines whether or not the painted surface of the object is abnormal according to the information regarding each area specified by the area identification unit 13.

(Step S104) The area output unit 14 outputs each area specified by the area identification unit 13, and the determination result output unit 16 outputs the determination result by the determination unit 15. Then, a series of processes such as identification of the area related to the painted surface image and determination are completed, and the process returns to step S101.

The order of processing in the flowchart of FIG. 2 is an example, and the order of each step may be changed as long as the same result can be obtained. Further, in the flowchart of FIG. 2, the processing is terminated by an interrupt of power off or processing termination.

FIG. 3 is a flowchart showing the details of the region-specific process (step S102) in the flowchart of FIG. In the flowchart of FIG. 3, the identification of the region using the rule will be described, and the identification of the region performed by using the learning device will be described later.

(Step S201) The area specifying unit 13 specifies a black rust region. The region specifying unit 13 may specify, for example, a region having low lightness or a region having low lightness and low saturation as a black rust region. Specifically, the region specifying unit 13 may use a region whose brightness is smaller than the first threshold value as a black rust region. Further, the region specifying unit 13 may use a region whose brightness is smaller than the second threshold value and whose saturation is smaller than the third threshold value as the black rust region. The first threshold value is preferably smaller than the second threshold value, and may be set to, for example, a value of about the maximum value of the brightness of black rust. Further, the second threshold value may be set to, for example, a value of about the maximum value of the lightness of the black rust whose saturation is a value near the third threshold value. Further, the third threshold value may be, for example, the smaller value of a predetermined value and the minimum value of the saturation of the painted area (this minimum value will be described later). The predetermined value may be set to a value of about the maximum value of the saturation of black rust. When the minimum value of the saturation of the painted area becomes "0", the area specifying portion 13 may set the area where the lightness is smaller than the second threshold value as the black rust area. Further, the region specifying unit 13 may exclude the specified minute region from the specific result. This process may be referred to as "micro area exclusion process". Specifically, when the area of one specified black rust area (for example, the number of pixels) is smaller than a predetermined threshold value, the area specifying unit 13 may exclude the area from the black rust area. Good. Further, the region specifying unit 13 may incorporate a minute region existing between the specified regions or within the region into the specified region. This process is sometimes called "process of filling the gap". Specifically, when the region specifying unit 13 has a region other than the black rust region smaller than a predetermined threshold value in the specified black rust region, the region specifying unit 13 may use the small region as the black rust region. Good. Further, when the distance between the specified two or more black rust regions is smaller than a predetermined threshold value, the region specifying unit 13 may connect the two or more black rust regions to form one region.

(Step S202) The region specifying unit 13 specifies a region of white rust. The region specifying unit 13 may specify, for example, a region in which roughness has occurred and has high brightness in a region that has not been specified as a rust region, as a white rust region. Specifically, the region specifying portion 13 may have a region where roughness is generated and the brightness is higher than the average brightness of the painted region as a region of white rust. The portion of the painted surface corresponding to the region where the roughness is generated may be, for example, a portion where fine irregularities are generated on the surface. The region where the roughness occurs may be, for example, an region of an image in which a large number of point shapes are scattered like noise. Since the region where the roughness occurs is an region where fine edges are continuous on the image, the region identification unit 13 performs edge detection, and a process (closing) of agglomerating the detected edges to fill the gap. The result of the processing) may be the area where the roughness is generated. If the result of the closing process does not exceed the predetermined threshold value, the area resulting from the closing process does not have to be the area where the roughness occurs. The edge detection may be performed by, for example, the Canny method, or may be performed by another method of edge detection. The average brightness of the painted area may be obtained by tentatively specifying the painted area at this point and calculating the average brightness of each pixel of the tentatively specified painted area. When the average brightness of the painted area cannot be obtained (for example, when the painted area cannot be specified), a predetermined brightness threshold value may be used instead of the average brightness. It should be noted that the white rust region may also be subjected to a micro region exclusion process or a process of filling a gap.

(Step S203) The region specifying unit 13 specifies a region of tea rust. For example, the region specifying unit 13 may specify a region that is the color of brown rust as a region of brown rust in a region that has not been specified as a region of rust. Specifically, the region specifying unit 13 may use a region where the saturation is larger than the fourth threshold value and the hue is within a predetermined range as the brown rust region. It is preferable that the fourth threshold value is set to a value such that the difference in color becomes clear when the saturation is larger than the fourth threshold value. Further, it is preferable that the predetermined range of hue is set to the range of hue including the color of brown rust (for example, brown). It should be noted that the tea rust region may also be subjected to a micro region exclusion process or a process of filling the gap.

