Image correction device, image correction method, and image correction program

Abstract

To provide an image correction device, an image correction method and an image correction program, which can correct a captured image to an image suitable for a diagnosis in order to perform state recording or state diagnosis of an engine by using the captured image obtained by imaging a piston of the engine that is an inspection target.SOLUTION: An image correction device corrects a captured image for diagnosing an engine by using the captured image obtained by imaging a piston of the engine that is an inspection target or a peripheral structure of the piston. The image correction device includes: a storage unit storing a reference parameter related to a predetermined appearance reference part of the piston, which is predetermined in accordance with the engine; and a computing unit that performs image correction operation of the captured image. The computing unit acquires the captured image, acquires the reference parameter of the engine corresponding to the captured image, extracts a characteristic part corresponding to the appearance reference part from the captured image, generates a characteristic parameter corresponding the reference parameter from the characteristic part, and corrects the captured image on the basis of a comparison between the characteristic parameter and the reference parameter.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present disclosure relates to an image correction device, an image correction method, and an image correction program. [Background technology]

[0002] In large engines such as marine two-stroke diesel engines, periodic checking of cylinder condition is important to maintain engine performance during operation. Cylinder condition refers to the condition of sliding parts, including pistons and cylinder liners that slide as the pistons move up and down, combustion chamber components, including pistons, cylinder liners, and cylinder covers, as well as the condition of the oil supply system. However, in large marine engines, the pistons are large, making it difficult to remove them from the cylinders. Therefore, to check the cylinder condition in such engines, images of the pistons inside the cylinders are taken through scavenging ports on the sides of the cylinders, and the conditions of the pistons and piston rings attached to the pistons are recorded or diagnosed based on the captured images.

[0003] More specifically, the internal space of the cylinder of such an engine is connected to a scavenging pipe via a scavenging port formed on the side of the lower part of the cylinder. The scavenging pipe is large enough for a person to enter. A photographer, such as a ship's crew member or maintenance worker, enters the scavenging pipe when the engine is stopped and photographs the piston and piston rings inside the cylinder through the scavenging port. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 3640726 [Patent Document 2] Japanese Patent Application Laid-Open No. 2003-157408 Summary of the Invention [Problem to be solved by the invention]

[0005] The scavenging ports are oriented at a predetermined angle in the circumferential direction relative to the radial direction of the cylinder to ensure smooth scavenging air flow. Therefore, when photographing the piston and piston rings inside the cylinder through the scavenging ports, the photographer must photograph them at an angle relative to the radial direction of the cylinder to match the orientation of the scavenging ports and prevent the piston and piston rings from being shaded by the scavenging ports. As a result, the piston and piston rings seen through the scavenging ports in the photographed image have a difference in perspective in the horizontal direction. Furthermore, because the subjects being photographed are a cylindrical piston and annular piston rings, the photographed image contains distortion of the curved surfaces.

[0006] For these reasons, it becomes difficult to detect sliding scratches, which are important in diagnosing the condition of piston rings, and there is a risk that the condition of piston rings cannot be properly diagnosed.

[0007] In diagnosing the condition using such images, Patent Document 1 discloses a method of photographing the annular component to be inspected by rotating the component or by revolving the camera around the component, and then detecting edges from the photographed images. However, Patent Document 1 is based solely on the premise of measuring the accuracy at the time of product shipment and does not anticipate inspections after the component is installed on a ship. In other words, photographing the piston in the narrow scavenging pipe while it is still in the cylinder makes it impossible to photograph the entire outer circumference of the piston. Therefore, this method cannot be used to check the cylinder condition during operation.

[0008] Furthermore, Patent Document 2 discloses a method for associating each feature point in an undistorted reference image with a corresponding point in an input image that includes distortion. However, this method is based on the assumption that the position of a feature point in the reference image and the position of the input image corresponding to the feature point are located at the same position. Therefore, this method cannot be applied to images taken for engine condition diagnosis, such as those described above, where the position and orientation of the camera relative to the object of inspection are likely to change with each image capture.

[0009] Therefore, the present disclosure aims to provide an image correction device, an image correction method, and an image correction program that can correct captured images of the pistons of the engine being inspected into images suitable for diagnosis in order to record or diagnose the condition of the engine using the captured images. [Means for solving the problem]

[0010] An image correction device according to one aspect of the present disclosure is an image correction device that corrects a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine using the photographed image, and includes: a memory that stores reference parameters related to a predetermined appearance reference part of the piston that is predetermined according to the engine; and a calculator that performs image correction calculations for the photographed image, and the calculator acquires the photographed image and calculates the Related to the specified appearance standard part of the piston and acquiring reference parameters from the captured image. The piston's predetermined A feature portion corresponding to the appearance reference portion is extracted, and the Related to the specified appearance standard part of the piston generating feature parameters corresponding to the reference parameters, and based on a comparison between the feature parameters and the reference parameters, so that the generated feature parameter value becomes the corresponding reference parameter value, The captured image is corrected.

[0011] An image correction method according to another aspect of the present disclosure is an image correction method for correcting a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine using the photographed image, the image correction method comprising: acquiring the photographed image; acquiring reference parameters related to a predetermined appearance reference part of the piston that is predetermined for the engine corresponding to the photographed image; and correcting the reference parameters from the photographed image. The piston's predetermined A feature portion corresponding to the appearance reference portion is extracted, and the Related to the specified appearance standard part of the piston generating feature parameters corresponding to the reference parameters, and based on a comparison between the feature parameters and the reference parameters, so that the generated feature parameter value becomes the corresponding reference parameter value, The captured image is corrected.

[0012] An image correction program according to another aspect of the present disclosure is an image correction program for correcting a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine using the photographed image, the image correction program including: a computer; acquiring the photographed image; acquiring reference parameters related to a predetermined appearance reference portion of the piston that is predetermined for the engine corresponding to the photographed image; and correcting the reference parameters from the photographed image. The piston's predetermined A feature portion corresponding to the appearance reference portion is extracted, and the Related to the specified appearance standard part of the piston generating feature parameters corresponding to the reference parameters, and based on a comparison between the feature parameters and the reference parameters, so that the generated feature parameter value becomes the corresponding reference parameter value, The photographed image is corrected. [Effects of the Invention]

[0013] According to the present disclosure, in order to record or diagnose the condition of an engine using a photographed image of a piston of the engine being inspected, the photographed image can be corrected to an image suitable for diagnosis. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of an engine diagnosis system to which an image correction device according to an embodiment of the present disclosure is applied. [Figure 2] FIG. 2 is a diagram showing an example of an engine to be inspected. [Figure 3] FIG. 3 is a diagram showing a schematic image of a piston and a piston ring photographed through a scavenging port. [Figure 4] FIG. 4 is a schematic diagram showing the correspondence between the appearance reference parts and the characteristic parts of the captured image in this embodiment. [Figure 5] FIG. 5 is a flowchart showing the flow of the image correction process in this embodiment. [Figure 6] FIG. 6 is a conceptual diagram of the process of specifying the determination reference area in this embodiment. [Figure 7] FIG. 7 is a conceptual diagram showing the interpolation process in this embodiment. [Figure 8] FIG. 8 is a conceptual diagram showing the third conversion process in this embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that, in the following, the same or corresponding elements will be denoted by the same reference numerals throughout the drawings, and redundant explanations will be omitted.