(Step S204) The area specifying portion 13 specifies a painted area. For example, the region specifying unit 13 may specify a region in the range of hue, saturation, and lightness according to painting as a painting region in a region that has not been specified as a rust region so far. Specifically, when the hue range, saturation range, and lightness range according to the painting are set in advance, the area specifying unit 13 uses the area included in these ranges as the painting area. May be good. Further, for example, the region specifying unit 13 receives a painted region (for example, a typical painted region may be designated by the user) instructed by the user in the painted surface image, and the hue is derived from the accepted region. , Saturation, and the upper and lower limits of lightness are acquired, and the range in which the upper and lower limits are expanded by predetermined values is used as the range of hue, saturation, and lightness as described above. The area of painting may be specified. The lower limit of the saturation range used in specifying the painted area may be used as the minimum value of the saturation of the painted area in step S201. As for the painted area, a process for excluding a minute area or a process for filling a gap may be performed.

(Step S205) The region specifying portion 13 may specify a region other than the specified rust-colored region and the painted region as a cracked and peeled region of the coating on the painted surface which is the target of specifying the region. This is because the region excluding the rust-colored region and the painted region is usually a cracked or peeled region of the paint. Specifically, the region specifying portion 13 may specify a region other than the black rust region, the white rust region, the brown rust region, and the coating region as the cracking and peeling regions of the coating.

(Step S206) In the region specified as the coating region in step S204, the region specifying unit 13 may specify a region in which the variance of brightness is larger than the fifth threshold value as a coating swelling region. When the coating has bulges, the change in brightness is usually large at the bulges. This is because the reflection of light changes due to the change in the surface shape according to the swelling, and as a result, the brightness changes. Therefore, since the brightness dispersion is large for the bulging portion, it is possible to specify the bulging portion by specifying the portion having such a large luminance dispersion. Specifically, the area specifying portion 13 divides the painted area into a range of a predetermined size (for example, a square of 20 pixels × 20 pixels), and each of the divided ranges. Then, the variance value of the brightness is calculated, and if the calculated variance value is larger than the fifth threshold value, it is determined that the range is the bulging region, and if not, the range is not the bulging region. It may be determined that such a process is repeated for each divided range. In that way, for each predetermined size range, it will be specified whether or not it is a swelling area of the paint. In the boundary portion with the rust-colored region, the change in brightness becomes large, and as a result, it may be erroneously specified as a swelling region of the coating. Therefore, it is possible not to perform the specific treatment of the swelling region of the coating on the boundary portion between the coating region and the rust-colored region. Further, before calculating the variance value of the luminance, for example, the averaging filter processing may be performed at a predetermined size. Further, since the bulging region is specified in the coating region, the region specifying portion 13 may exclude the bulging region specified in step S206 from the coating region specified in step S204. Therefore, strictly speaking, in step S204, the tentatively specified coating area is tentatively specified, and in step S206, the result of removing the swelling area of the coating from the tentatively specified coating area is the final result of the coating area. Specific results may be obtained. In this way, in the painted surface image, each region can be specified based on the rules. Then, the process returns to the flowchart of FIG.

The area specifying unit 13 may calculate the ratio of the area of each area after the identification of each area is completed. For example, the area specifying unit 13 may calculate the ratio of the area of each specified area to the area of the painted surface. Therefore, the area specifying unit 13 may calculate the area of the painted surface and the area of each specified area, and calculate the ratio of the area of each of the specified areas.

Further, the set value such as the threshold value used in specifying each region may be appropriately adjusted. Then, the threshold value and the like may be set so that suitable identification can be performed by comparing the identification result of each region by the region identification unit 13 with the actual painted surface. Further, the method for specifying each region is not limited to the above description, and it goes without saying that the method may be specified by another method. For example, in specifying the region of black rust and the region of brown rust, the occurrence of roughness may be used as a condition. This is because it is considered that the area where rust is generated is often rough. Needless to say, the rust-colored region may be specified by a method other than the above.

Next, the operation of the painted surface inspection device 1 according to the present embodiment will be described with reference to specific examples.
First, it is assumed that a photograph of the bottom surface of the pole transformer is taken, the bottom surface of the pole transformer is corrected to be circular by circular correction, and only the bottom surface portion is trimmed. FIG. 4 is a schematic view showing an example of such a painted surface image. In FIG. 4, the circular portion is the painted surface. When the user inputs the painted surface image of FIG. 4 into the painted surface inspection device 1, the painted surface image is received by the reception unit 11 and passed to the area identification unit 13 (step S101).

Upon receiving the painted surface image, the region specifying unit 13 first identifies the black rust region as described above (steps S102 and S201). The area specifying unit 13 may, for example, store information indicating the outline of the specified area in a recording medium (not shown).

Similarly, the region specifying unit 13 identifies the white rust region and the brown rust region (steps S202 and S203). Note that FIG. 5 is a diagram showing the result of specifying the region by the region specifying portion 13 in the painted surface image. In FIG. 5, the identified black rust region is the black region 52, the identified white rust region is the white region 54, and the identified brown rust region is the shaded region 53. Suppose it was.

Next, the region specifying unit 13 specifies a painted region in a region other than the rust identified so far (step S204). In FIG. 5, it is assumed that the painted area is the shaded area 51. After that, the region specifying portion 13 specifies a rust-colored region and a coating cracking and peeling region other than the coating region (step S205). In FIG. 5, it is assumed that the cracked and peeled area of the coating is the shaded area 55.

Next, the region specifying unit 13 identifies a bulging region, which is a region having a large brightness dispersion, in the painted region (step S206). In FIG. 5, it is assumed that the bulging region is the region 56 whose outline is indicated by the broken line. Then, the region specifying unit 13 excludes the region 56 from the painted region. Then, the area specifying unit 13 calculates the ratio of the area of each specified area.

After that, the determination unit 15 determines whether or not the painted surface is abnormal by using the ratio of each region (step S103). Here, it is assumed that the painted surface is not abnormal according to the specific result of FIG. Then, the area output unit 14 outputs the specific result of the area, and the determination result output unit 16 outputs the determination result that the painted surface is not abnormal (step S104). Specifically, as shown in FIG. 6, the area output unit 14 may display a painted surface image, an image showing a specific result, and information showing a ratio of each specified area. Further, the determination result output unit 16 may display a determination result indicating that the painted surface is not abnormal together with or separately from the specific result of the area.

As described above, according to the painted surface inspection device 1 according to the present embodiment, the deteriorated region of the coating other than the rust color, for example, the cracking and peeling region of the coating and the swelling region of the coating are specified by using the photographed image. can do. In addition, since it is possible to make a judgment on the painted surface using the deteriorated region of the coating other than the rust color, the judgment can be made with higher accuracy than the case where the judgment is made based only on the rust color. Will be able to. Furthermore, since those areas can be automatically identified and determined, it becomes possible to realize an inspection that does not depend on the sense or experience of the observer. In addition, when the object is a pole transformer, by taking a picture of the bottom surface from the ground without climbing the utility pole, it becomes possible to identify the area, determine an abnormality, and so on. The burden on the poles will be significantly reduced.

Next, a modification of the painted surface inspection device 1 according to the present embodiment will be described.
[Billing process]
The painted surface inspection device 1 may receive the painted surface image, specify the region and determine the painted surface using the received painted surface image, and transmit the results. Then, a charge may be made according to the transmission. In that case, as shown in FIG. 7, the painted surface inspection device 1 may further include a charging unit 17. In FIG. 7, the reception by the reception unit 11 may be the reception of the reception target, and the output by the area output unit 14 and the determination result output unit 16 may be the transmission of the output target. The destination of the output target may be, for example, the source of the painted surface image, or may be another destination. In the latter case, for example, the address of the transmission destination to be output may be received by the reception unit 11 together with the painted surface image. Further, in order to specify the billing destination, the reception unit 11 may receive the requester identifier that identifies the requester together with the painted surface image.

The billing unit 17 charges according to the output of the specific result of the area and the determination result. The billing destination may be, for example, a requester corresponding to the destination to be output, and if the requester identifier is accepted by the reception unit 11, the requester is identified by the requester identifier. You may. Specifically, the billing unit 17 may output the requester identifier that identifies the requester of the transmission destination and the amount of money corresponding to the transmission in association with each other in response to the transmission of the specific result of the area or the like. .. Then, depending on the output information, a request to the requester may be made as appropriate. Further, when the information such as the credit card of the requester is registered in advance, the billing unit 17 corresponds to the requester identifier that identifies the requester of the destination in response to the transmission of the specific result of the area or the like. Information such as a credit card may be read out, and the read information may be used to perform billing processing corresponding to the amount of billing. The billing process is already known, and detailed description thereof will be omitted.

In this case, the painted surface inspection device 1 may be a server capable of communicating with the information processing terminal 5 via a wired or wireless communication line 500, as shown in FIG. Then, the information processing terminal 5 of the client who requests the inspection of the painted surface receives the painted surface image, and the information processing terminal 5 of the requester receives the specific result and the determination result which are the inspection results of the received painted surface image. May be sent to. The communication line 500 may be, for example, the Internet, an intranet, a public telephone line network, or the like. Further, in each of the painted surface inspection devices 2 and 3 described later, a charge may be made according to the output.

[Identification of area using learner]
As described above, the area specifying unit 13 may specify the area by using the learning device. Here, the identification of the region using the learner will be described. First, the learning device used to identify the area will be described. This learning device includes training input information which is a photographed image of a painted surface of an object having a painted metal plate, and training output information including a teacher figure showing a deteriorated area of painting other than rust color in the photographed image. It is the learning result of multiple sets of. The combination of the training input information and the training output information is sometimes called training information. The learner may be, for example, a learning result of a neural network (NN), or may be a learning result of other machine learning.

The neural network may be, for example, a convolutional neural network (CNN) or another neural network (for example, a neural network composed of fully connected layers). A convolutional neural network is a neural network having one or more convolutional layers. Further, when the neural network has at least one intermediate layer (hidden layer), the learning of the neural network may be considered as deep learning (deep learning). In addition, when a neural network is used for machine learning, the number of layers of the neural network, the number of nodes in each layer, the type of each layer (for example, convolution layer, fully connected layer, etc.), etc. are appropriately selected. May be good. The number of nodes in the input layer and the output layer is usually determined by the number of pixels of the training input information included in the training information and the number of pixels of the training output information. In the present embodiment, the case where the learning device is the learning result of CNN will be mainly described.