[0016] [Overall configuration] 1 is a block diagram showing a schematic configuration of an engine diagnostic system to which an image correction device according to an embodiment of the present disclosure is applied. The engine diagnostic system 1 in this embodiment records or diagnoses the condition of pistons and piston rings of a two-stroke diesel engine installed on a ship. The engine diagnostic system 1 includes an image correction device 2, an image diagnostic device 3, a captured image receiver 4, and a diagnostic result output device 5. These components 2 to 5 transmit and receive data to and from each other via a signal bus 6.

[0017] The image correction device 2 includes a memory 21 and a computing unit 22. The image diagnostic device 3 also includes a memory 31 and a computing unit 32. The memory 21 of the image correction device 2 and the memory 31 of the image diagnostic device 3 may be configured as a common storage device such as a data server. Alternatively, the memories 21 and 31 may be configured as separate storage devices. The memories 21 and 31 may be configured as a cloud server. Furthermore, the computing unit 22 of the image correction device 2 and the computing unit 32 of the image diagnostic device 3 may be configured as a common computing device, or may be configured as separate computing devices.

[0018] When the storage device 21 is configured as a cloud server, the image correction device 2 may be configured as a communication terminal such as a personal computer that has a computing device 22 and can communicate with the cloud server that is the storage device 21. The image diagnostic device 3 may also have a similar configuration.

[0019] The storage 21 of the image correction device 2 stores an image correction program for performing image correction processing, photographed image data acquired through the photographed image receiver 4, photographed image data after image correction processing (corrected image data), engine model information, and reference parameters related to appearance reference parts associated with the engine model information. The storage 31 of the image diagnostic device 3 stores a diagnostic program for performing engine diagnostic processing, corrected image data, diagnostic result data, etc.

[0020] The captured image receiver 4 is a communication interface that receives captured images taken by the photographing device 8 via a predetermined communication network 7. The photographing device 8 is, for example, a communication terminal equipped with a camera such as a smartphone. Alternatively, the photographing device 8 may be a camera with communication capabilities. The photographer photographs the engine to be inspected that is installed on the ship, and transmits the captured image data from the photographing device 8 to the engine diagnostic system 1 via the communication network 7. In cases where the photographing device 8 does not have a communication capability, for example, the photographer may import the data of the captured images taken by the photographing device 8 into a communication terminal such as a personal computer or smartphone, and transmit the captured image data from the communication terminal to the engine diagnostic system 1.

[0021] The image correction device 2 performs image correction processing, which will be described later, on the captured image received by the captured image receiver 4, and corrects the image into an image suitable for image diagnosis performed by the image diagnosis device 3. The image diagnosis device 3 uses the corrected image corrected by the image correction device 2 to diagnose the engine.

[0022] The captured image includes the circumferential surface of the engine piston and the piston ring attached to the piston. The exposed portion of the piston circumferential surface, i.e., the portion where the piston ring is not attached, is called the ring land. The captured image shows the ring land and piston ring as a strip. When the engine is operating, the piston moves up and down within the cylinder, forming scratches extending vertically in the sliding direction on the piston ring sliding along the inner circumferential surface of the cylinder. The diagnostic imaging device 3 analyzes the condition of the scratches formed on the piston ring from the captured image and performs a diagnosis such as a pass / fail determination. The diagnostic imaging device 3 also analyzes the degree of contamination on the ring land surface and performs a diagnosis such as a pass / fail determination. The diagnostic imaging device 3 can comprehensively evaluate the cylinder condition, including sliding parts, combustion chamber components, and the oil injection system, based on the condition of the piston ring and the condition of the ring land.

[0023] The captured image may include the structure surrounding the piston. The image correction device 2 may identify the position of the piston from the structure surrounding the piston, or may determine whether or not a diagnosis of the piston is possible from the captured image by identifying the structure surrounding the piston.

[0024] The diagnostic result output device 5 transmits the diagnostic results from the image diagnostic device 3 to a predetermined result display device. The result display device may include, for example, a communication terminal which is the imaging device 8, a predetermined computer mounted on the ship 9 on which the engine to be inspected is mounted, or a management device 10 provided on land or the like. In addition to or instead of transmitting the diagnostic results to the result display device, the diagnostic results may be stored as a diagnostic history in the memory 31 together with information about the engine to be inspected. The result display device may be able to access the engine diagnostic system 1 each time by performing a predetermined operation and obtain diagnostic history information for a desired engine.

[0025] The computing unit 22 of the image correction device 2 and the computing unit 32 of the image diagnostic device 3 each include a computer such as a microcontroller, a personal computer, etc. For example, the computing units 22 and 32 each include a CPU, a main memory such as a RAM, a communication interface, etc.

[0026] It should be noted that the functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, special-purpose processors, integrated circuits, ASICs (Application Specific Integrated Circuits), conventional circuits, or combinations thereof, configured or programmed to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors and other circuitry. In this specification, a circuit, unit, or means is hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be hardware disclosed herein or other known hardware that is programmed or configured to perform the recited functions. Where the hardware is a processor, which is considered a type of circuit, the circuit, unit, or means is a combination of hardware and software, and the software is used to configure the hardware or processor.

[0027] [Configuration of engines to be inspected] Fig. 2 is a diagram showing an example of an engine to be inspected. The engine 100 shown in Fig. 2 is a large two-cycle engine that is the main engine for propelling a ship. A two-cycle engine is also called a two-stroke engine.

[0028] The engine 100 has multiple cylinders 110 extending vertically. Note that only one cylinder 110 is shown in FIG. 2. In particular, in a uniflow two-stroke engine, each cylinder 110 has scavenging ports 120 on the circumferential surface of its lower portion and an exhaust port 130 on its upper portion. The engine 100 includes a piston 140, a fuel injection valve 150, and an exhaust valve 160 for each cylinder 110. The piston 140 has a cylindrical shape with an outer diameter corresponding to the inner diameter of the cylindrical cylinder 110. The piston 140 slides within the cylinder 110 across the scavenging ports 120. In other words, the sliding direction of the piston 140 is the vertical direction of the engine 100. An annular piston ring 210 is assembled to the circumferential surface of the piston 140. A piston rod 290 is fixed to the lower end of the piston 140. The piston rod 290 is connected to the crankshaft 180 via the connecting rod 170 .