As the learning result of CNN, for example, the learning result of the neural network used in segmentation may be used. That is, the structure of the neural network may be similar to that of segmentation. The segmentation neural network has, for example, a plurality of convolution layers in the front stage, and one or more layers such as an unpooling layer and a deconvolution layer in the rear stage for enlarging the image. You may have. In this learner, the input and the output need to have the same number of pixels, but the number of pixels usually decreases due to the convolution layer in the previous stage, so it is necessary to increase the number of pixels in the latter stage. Because.

In addition, bias may or may not be used in each layer. Whether or not to use the bias may be decided independently for each layer. The bias may be, for example, a layer-by-layer bias or a filter-by-filter bias. In the former case, one bias is used in each layer, and in the latter case, one or more biases (the same number as the filter) are used in each layer. When bias is used in the convolution layer, the result of multiplying each pixel value by the parameter of the filter and adding them together, and adding the bias, is input to the activation function. The same applies to other learners.

In addition, each setting in the neural network may be as follows. The activation function may be, for example, a ReLU (normalized linear function), a sigmoid function, or another activation function. Further, in learning, for example, the error back propagation method may be used, or the mini-batch method may be used. Further, the loss function (error function) may be a mean square error. The number of epoches (number of parameter updates) is not particularly limited, but it is preferable to select the number of epoches that do not cause overfitting. As a learning method in machine learning, a known method can be used, and detailed description thereof will be omitted. The same applies to other learners.

The learning device used to identify the region may be stored in the storage unit 12. Note that the learning device is stored in the storage unit 12, for example, even if the learning device itself (for example, a function that outputs a value with respect to an input, a model of a learning result, etc.) is stored. Often, information such as parameters required to configure the learner may be stored. Even in the latter case, since the learning device can be configured by using the information such as the parameters, it can be considered that the learning device is substantially stored in the storage unit 12. In the present embodiment, the case where the learning device itself is stored in the storage unit 12 will be mainly described. It is assumed that this also applies to other learning devices stored in other storage units.

Here, the generation of the learner will be described. As described above, the training input information is a photographed image of the painted surface of the object having the painted metal plate. In the training input information, it is preferable that correction such as circular correction or keystone correction, trimming, etc. are performed. The training output information may include a teacher figure indicating each region manually specified in the captured image of the training input information to be paired. That is, the result of setting the teacher figure in each area in the captured image may be the training output information. For example, as shown in FIG. 9, the training output information includes a coating area 51, a black rust area 52, a brown rust area 53, a white rust area 54, a coating cracking and peeling area 55, and a coating swelling. The area 56 may include a teacher figure showing different types of shading and other attributes. It is preferable that different regions are learned using teacher figures with different attributes. The attribute may be, for example, a pattern such as shading, a color (hue), lightness or saturation, or other attributes. It is preferable that the training output information includes a teacher figure relating to the area to be specified. For example, when it is desired to specify a rust-colored region, a paint deterioration region other than the rust-colored region, and a paint region, the three types of teacher figures may be included in the training output information.

A learner is manufactured by learning a plurality of sets of training input information and training output information. Then, when the painted surface image, which is a photographed image of the painted surface of the object having the painted metal plate, is applied to the learner, for example, an image showing a deteriorated region of the paint other than the rust color included in the painted surface image. Can be obtained. Specifically, the area identification unit 13 acquires an image divided into areas for each attribute by applying the painted surface image received by the reception unit 11 to the learning device stored in the storage unit 12. can do. That is, when the painted surface image is input to the learning device, the image divided into the areas for each attribute is output (for example, the same image as the training output information of FIG. 9 is output. ). The area specifying unit 13 can specify each area by specifying the area for each attribute in the output. Specifically, the black rust region can be specified by specifying the attribute region corresponding to the attribute of the teacher figure corresponding to the black rust region in the output image of the learner. The same applies to other areas. Further, the area specifying unit 13 may calculate, for example, the ratio of the area in the area of the painted surface for each specified area.

In FIG. 9, the case where the training output information includes a plurality of types of teacher figures has been described, but this may not be the case. One training output information may include only one type of teacher figure. In that case, for example, training output information including only the teacher figure corresponding to the black rust region, training output information including only the teacher figure corresponding to the cracked and peeled paint region, and the like will be used. ..

By specifying the area using the learner in this way, it is not necessary to set rules and conditions for specifying the area and adjust the threshold value, etc., so that the burden on them is reduced. The merit of being done is obtained. In addition, the accuracy can be improved by increasing the number of training information used in learning.

[Judgment using a learning device]
As described above, the determination unit 15 may make a determination using the learning device. Here, the determination using the learner will be described. First, the learning device used in the determination will be described. This learning device is used for training input information, which is information on a rust-colored region and a non-rust-colored coating deterioration region on the painted surface of an object having a painted metal plate, and whether or not the painted surface of the object is abnormal. It is a learning result of a plurality of sets with the training output information which is a judgment result. Regarding the training input information, the information regarding the rust-colored region may be, for example, information for each of the black rust region, the white rust region, and the brown rust region, and relates to the deteriorated region of the coating other than the rust color. The information may be, for example, information for each of the cracked and peeled areas of the coating and the swelling area of the coating. The training input information may also include information about the area of painting. The information regarding each region may be information indicating the ratio of the area of each region on the painted surface. The learner may be, for example, a learning result of a neural network, a learning result of a support vector machine (SVM), or a learning result of other machine learning.

The neural network may be, for example, a CNN, a neural network composed of fully connected layers, or another neural network. When a neural network is used for machine learning, the number of layers of the neural network, the number of nodes in each layer, the type of each layer (for example, convolution layer, fully connected layer, etc.) and the like may be appropriately selected. .. The number of nodes in the input layer and the output layer is usually determined by the number of training input information and the number of training output information included in the training information. Since the output of this learner is abnormal or not, the number of nodes in the output layer may be one or two. When the number of nodes in the output layer is 1, for example, a determination result (for example, normal or abnormal) may be shown depending on whether the output value is 0 or 1. On the other hand, when the number of nodes in the output layer is two, for example, it is determined depending on whether the output value of the node corresponding to the normal is 1 or the output value of the node corresponding to the abnormality is 1. Results may be shown. As described above, when the number of nodes in the output layer is two, for example, a softmax layer may be provided in front of the output layer. Here, first, the case where the learner is the learning result of the neural network will be described, and the case where the learner is the SVM will be described later.

Here, the generation of the learner will be described. As described above, the training input information may be information indicating the ratio of the area of each region. The training output information is information indicating whether or not the painted surface corresponding to the training input information is abnormal. As the training information, for example, information on each area specified by a person and a result determined by the person may be used. A learner is manufactured by learning a plurality of sets of training input information and training output information. Then, when information on the rust-colored region on the painted surface of the object to be judged and the deteriorated region of the paint other than the rust-colored is applied to the learner, whether or not the painted surface of the object to be judged is abnormal. The judgment result can be acquired. That is, when information indicating the ratio of the area of each region is input to the learner, information indicating whether or not it is abnormal is output. In this way, the determination using the learner can be performed.

In the learning of SVM, the training information is the same as the learning of the neural network. Then, even when the determination is performed using the learner which is the learning result of the SVM, the determination result is output by inputting the information about each region in the same manner as the neural network. Since the SVM is usually a binary determination device, it is suitable for such a determination. In this way, the determination regarding the painted surface can also be made by using the SVM.

In this way, by making a judgment using a learning device, it is not necessary to set rules and conditions for judgment and adjust the set values, so that the burden on them is reduced. Is obtained. In addition, the accuracy can be improved by increasing the number of training information used in learning.

[Painted surface inspection device for judgment]
In the present embodiment, the case where the painted surface inspection device 1 makes a determination together with the identification of the region has been described, but it is not necessary. The painted surface inspection device does not have to specify the area. In that case, as shown in FIG. 10, the painted surface inspection device 2 may include a reception unit 21, a storage unit 22, a determination unit 15, and a determination result output unit 16. .. The determination unit 15 and the determination result output unit 16 are the same as those described above, and detailed description thereof will be omitted.

The reception unit 21 receives information on a rust-colored region and a non-rust-colored coating deterioration region on the painted surface of the object having the painted metal plate. The information regarding the rust-colored region received by the reception unit 21 may be, for example, information for each of the black rust region, the white rust region, and the brown rust region. Further, the information regarding the deterioration region of the coating other than the rust color received by the reception unit 21 may be, for example, information for each of the cracking and peeling regions of the coating and the swelling region of the coating. The reception unit 21 may also receive information regarding the area of painting. The information regarding each region may be information indicating the ratio of the area of each region on the painted surface. The information regarding each area received by the reception unit 21 is the same as the information acquired by the area identification unit 13, and detailed description thereof will be omitted.

The reception unit 21 may receive, for example, information input from an input device (for example, a keyboard, mouse, touch panel, etc.), or may receive information transmitted via a wired or wireless communication line. Information read from a predetermined recording medium (for example, a magnetic disk, a semiconductor memory, etc.) may be accepted. The reception unit 21 may or may not include a device for receiving (for example, a modem or a network card). Further, the reception unit 21 may be realized by hardware, or may be realized by software such as a driver that drives a predetermined device.

The storage unit 22 stores information used for determination. For example, the conditions used in the determination, the learning device, and the like may be stored in the storage unit 22. The process of storing information in the storage unit 22 does not matter. For example, the information may be stored in the storage unit 22 via the recording medium, the information transmitted via the communication line or the like may be stored in the storage unit 22, or The information input via the input device may be stored in the storage unit 22. The storage in the storage unit 22 may be temporary storage in RAM or the like, or long-term storage. The storage unit 22 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, or the like).