[0029] The fuel injection valve 150 is located at the upper end portion of the cylinder 110, and is supplied with fuel from a fuel supply device 190. The fuel supply device 190 can change the fuel injection pattern and fuel injection timing of the fuel injection valve 150. The exhaust valve 160 is a valve that opens and closes the exhaust port 130, and is driven by an exhaust valve drive device 200. The exhaust valve drive device 200 can change the opening and closing pattern and opening and closing timing of the exhaust valve 160.

[0030] The internal space of the cylinder 110 is connected to a scavenging pipe 220 via a scavenging port 120. The scavenging pipe 220 and the scavenging port 120 are connected via a scavenging chamber 230. The scavenging pipe 220 extends horizontally and is connected to the plurality of cylinders 110. Air is supplied to the engine 100 through the scavenging pipe 220 as scavenging air.

[0031] A supercharger 240 is connected to the scavenging pipe 220. The supercharger 240 is configured to adiabatically compress the air to be supplied to the engine 100. The supercharger 240 includes a compressor 260 having a plurality of compressor blades and rotating around a rotary shaft 250 to compress air introduced from the outside, and a turbine 270 connected to the rotary shaft 250 of the compressor 260 to generate rotational power for the compressor 260.

[0032] The air compressed by the compressor 260 is supplied as scavenging air to the engine 100 via the scavenging pipe 220 and the scavenging chamber 230. Exhaust gas discharged from the exhaust port 130 due to combustion in the engine 100 flows into the turbine 270 of the turbocharger 240. The turbine 270 has a plurality of turbine blades, and the turbine 270 is rotated around the rotary shaft 250 by the exhaust gas that has flowed in, thereby rotating the compressor 260.

[0033] In such a marine engine 100, the scavenging pipe 220 is large enough for a person to enter. A photographer, such as a crew member or mechanic of the marine vessel, enters the scavenging pipe 220 or the scavenging chamber 230 through a predetermined maintenance hatch when the engine is stopped, and photographs the piston 140 and the piston ring 210 in the cylinder 110 through the scavenging port 120. The engine 100 is configured so that the vertical position of the piston 140 when the engine 100 is stopped can be adjusted. When photographing the piston 140 and the piston ring 210, the piston 140 is adjusted to be located at an inspection position P1 where the piston 140 and the piston ring 210 can be seen through the scavenging port 120. For example, the inspection position P1 is set to the lowest position of the piston 140 in the cylinder 110 or a position a predetermined distance above the lowest position.

[0034] [How the engine appears in the photograph] FIG. 3 is a diagram schematically illustrating a captured image of a piston and a piston ring photographed through a scavenging port. As shown in FIG. 3, when a photographer photographs the piston as described above, the captured image G includes at least one scavenging port 120 and a portion of the piston 210 and the piston ring 210 that is visible through the scavenging port 120. In the example of FIG. 3, three scavenging ports 120 are captured in the captured image G. The scavenging port 120 has an elliptical shape that extends vertically, that is, in the sliding direction of the piston 140. The elliptical shape is composed of arcs at the upper and lower ends and a straight portion connecting the two arcs. In other words, the scavenging port 120 has an elliptical shape that has a straight portion in the longitudinal direction.

[0035] 3, the captured image G shows four piston rings 210 assembled to the circumferential surface of the piston 140 and five ring lands 280 that are exposed portions on the circumferential surface of the piston 140. The outer circumferential surfaces of the piston rings 210 slide in contact with the inner circumferential surface of the cylinder 110, so in the captured image G, the piston rings 210 have a higher brightness than the ring lands 280. On the other hand, the outer circumferential surfaces of the ring lands 280 do not contact the inner circumferential surface of the cylinder 110, so that combustion residue, fuel residue, and the like accumulate on the outer circumferential surface. Therefore, in the captured image G, the ring lands 280 have a lower brightness than the piston rings 210.

[0036] Therefore, the captured image G has a feature portion in the shape of horizontal stripes, in which a plurality of rectangular piston ring portions and a plurality of rectangular ring land portions are alternately arranged in the vertical direction. The diagnostic imaging device 3 analyzes this feature portion through image processing to diagnose the condition of the engine 100, more specifically, the condition of the piston ring 210 and the ring land 280.

[0037] Here, the distance between the photographer and the piston 140 being photographed may change for each photograph, so the size of the piston 140 in the photographed image G is not constant. For this reason, in the past, a ruler or the like was included in the photographed image G so that the size of the piston 140 or the like could be seen.

[0038] Furthermore, the scavenging ports 120 are oriented in a direction inclined at a predetermined angle in the circumferential direction with respect to the radial direction of the cylinder 110 so that the scavenging air can flow smoothly. For example, the radially outer end of the scavenging port 120 is located at a position that is further clockwise with respect to the central axis of the cylinder 110 than the radially inner end.

[0039] Therefore, when photographing the piston 140 and the piston ring 210 in the cylinder 110 through the scavenging ports 120, in order to prevent the shadow of the scavenging ports 120 from being cast on the piston ring 210 and the ring land 280, the photographer needs to photograph from an oblique angle relative to the radial direction of the cylinder 110 in accordance with the direction of the scavenging ports 120. As a result, the piston ring 210 and the ring land 280 photographed through the scavenging ports 120 in the photographed image G have a perspective difference in the horizontal direction. Furthermore, since the photographed objects are the cylindrical piston 140 and the annular piston ring 210, the photographed image G includes distortion of the curved surface.

[0040] Therefore, depending on the photographed image G, it may be difficult to detect sliding scratches, which are important in diagnosing the condition of the piston ring 210, and it may not be possible to properly diagnose the condition of the piston ring 210. Therefore, the image correction device 2 detects characteristic parts in the photographed image G and performs image processing to deform the shapes of the characteristic parts, thereby correcting the entire photographed image G.

[0041] The scavenging pipe 220 or the scavenging chamber 230 where the photographed image G is taken is a space through which scavenging air flows, and is therefore contaminated with combustion debris, fuel residue, and the like. Furthermore, the scavenging pipe 220 or the scavenging chamber 230 is originally not a place where people can enter, so it is narrow and dark without any lighting. Therefore, the photographer takes the photograph by turning on the light or strobe of the photographing device 8. For this reason, the photographing conditions for the photographed image G are generally not good. For example, the shadow of the scavenging port 120 falls on the piston ring 210 or the ring land 280, and overexposure due to lighting is likely to occur. Therefore, there is a limit to image correction using only the photographed image G.