The determination unit 15 is the same as the determination unit 15 except that the determination is performed using the information about each area received by the reception unit 21 instead of the information about each area specified by the area identification unit 13. Yes, and the detailed description thereof will be omitted. As described above, the determination unit 15 may make a determination using, for example, a rule, or may make a determination using a learner.

Next, the operation of the painted surface inspection device 2 will be described with reference to the flowchart of FIG. In the processes other than step S301, the determination unit 15 makes a determination using the information about each area received by the reception unit 21 instead of the information about each area specified by the area identification unit 13. It is the same as the flowchart of FIG. 2, and the description thereof will be omitted.

(Step S301) The reception unit 21 determines whether or not information regarding each region on the painted surface of the object having the painted metal plate has been received. Then, when the information regarding each area is received, the process proceeds to step S105, and when not, the process of step S301 is repeated until the information about each area is received.

The order of processing in the flowchart of FIG. 11 is an example, and the order of each step may be changed as long as the same result can be obtained. Further, in the flowchart of FIG. 11, the process is terminated by the power off or the interrupt of the process termination.

[Other]
In the present embodiment, the case where the painted surface inspection device 1 includes the area output unit 14 has been described, but it is not necessary. When it is not necessary to output for each of the specified regions, the painted surface inspection device 1 may not include the region output unit 14.

Further, in the present embodiment, the case where the painted surface inspection device 1 includes the determination unit 15 and the determination result output unit 16 has been described, but it is not necessary. When it is not necessary to perform the determination, the painted surface inspection device 1 may not include the determination unit 15 and the determination result output unit 16.

Further, in the present embodiment, the case where the region specifying portion 13 specifies both the cracked and peeled region of the coating and the swelling region of the coating as the deteriorated region of the coating other than the rust color has been described. It does not have to be. The area specifying unit 13 may specify one of the areas. Further, when specifying only the swelling region of the coating, the region specifying portion 13 does not have to specify the rust-colored region. Further, in the present embodiment, as an example, a case where the rust-colored region includes a black rust region, a white rust region, and a brown rust region has been described, but the rust-colored region is at least one of them. It is not necessary to include the region of rust, and the region of rust of other colors (for example, the region of red rust, blue rust, etc.) may be included.

(Embodiment 2)
The painted surface inspection apparatus according to the second embodiment of the present invention will be described with reference to the drawings. The painted surface inspection device according to the present embodiment determines whether or not the painted surface of the object is abnormal by applying a photographed image of the object having the painted metal plate to the learning device.

FIG. 12 is a block diagram showing the configuration of the painted surface inspection device 3 according to the present embodiment. The painted surface inspection device 3 according to the present embodiment includes a reception unit 11, a storage unit 31, a determination unit 32, and a determination result output unit 33. The configuration and operation of the reception unit 11 are the same as those in the first embodiment, and detailed description thereof will be omitted.

In the storage unit 31, the learning device used for the determination by the determination unit 32 is stored. The learning device has a plurality of sets of training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information which is a judgment result of whether or not the painted surface of the object is abnormal. It is a learning result learned by using. The learner may be, for example, a learning result of a neural network or a learning result of other machine learning. The neural network may be, for example, CNN or another neural network (for example, a neural network composed of fully connected layers). The learner may be a CNN learning result in which the number of nodes in the output layer is one or two. This CNN may have the same configuration as the CNN used in object recognition, for example.

Here, the generation of the learner will be described. The training input information is a photographed image of the painted surface of the object having the painted metal plate. In the training input information, it is preferable that correction such as circular correction or keystone correction, trimming, etc. are performed. Further, the training output information is a determination result indicating whether or not the painted surface of the object corresponding to the training input information is abnormal. The determination result of the training output information may be, for example, a result determined by a person. A learner is manufactured by learning a plurality of sets of training input information and training output information.

The process in which the learning device is stored in the storage unit 31 does not matter. For example, the learning device may be stored in the storage unit 31 via the recording medium, or the learning device transmitted via the communication line or the like may be stored in the storage unit 31. The storage in the storage unit 31 may be temporary storage in RAM or the like, or long-term storage. The storage unit 31 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, or the like).

The determination unit 32 determines whether or not the painted surface of the object to be determined is abnormal by applying the painted surface image to the learning device stored in the storage unit 31. That is, when the painted surface image is input to the learner, information indicating whether or not it is abnormal is output. In this way, the determination using the learner can be performed.

The determination result output unit 33 outputs the determination result by the determination unit 32. The determination result output unit 33 is the same as the determination result output unit 16 except that the determination result of the determination unit 32 is output instead of the determination result output of the determination unit 15, and a detailed description thereof will be given. Omit.

Next, the operation of the painted surface inspection device 3 will be described with reference to the flowchart of FIG. The process of step S101 is the same as the flowchart of FIG. 2, and the description thereof will be omitted.
(Step S401) The determination unit 32 acquires the determination result by applying the painted surface image received by the reception unit 11 to the learning device stored in the storage unit 31.