[0042] Therefore, in this embodiment, characteristic portions of the piston 140 of the engine 100 to be inspected that are captured in the photographed image G are converted into data in advance as appearance reference portions. Feature parameters obtained from the characteristic portions corresponding to the appearance reference portions in the photographed image G are compared with reference parameters related to the appearance reference portions, and the photographed image G is corrected based on the comparison results. For this purpose, the memory 21 of the image correction device 2 stores reference parameters related to predetermined appearance reference portions of the piston 140 that are predetermined according to the engine 100 to be inspected.

[0043] FIG. 4 is a schematic diagram showing the correspondence between the appearance reference portion and the characteristic portion of the captured image in this embodiment. The diagram on the left side of FIG. 4 is a virtual ideal diagram created based on the positional relationship and dimensions of the appearance reference portion EB in the actual engine 100. The diagram on the right side of FIG. 4 is a schematic diagram showing an example of the appearance reference portion, i.e., the characteristic portion CH, on the captured image G when the appearance reference portion EB is captured. Note that the diagram on the right side of FIG. 4 does not take into account distortion due to the piston 140 and piston ring 210 having curved outer peripheries that appear in the captured image G. Furthermore, the diagram on the right side of FIG. 4 shows exaggerated distortion of the characteristic portion CH due to differences in perspective.

[0044] The appearance reference portion includes a pair of first boundary lines BL1 (up and down) between the piston ring 210 and the ring land 280 when the piston 140 is viewed through the scavenging port 120 of the cylinder 110, a pair of second boundary lines BL2 (left and right) between the piston ring 210 or the ring land 280 and the scavenging port 120, or an intersection point BP of the first boundary line BL1 and the second boundary line BL2.

[0045] For example, as shown in Fig. 4, when there are four piston rings 210, there are eight first boundary lines BL1. However, when the topmost piston ring 210 in Fig. 4 is the object of inspection or correction, a total of two first boundary lines BL1, namely, a boundary line between the piston ring 210 and the ring land 280 that contacts the upper side of the piston ring 210 and a boundary line between the piston ring 210 and the ring land 280 that contacts the lower side of the piston ring 210, serve as the basis for image correction. In addition, there are two second boundary lines BL2 for each scavenging port 120. The first boundary lines BL1 and the second boundary lines BL2 are perpendicular to each other.

[0046] Therefore, one portion of the piston ring 210 indicated by diagonal lines in FIG. 4 is defined as a rectangle having vertices at four intersections BP where two first boundary lines BL1 and two second boundary lines BL2 intersect. Hereinafter, this rectangle will be referred to as a reference rectangle BB. The thickness of the piston ring 210 and the opening width of the scavenging port 120 are known in advance depending on the engine 100. Therefore, one or more parameters including the length of the long side, the length of the short side, or the ratio between the lengths of the long side and the short side of this reference rectangle BB, the angle between the first boundary line BL1 and the second boundary line BL2 being a right angle, the fact that the first boundary line BL1 extends horizontally, etc., are stored in the memory 21 as reference parameters. The reference parameters are converted from actual values ​​of the corresponding portion of the engine 100 to values ​​on an image region to be compared with the photographed image G, and the converted values ​​are stored in the memory 21.

[0047] Since the outer peripheral surfaces of the piston 140 and the piston ring 210 are curved, the length of the long side of the reference rectangle BB set as the reference parameter is set as the length of the arc of the piston 140 or the piston ring 210 bounded by the scavenging ports 120. In other words, the length of the long side of the reference rectangle BB is set as the horizontal length when the curved surface portions of the piston 140 and the piston ring 210 visible through the scavenging ports 120 are developed on a plane.

[0048] The length of the long side of the reference rectangle BB can also be defined as a first distance W1 between two intersections where one first boundary line BL1 intersects with two second boundary lines BL2. The length of the short side of the reference rectangle BB can also be defined as a second distance W2 between two intersections where two vertically adjacent first boundary lines BL1 intersect with one second boundary line BL2. The reference rectangle BB may be the topmost piston ring 210 alone, or may be a rectangle corresponding to other portions of the piston ring 210 in addition to or instead of the topmost piston ring 210. A rectangle corresponding to a portion of the ring land 280 may also be defined in a similar manner, and a rectangle corresponding to one or more portions of the ring land 280 may be used as the reference rectangle BB. In other words, the image correction process in this embodiment can be performed not only for diagnosing the condition of the piston ring 210 but also for diagnosing the condition of the ring land 280. The appearance reference portion EB, which serves as a reference for image correction of the reference rectangle BB and the like, may be set or changed depending on the portion to be diagnosed by the image diagnostic device 3.

[0049] In the photographed image G of the vicinity of the piston 140 including the appearance reference portion EB, the reference rectangle BB does not appear as a rectangle due to the above-mentioned photographing constraints. For example, as shown in FIG. 4 , the straight line portion on the left side of the scavenging port 120 is located closer to the viewer than the straight line portion on the right side. Therefore, the feature region BC, which is the feature portion CH corresponding to the reference rectangle BB, is a rectangle in which the left side extending in the vertical direction is longer than the right side extending in the vertical direction. Furthermore, by photographing the scavenging port 120 from an oblique direction, the left-right width of the scavenging port 120 appears shorter in the photographed image G. As a result, the ratio of the lengths of the short sides and long sides of the rectangle in the feature region BC differs from the actual ratio in the appearance reference portion EB.

[0050] The calculator 22 of the image correction device 2 corrects the photographed image G so that the characteristic portion CH in the photographed image G has a shape and size defined as the appearance reference portion EB. For example, the calculator 22 corrects the photographed image G so that the characteristic region BC on the photographed image G becomes a reference rectangle BB having a long side of a first distance W1 and a short side of a second distance W2.

[0051] [Image correction processing] The image correction process will be described in more detail below. FIG. 5 is a flowchart showing the flow of the image correction process in this embodiment. The calculator 22 acquires the captured image G transmitted from the image capture device 8 via the communication network 7 (step S1). The data of the captured image G transmitted from the image capture device 8 includes data on the image capture device 8, such as the focal length, and engine identification data for identifying the engine 100 being captured. The data of the captured image G may also include data on the orientation of the captured image G, indicating whether the captured image G is vertical or horizontal, and situation data at the time of capture, such as the shooting distance. The engine identification data may include, for example, model information of the engine 100, a customized configuration of the engine 100, or model information of the ship on which the engine 100 is installed. The engine identification data may also include the individual identification number of the engine 100 or the individual identification number of the ship.

[0052] The calculator 22 acquires the reference parameters of the engine 100 corresponding to the acquired photographed image G (step S2). The calculator 22 identifies the engine 100 corresponding to the photographed image G from the engine identification data accompanying the photographed image G. The calculator 22 reads out the reference parameters of the identified engine 100 from the memory 21. In this embodiment, the reference parameters stored in the memory 21 are associated with engine identification data such as model information of the engine 100. Therefore, the reference parameters of the engine 100 corresponding to the photographed image G can be easily acquired based on the engine identification data.

[0053] The calculator 22 extracts a characteristic portion CH from the captured image G (step S3). The calculator 22 extracts, as the characteristic portion CH, a characteristic region BC from the captured image G, which is a quadrangle having vertices each consisting of a first characteristic line CL1 corresponding to the first boundary line BL1, a second characteristic line CL2 corresponding to the second boundary line BL2, an intersection CP between the first characteristic line CL1 and the second characteristic line CL2, or four intersections CP where two first characteristic lines CL1 and two second characteristic lines CL2 intersect.

[0054] The calculator 22 determines whether or not a characteristic portion CH is included in the photographed image G. To this end, the calculator 22 estimates a first characteristic line CL1 from the luminance distribution in the vertical direction of the photographed image G. The calculator 22 also estimates a second characteristic line CL2 from the luminance difference distribution in the horizontal direction of the photographed image G. In this embodiment, the following process for identifying the determination reference area BD is performed.

[0055] [Decision-making criteria area identification process] 6 is an image diagram of the process of specifying the judgment reference area in this embodiment. In FIG. 6, the first characteristic line CL1 in the captured image G is shown as a straight line extending substantially horizontally, and the second characteristic line CL2 is shown as a straight line extending vertically.

[0056] The brightness distribution in the vertical direction of the captured image G is obtained as a first distribution D1. The calculator 22 calculates the sum of the brightness of multiple pixels arranged in the horizontal direction for the entire captured image G. The first distribution D1 represents the vertical distribution of the sum of the horizontal brightness. As described above, the piston ring 210 slides against the inner circumferential surface of the cylinder 110, so the area corresponding to the piston ring 210 in the captured image G has higher brightness than the area corresponding to the ring land 280, which does not contact the inner circumferential surface of the cylinder 110. As a result, the first distribution D1 alternates between vertical regions with high brightness and vertical regions with low brightness. The calculator 22 identifies, as a first identified region E1, the position of a vertical region in the first distribution D1 in the captured image G where the brightness is equal to or greater than a first threshold value Th1.

[0057] The distribution of brightness differences in the horizontal direction of the captured image G is obtained as the second distribution D2. The calculator 22 calculates the sum of the brightness differences of multiple pixels lined up in the vertical direction for the entire captured image G. For example, the calculator 22 calculates the difference between the brightness of a first pixel in a certain vertical column and the brightness of a second pixel located one unit pixel below the first pixel. The calculator 22 further calculates the difference between the brightness of the second pixel and the brightness of a third pixel located one unit pixel below the second pixel. In the same manner, the calculator 22 calculates all brightness differences of adjacent pixels in a certain vertical column and adds them together. The second distribution D2 indicates the horizontal distribution of the sum of the brightness differences in the vertical direction.

[0058] As described above, in the appearance reference portion EB of the piston 140, the horizontally extending piston rings 210 and ring lands 280 alternate in the vertical direction. Therefore, the region in the photographed image G where the piston rings 210 and ring lands 280 exist includes a region with a large vertical brightness difference. As a result, the second distribution D2 includes horizontal regions with a large brightness difference and horizontal regions with a small brightness difference. The calculator 22 identifies, as the second identified region E2, the position of the horizontal region in the second distribution D2 in the photographed image G where the brightness difference is equal to or greater than the second threshold value Th2.

[0059] A determination reference region BD where the first specific region E1 and the second specific region E2 overlap is estimated to be a region in the photographed image G where the piston ring 210 is captured. In this way, by using the alternating pattern of brightness in the photographed image G to estimate the region where the piston ring 210 exists, including the characteristic portion CH, the position and number of the region where the piston ring 210 exists in the photographed image G can be easily identified.

[0060] When calculating the first distribution D1 and the second distribution D2, addition or calculation of the luminance difference may be performed using one pixel as a calculation unit, or the luminance of a unit pixel region including multiple pixels, for example, 3 pixels x 3 pixels, may be added. The luminance of a unit pixel region may be calculated as the average value or sum of the multiple pixels included in the unit pixel region.

[0061] The calculator 22 estimates the longitudinal boundary line of the first specific region E1, i.e., the horizontal boundary line, as the first characteristic line CL1. The calculator 22 also estimates the longitudinal boundary line of the second specific region E2, i.e., the vertical boundary line, as the second characteristic line CL2. However, as described above, in the captured image G, the actual first characteristic line CL1 may be inclined with respect to the horizontal direction and may be distorted. Similarly, in the captured image G, the actual second characteristic line CL2 may be inclined with respect to the vertical direction.

[0062] Therefore, hereinafter, the estimated first and second characteristic lines CL1 and CL2 are referred to as first and second estimated lines IL1 and IL2 to distinguish them from the actual first and second characteristic lines CL1 and CL2 in the captured image G. The calculator 22 performs a predetermined detection process in a predetermined determination area based on a determination reference area BD, which is a rectangle surrounded by the first and second estimated lines IL1 and IL2, to identify a characteristic portion CH including the first and second characteristic lines CL1 and CL2 or an intersection CP between the first and second characteristic lines CL1 and CL2. The determination area may be set to the same area as the determination reference area BD, or may be set to an area larger or smaller than the determination reference area BD. The data for the determination reference area BD includes, for example, X-coordinate data indicating the horizontal positions of the two second estimated lines IL2, data on the lengths of the short sides, which are the distances between the two first estimated lines IL1, and Y-coordinate data indicating the vertical positions of the midpoints of the short sides.

[0063] As described above, the determination reference region BD indicates the approximate position where the piston ring 210 is captured in the captured image G. For example, in FIG. 6, six determination reference regions BD are identified, and it is estimated that the piston ring 210 is captured in these six regions in the captured image G. A predetermined detection process may be performed on all of these six regions, or on one or more predetermined regions among them. The user may specify the locations where the detection process is performed, i.e., the locations where the diagnosis is performed. Furthermore, a predetermined detection process may be performed on the region between two vertically adjacent determination reference regions BD.