(Step S402) The determination result output unit 33 outputs the determination result acquired in step S401. Then, the process returns to step S101.
In the flowchart of FIG. 13, the process ends when the power is turned off or an interrupt for the end of the process occurs.

As described above, according to the painted surface inspection device 3 according to the present embodiment, by applying the painted surface image to the learning device, it is possible to determine whether or not the painted surface corresponding to the painted surface image is abnormal. Can be obtained. In this way, since the determination can be made automatically, it becomes possible to realize an inspection that does not depend on the sense or experience of the observer. In addition, by making a judgment using a learning device, it is not necessary to set rules and conditions for judgment and adjust the set values, which has the advantage of reducing the burden on them. Be done. In addition, the accuracy can be improved by increasing the number of training information used in learning.

In each of the above embodiments, each process or each function may be realized by centralized processing by a single device or a single system, or distributed processing by a plurality of devices or a plurality of systems. It may be realized by being done.

Further, in each of the above-described embodiments, the transfer of information performed between the components is, for example, one component when the two components that transfer the information are physically different. It may be performed by the output of information by and the reception of information by the other component, or when the two components that pass the information are physically the same, one component is used. It may be performed by moving from the processing phase corresponding to the element to the processing phase corresponding to the other component.

Further, in each of the above embodiments, information related to the process executed by each component, for example, received, acquired, selected, generated, transmitted, or received by each component. Information, threshold values, mathematical formulas, addresses, and other information used by each component in processing may be temporarily or for a long period of time in a recording medium (not shown), even if they are not specified in the above description. .. In addition, each component or a storage unit (not shown) may store information on a recording medium (not shown). Further, the information may be read from the recording medium (not shown) by each component or a reading unit (not shown).

Further, in each of the above embodiments, when the information used in each component or the like, for example, the information such as the threshold value and the address used in the processing by each component and various setting values may be changed by the user, Although not specified in the above description, the user may or may not be able to change the information as appropriate. When the information can be changed by the user, the change is realized by, for example, a reception unit (not shown) that receives a change instruction from the user and a change unit (not shown) that changes the information in response to the change instruction. You may. The reception unit (not shown) may accept the change instruction from, for example, an input device, information transmitted via a communication line, or information read from a predetermined recording medium. ..

Further, in each of the above embodiments, when the two or more components included in the painted surface inspection devices 1, 2, and 3 have a communication device, an input device, or the like, the two or more components are physically single. It may have a device, or it may have a separate device.

Further, in each of the above embodiments, each component may be configured by dedicated hardware, or a component that can be realized by software may be realized by executing a program. For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. At the time of execution, the program execution unit may execute the program while accessing the storage unit or the recording medium. The software that realizes the painted surface inspection devices 1, 2, and 3 according to the above embodiment may be, for example, the following program. That is, the program uses the computer as a reception unit that receives a painted surface image that is a photographed image of a painted surface of an object having a painted metal plate, and an area that at least specifies a deteriorated region of paint other than rust color in the painted surface image. It may be for functioning as a specific part. In addition, the program uses the computer as a reception unit that receives information on the rust-colored region and the deterioration region of the coating other than the rust-colored region on the painted surface of the object having the painted metal plate, and information on each region received by the reception unit. Depending on the situation, it may function as a determination unit for determining whether or not the painted surface of the object is abnormal, and a determination result output unit for outputting the determination result by the determination unit. In addition, the program uses the computer as a reception unit that receives a painted surface image that is a photographed image of the painted surface of an object having a painted metal plate, and training input information that is a photographed image of the painted surface of an object having a painted metal plate. By applying the painted surface image to a learner learned by using a plurality of pairs of the painted surface and the training output information which is the result of determining whether the painted surface of the object is abnormal, the object to be determined is determined. It may function as a determination unit for determining whether or not the painted surface is abnormal, and a determination result output unit for outputting the determination result by the determination unit.

In the above program, the functions realized by the above program do not include functions that can be realized only by hardware. For example, functions realized only by hardware such as a modem and an interface card in a reception unit that receives information and an output unit that outputs information are not included in at least the functions realized by the above program.

Further, the program may be executed by being downloaded from a server or the like, or may be executed by reading a program recorded on a predetermined recording medium (for example, an optical disk, a magnetic disk, a semiconductor memory, etc.). Good. Further, this program may be used as a program constituting a program product. Further, the number of computers that execute the program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

Further, it goes without saying that the present invention is not limited to the above embodiments, and various modifications can be made, and these are also included in the scope of the present invention.

From the above, according to the painted surface inspection device or the like according to the present invention, it is possible to automatically inspect the painted surface of the object. For example, it can be determined whether or not the painted surface of the object is abnormal. It is useful as a judgment device or the like.