[0064] [First detection process] For example, as the first detection process, the calculator 22 may perform line detection processing using a Hough transform or the like on the estimated region. The calculator 22 may extract line components adjacent to the first estimated line IL1 from the line components detected in the line detection process and perform edge detection processing based on the extracted line components to identify the first characteristic line CL1. Similarly, the calculator 22 may extract line components adjacent to the second estimated line IL2 from the line components detected in the line detection process and perform edge detection processing based on the extracted line components to identify the second characteristic line CL2. When the first characteristic line CL1 and the second characteristic line CL2 have been identified, the calculator 22 identifies their intersection points CP and identifies a characteristic region BC, which is a quadrangle with vertices at the four intersection points CP where two first characteristic lines CL1 and two second characteristic lines CL2 intersect.

[0065] [Second detection process] If the first characteristic line CL1 or the second characteristic line CL2 in the captured image G has a large distortion, it may be impossible to extract a straight line component close to the first estimated line IL1 or the second estimated line IL2 in the line detection process. In this case, the calculator 22 may perform a corner detection process on the determination area as the second detection process. If the intersection point CP is identified by the corner detection process, the calculator 22 performs a third conversion process, which will be described later.

[0066] [Interpolation] Depending on the photographed image G, the number of determination reference regions BD identified may not be the same as the number of piston rings 210 that are actually present, for example, because some or all of the piston rings 210 are in the shadow of the scavenging ports 120. In such a case, it is not clear whether the determination reference region BD located at the top in the photographed image G corresponds to the piston ring 210 located at the top among the multiple piston rings 210 assembled to the piston 140.

[0067] In anticipation of such a case, the following processing may be performed. In this processing, piston configuration data including information such as the number, position, width, or inter-ring spacing of piston rings 210 assembled to piston 140 of engine 100 to be inspected is stored in memory 21. After specifying the judgment criterion region BD, calculator 22 reads the corresponding piston configuration data of engine 100 and determines whether the specified number, position, or spacing between judgment criterion regions BD corresponds to the piston configuration data.

[0068] If the calculator 22 determines that the judgment reference region BD is not identified in correspondence with the piston configuration data, the calculator 22 may perform an interpolation process in the detection process of the characteristic part CH to estimate the characteristic part CH that could not be detected from some of the detected characteristic parts CH.

[0069] Fig. 7 is an image diagram showing the interpolation process in this embodiment. In the example of Fig. 7, although 12 points should be identified as intersection points CP of the characteristic portion CH from the piston configuration data, three points cannot be identified because the upper part of the piston 140 in the photographed image G is shaded by the scavenging port 120.

[0070] The calculator 22 compares the information on the first characteristic line CL1, the second characteristic line CL2, the intersection points CP, and the characteristic region BC identified in the photographed image G with the piston configuration data, thereby identifying missing portions in the photographed image G. In the example of Fig. 7, of the four intersection points CP that constitute the four vertices of the characteristic region BC corresponding to the top piston ring 210, three intersection points CPc are identified as missing.

[0071] The missing intersection CPc is estimated to be located on an extension of the identified first characteristic line CL1. The calculator 22 estimates the position of the missing intersection CPc from the positional relationship of the multiple intersections CL identified on the first characteristic line CL1 and the piston configuration data. That is, the position of the missing intersection in the photographed image G is estimated based on the correlation between the distances between the multiple intersections CL in the photographed image G and the distances in the corresponding piston parts or piston ring parts.

[0072] 7, the positions of the intersection points CP on the photographed image G corresponding to the second and third piston rings 210 from the top are identified, and the positional relationship of the three piston rings 210 is acquired from the piston configuration data. Therefore, the calculator 22 can extrapolate and determine the position of the missing intersection point CPc based on the position of the intersection point CP on the photographed image G. Similarly, missing portions CL1c, CL2c on the first characteristic line CL1 or the second characteristic line CL2 can also be estimated.

[0073] [Auxiliary Processing] In addition to or instead of performing such interpolation processing, the following auxiliary processing may be performed in the processing for specifying the determination reference region BD using the first distribution D1 and the second distribution D2.

[0074] For example, the calculator 22 may perform differential image processing to obtain a vertical differential image of the luminance components for the entire captured image G, and estimate the area where the piston ring 210 and the ring land 280 exist from the alternating pattern of luminance differences that appear alternately in the vertical direction in the obtained differential image. For example, the calculator 22 may subtract the luminance of a first pixel in a certain vertical column from the luminance of a second pixel located one unit pixel below the first pixel, and determine the luminance after the subtraction as the luminance of the first pixel. Similarly, the calculator 22 may subtract the luminance of the second pixel from the luminance of a third pixel located one unit pixel below the second pixel, and determine the luminance after the subtraction as the luminance of the third pixel. The calculator 22 performs such calculations for the entire captured image G to generate a differential image expressed in the luminance after the subtraction.

[0075] As a result, the difference image is an image in which the boundary between the piston ring 210 and the ring land 280, i.e., the first characteristic line CL1, is more emphasized. In other words, the difference image is more clearly defined than the photographed image G, in which a high-brightness area corresponding to the piston ring 210 and a low-brightness area corresponding to the ring land 280 are repeated in the vertical direction. Furthermore, as shown in FIG. 3 , when the photographed image G includes two or more scavenging ports 120, two or more alternating pattern groups appear in the horizontal direction in the photographed image G and the difference image. In this case, the calculator 22 may set a predetermined alternating pattern group from among the two or more alternating pattern groups as the region for which the determination process of the determination criterion region BD is performed. For example, when three alternating pattern groups are included, the calculator 22 sets the central alternating pattern group as the region for which the determination process is performed. Alternatively, the calculator 22 may set the alternating pattern group with the longest horizontal length from among the two or more alternating pattern groups as the region for which the determination process of the determination criterion region BD is performed.

[0076] The calculator 22 may perform a process of specifying the determination reference region BD based on the difference image obtained in this way.

[0077] Furthermore, when the arc portion of the scavenging port 120 is included in the photographed image G, the calculator 22 may identify the judgment reference region BD by using the shape of the arc portion of the scavenging port 120. If light is shining on the inner peripheral surface of the scavenging port 120 during photography, the piston ring 210 and the ring land 280 may appear on the inner peripheral surface of the scavenging port 120 in the photographed image G, and the second characteristic line CL2, which is the boundary between the scavenging port 120 and the piston ring 210 or the ring land 280, may not be directly identified. Furthermore, if the scavenging port 120 is contaminated or if the scavenging port 120 is in shadow, the second characteristic line CL2 may not be directly identified. In such cases, the following processing may be performed.

[0078] As shown in Fig. 3, the arc portion of the scavenging port 120 has a shape such that the diameter is larger on the radially outer side of the cylinder 110 than on the radially inner side. The photographed image G includes a first arc portion 121 at the radially outer end and a second arc portion 122 at the radially inner end. The calculator 22 extracts the first arc portion 121 by ellipse detection processing. The memory 21 stores structural data including the diameter of the arc portion on the radially outer side of the scavenging port 120, the diameter of the arc portion on the radially inner side, the depth of the scavenging port 120, or the inclination of the scavenging port 120 with respect to the radial direction.

[0079] The computing unit 22 determines the relative position of the scavenging port 120 and the imaging device 8 based on the structural data of the scavenging port 120 from the inclination or circularity of the ellipse in the first arc portion 121, etc. The computing unit 22 estimates the position and attitude of the second arc portion 122 from the structural data and the relative position of the scavenging port 120, and estimates the boundary between the scavenging port 120 and the piston ring 210 or the ring land 280 in the determination reference region BD, i.e., the position of the second characteristic line CL2, from the two upper and lower second arc portions 122 included in one scavenging port 120.

[0080] For this purpose, the calculator 22 may, for example, find a common tangent to the two upper and lower second arc portions 122 and estimate the common tangent as the second characteristic line CL2. Alternatively, the calculator 22 may, for example, perform edge detection on the two upper and lower second arc portions 122 and estimate the second characteristic line CL2 by connecting the edges detected in the two second arc portions 122. In this embodiment, the scavenging port 120 has a straight line portion connecting the two second arc portions 122. However, the scavenging port 120 may have a special shape without a straight line portion, such as a shape formed by a plurality of circular holes formed by machining, such as grinding or cutting. The above estimation method is effective for estimating the second characteristic line CL2 for a scavenging port 120 having such a special shape.

[0081] The calculator 22 generates feature parameters corresponding to the reference parameters from the feature portion CH extracted as described above (step S4). For example, if the reference parameters include the lengths of the long and short sides of the reference rectangle BB and the ratio between the lengths of the long and short sides, the calculator 22 calculates the lengths of the long and short sides and the ratio between the lengths of the long and short sides of a rectangle in a feature region BC, which is the feature portion CH in the captured image G. Furthermore, the calculator 22 may include, as feature parameters, the inclination of the long sides of the rectangle in the feature region BC relative to the horizontal direction, the inclination of the short sides of the rectangle in the feature region BC relative to the vertical direction, the positions of the vertices of the rectangle in the feature region BC, and the like. The calculator 22 may also generate feature parameters for the captured image G after a third conversion process, which will be described later, has been performed.

[0082] The computing unit 22 corrects the photographed image G based on a comparison between the generated feature parameters and the reference parameters (step S5). For example, the photographed image G is corrected so that the shape of the feature portion CH on the photographed image G shown on the right side of Fig. 4 matches the appearance reference portion EB shown on the left side of Fig. 4, i.e., so that the feature parameters match the reference parameters.

[0083] [Example of correction processing] The calculator 22 executes at least one of the following processes as a correction process for the captured image G. For example, the calculator 22 may correct the captured image G by performing a first conversion process that converts a quadrangular feature region BC in the captured image G into a rectangle. The first conversion process includes, for example, a homography conversion. The first conversion process corrects the horizontal perspective difference in the feature portion CH of the captured image G.

[0084] The calculator 22 may also correct the photographed image G by performing a second conversion process to match the ratio of the lengths of the two intersecting sides in the rectangular feature area BC on the photographed image G, i.e., the ratio of the lengths of the long and short sides of the rectangle, with the ratio in the reference parameters. This allows the feature area CH in the photographed image G to be converted to an aspect ratio of an appropriate actual size. The second conversion process may be performed after the first conversion process, or may be performed without the first conversion process if, for example, the first conversion process does not need to be performed on the photographed image G.

[0085] The calculator 22 may also perform a correction to match the magnification of the characteristic region BC in the photographed image G with the reference parameters. That is, the calculator 22 may perform a correction to match the lengths of the long and short sides of the characteristic region BC, which is a rectangle of the characteristic part CH in the photographed image G, with the lengths in the reference parameters.

[0086] Furthermore, the calculator 22 may correct the captured image G so that the first feature line CL1 in the captured image G becomes a straight line. For example, the calculator 22 may divide the feature region BC in the captured image G into a plurality of quadrangular parts, and perform a first conversion process on the quadrangular parts to perform a third conversion process that converts the entire feature region BC into a rectangle.

[0087] Fig. 8 is an image diagram showing the third conversion process in this embodiment. In the example shown in Fig. 8, the characteristic region BC in the captured image G is distorted due to the curved outer circumferential surface of the cylinder 110 or the piston ring 210. For this reason, the third conversion process is performed in cases such as when the first characteristic line CL1 cannot be detected in the detection process for the characteristic portion CH and the intersection point CP is detected by the corner detection process.

[0088] In the third conversion process, the calculator 22 acquires brightness difference information at one of the two upper intersection points CP among the four identified intersection points CP. In the example of FIG. 8, the calculator 22 acquires brightness difference information at the upper left intersection point CP. The brightness difference information is, for example, information indicating the difference between the brightness at a position a predetermined distance above the intersection point CP in the captured image G and the brightness at a position a predetermined distance below the intersection point CP. Since the intersection point CP is located on the boundary between the piston ring 210 and the ring land 280, it is estimated that the brightness difference in the vertical direction at the intersection point CP is somewhat large. The brightness difference information at the intersection point CP is stored in the memory 21 as a criterion for boundary search, which will be described later.

[0089] The calculator 22 moves the search point a predetermined distance horizontally from the intersection point CP where the brightness difference information was acquired, and acquires the brightness difference information at that position. In the example of FIG. 8, the calculator 22 sets the search point at a position that is a predetermined distance to the right, in a direction approaching the upper right intersection point from the upper left intersection point CP. The calculator 22 determines whether the acquired brightness difference is equal to or greater than a reference value determined based on the brightness difference at the intersection point CP. If the brightness difference at the search point is less than the reference value, the calculator 22 sets a position above or below the search point as a search point, and similarly acquires the brightness difference.

[0090] 8 is an area corresponding to the piston ring 210, i.e., an area with higher brightness than the surrounding area. Therefore, if the brightness at the search point is high, the calculator 22 sets a position above the search point as a new search point. On the other hand, if the brightness at the search point is low, the calculator 22 sets a position below the search point as a new search point.