1, 2, 3 Painted surface inspection device, 11, 21 Reception unit, 13 area identification unit, 14 area output unit, 15, 32 judgment unit, 16, 33 judgment result output unit, 17 billing unit

Claims (16)

The reception section that accepts the painted surface image, which is the photographed image of the painted surface of the object having the painted metal plate,
A painted surface inspection device including a region specifying portion for at least specifying a deteriorated region of coating other than rust color in the painted surface image. The deterioration area of the coating other than the rust color includes the cracking and peeling area of the coating.
The painted surface inspection apparatus according to claim 1, wherein the area specifying portion specifies a rust-colored area and a painted area in the painted surface image, and specifies a region other than the specified area as a cracked and peeled area of the paint. .. The deterioration area of the coating other than the rust color includes the swelling area of the coating.
The area specifying portion identifies a painted area in the painted surface image, and in the specified painted area, specifies a swelling area of the coating, which is a region in which the dispersion of brightness is larger than the threshold value, claim 1 or The painted surface inspection device according to claim 2. The area specifying portion includes training input information which is a photographed image of a painted surface of an object having a painted metal plate, and training output information including a teacher figure showing a deteriorated area of painting other than rust color in the photographed image. The painted surface inspection according to claim 1, wherein a deteriorated region of paint other than the rust color contained in the painted surface image is specified by applying the painted surface image to a learning device learned by using a plurality of pairs of apparatus. The painted surface inspection apparatus according to any one of claims 1 to 4, further comprising an area output unit that outputs an output for each area specified by the area identification unit. The region specifying portion also specifies a rust-colored region in the painted surface image.
A determination unit that determines whether or not the painted surface of the object is abnormal according to the information regarding each region specified by the region identification unit.
The painted surface inspection device according to any one of claims 1 to 5, further comprising a determination result output unit that outputs a determination result by the determination unit. A reception section that receives information on rust-colored areas and non-rust-colored paint deterioration areas on the painted surface of an object having a painted metal plate.
A determination unit that determines whether or not the painted surface of the object is abnormal according to the information regarding each area received by the reception unit.
A painted surface inspection device including a determination result output unit that outputs a determination result by the determination unit. The determination unit includes training input information which is information on a rust-colored region and a deterioration region of coating other than rust-colored on the painted surface of the object having a painted metal plate, and whether or not the painted surface of the object is abnormal. By applying information about each region on the painted surface of the object to be judged to the learner learned by using a plurality of sets of the output information for training which is the judgment result, the painted surface of the object to be judged The painted surface inspection device according to claim 6 or 7, wherein it is determined whether or not the rust is abnormal. The reception section that accepts the painted surface image, which is the photographed image of the painted surface of the object having the painted metal plate,
It was learned using a plurality of sets of training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information which is a judgment result of whether or not the painted surface of the object is abnormal. By applying the painted surface image to the learner, a determination unit that determines whether the painted surface of the object to be determined is abnormal, and a determination unit.
A painted surface inspection device including a determination result output unit that outputs a determination result by the determination unit. The reception is the reception of the reception target,
The output is a transmission to be output.
The painted surface inspection device according to any one of claims 5 to 9, further comprising a charging unit that charges according to the output. A plurality of sets of training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information including a teacher figure showing a deteriorated area of painting other than rust color in the photographed image. It is a learning device that is the learning result,
When a painted surface image, which is a photographed image of a painted surface of an object having a painted metal plate, is applied, it is possible to acquire an image showing a deteriorated region of the paint other than the rust color included in the painted surface image. vessel. Training input information that is information on a rust-colored region and a deterioration region of coating other than rust-colored on the painted surface of an object having a painted metal plate, and training that is a judgment result of whether or not the painted surface of the object is abnormal. It is a learner that is the learning result of multiple sets with the output information for
When information on the rust-colored area and the deterioration area of the coating other than the rust-colored area on the painted surface of the object to be judged is applied, the judgment result of whether or not the painted surface of the object to be judged is abnormal is acquired. A learning device that can be used. It is a learning result of a plurality of sets of training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information which is a determination result of whether or not the painted surface of the object is abnormal. It ’s a learning device,
A learning device capable of acquiring a determination result of whether or not the painted surface of the object is abnormal when the painted surface image, which is a photographed image of the painted surface of the object having the painted metal plate, is applied. Computer,
Reception desk that accepts painted surface images, which are captured images of painted surfaces of objects with painted metal plates,
A program for functioning as a region specifying portion for at least specifying a deteriorated region of coating other than rust color in the painted surface image. Computer,
A reception section that receives information on rust-colored areas and non-rust-colored paint deterioration areas on the painted surface of objects with painted metal plates.
A determination unit that determines whether or not the painted surface of the object is abnormal according to the information regarding each area received by the reception unit.
A program for functioning as a judgment result output unit that outputs a judgment result by the judgment unit. Computer,
Reception desk that accepts painted surface images, which are captured images of painted surfaces of objects with painted metal plates,
It was learned using a plurality of sets of training input information which is a photographed image of a painted surface of an object having a painted metal plate and training output information which is a judgment result of whether or not the painted surface of the object is abnormal. A determination unit that determines whether or not the coated surface of the object to be determined is abnormal by applying the painted surface image to the learner.
A program for functioning as a judgment result output unit that outputs a judgment result by the judgment unit.

Images