[0091] The calculator 22 determines whether the brightness difference at the new search point is equal to or greater than a reference value, and if so, determines that the new search point is the boundary point DP between the piston ring 210 and the ring land 280. In this way, the calculator 22 identifies the boundary point DP by moving the search point upward or downward to a position where the brightness difference is equal to or greater than the reference value. Furthermore, the calculator 22 performs a boundary search in which the identification of the boundary point DP is repeated until the boundary point DP reaches the horizontal position of the upper right intersection point CP. Similarly, the calculator 22 performs a similar boundary search for the intersection point CP below the characteristic region BC. Note that the horizontal positions of the search points on the first line segment LS1 and the second line segment LS2 are set to be equal to each other.

[0092] The calculator 22 divides the characteristic region BC into multiple rectangular portions BE by connecting the obtained boundary points DP with straight lines. The calculator 22 divides the captured image G at the horizontal positions of the dividing lines of the multiple rectangular portions BE extending vertically. The calculator 22 performs a first conversion process on the divided captured image portions to convert the rectangular portions BE included in the captured image portions into rectangles. The calculator 22 combines the multiple captured image portions after the first conversion process to generate a corrected image having a characteristic region BCc corrected to a rectangle. This corrects the distortion and enables the captured image G to be corrected into an image suitable for image diagnosis.

[0093] In the third conversion process described above, a search is performed starting from one of the two intersection points CP. However, the calculator 22 may set a search point on the perpendicular bisector of the line segment connecting the two upper or lower intersection points CP and determine the brightness difference. In this case, after identifying the boundary point DP on the perpendicular bisector, the calculator 22 sets a search point on the perpendicular bisector of the line segment connecting the identified boundary point DP and one of the intersection points CP and determines the brightness difference. By sequentially repeating this process, multiple boundary points DP are identified. Furthermore, the calculator 22 similarly sets a search point on the perpendicular bisector of the line segment connecting the identified boundary point DP on the perpendicular bisector of the line segment connecting the two intersection points CP to the other intersection point CP and determines the brightness difference.

[0094] The captured image G after the third conversion process may be further subjected to a second conversion process for correcting the aspect ratio.

[0095] Alternatively, the calculator 22 may adjust the tilt of the entire captured image G so that the first characteristic line CL1 is aligned horizontally. The calculator 22 detects the angle of the first characteristic line CL1 in the captured image G relative to the horizontal line, and performs an affine transformation to rotate the entire captured image G by the detected angle so that the first characteristic line CL1 coincides with the horizontal line.

[0096] If the first characteristic line CL1 is not a straight line, the calculator 22 may first perform the third conversion process and then perform the tilt adjustment, or may perform the tilt adjustment so that the center of the first characteristic line CL1 is horizontal. This allows the captured image G to be corrected to an image suitable for image diagnosis regardless of the photographer's shooting posture.

[0097] Additionally or alternatively, the calculator 22 may adjust the tilt of the entire captured image G so that the second characteristic line CL2 is aligned vertically. The calculator 22 detects the angle of the second characteristic line CL1 in the captured image G relative to the vertical line, and performs an affine transformation to rotate the entire captured image G by the detected angle so that the second characteristic line CL2 coincides with the vertical line.

[0098] By adjusting the tilt based on the second characteristic line CL2, the captured image G can be corrected to an image suitable for image diagnosis, regardless of the photographer's shooting posture. Furthermore, depending on the angle of view and shooting position of the imaging device 8, the second characteristic line CL2 may have higher linearity than the first characteristic line CL1 identified by the cylindrical piston 140 or the annular piston ring 210. In such cases, adjusting the tilt based on the second characteristic line CL2 allows for more accurate adjustment. The calculator 22 may perform both the tilt adjustment based on the first characteristic line CL1 and the tilt adjustment based on the second characteristic line CL2.

[0099] [effect] According to the above embodiment, the characteristic portion CH extracted from the photographed image G obtained by photographing the piston 140 of the engine 100 to be inspected is compared with the appearance reference portion EB of the piston 140 stored in advance in the memory 21, and the photographed image G is corrected so that the characteristic portion CH in the photographed image G appropriately corresponds to the appearance reference portion EB. Therefore, even if the photographed image G varies depending on the photographer, the photographing conditions, etc., the photographed image G can be corrected to an image suitable for image diagnosis of the engine 100.

[0100] As described above, according to the above embodiment, the captured image G can be made uniform to correspond to the dimensions or shape of the actual piston 140. This reduces oversights and false detections in image diagnosis of the engine 100, enabling highly accurate diagnosis. Furthermore, it also becomes easier to compare uniformed images. For example, it is easy to compare multiple engines 100 of the same model, or to determine the degree of aging deterioration of a single engine 100. Furthermore, because the dimensions of each part in the corrected captured image G are also made uniform according to the engine 100, the effort of incorporating a ruler into the captured image G is also eliminated.

[0101] Furthermore, according to the above embodiment, the appearance reference part EB is set based on the first boundary line BL1 between the piston ring 210 and the ring land 280 when the piston 140 in the cylinder 110 is viewed through the scavenging port 120, or based on the second boundary line BL2 between the piston ring 210 or the ring land 280 and the scavenging port 120. In this way, by setting a characteristic part of the appearance structure of the piston 140 as the appearance reference part EB, it becomes easier to extract the characteristic part CH corresponding to the appearance reference part EB in the photographed image G, and it is possible to reduce the frequency of erroneous detection or detection failure.

[0102] [Other embodiments] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments, and various improvements, changes, and modifications are possible.

[0103] For example, in the above embodiment, the first conversion process including homography conversion, the second conversion process for correcting the aspect ratio, the third conversion process for correcting a curve to a straight line, magnification correction, and tilt adjustment are exemplified as modes of correction processing for the captured image G. The calculator 22 of the image correction device 2 may be capable of executing all of these correction processes, or may be capable of executing only some of these correction processes. Furthermore, the calculator 22 of the image correction device 2 may be capable of executing correction processes other than those described above.

[0104] If the calculator 22 of the image correction device 2 is capable of executing multiple correction processes, the user may select at least one correction process to be executed from the multiple correction processes. Alternatively, the calculator 22 may determine whether or not each of the multiple correction processes needs to be executed, and execute the correction process determined to be necessary.

[0105] For example, the correction processes may be performed in the following order: First, the calculator 22 determines whether the first characteristic line CL1 in the captured image G is a straight line or whether the first characteristic line CL1 has been identified. If it is determined that the first characteristic line CL1 is not a straight line or that the first characteristic line CL1 cannot be identified, the calculator 22 performs the third conversion process.

[0106] Next, the calculator 22 determines whether the first characteristic line CL1, which is a straight line, is inclined with respect to the horizontal direction. If the first characteristic line CL1 is inclined with respect to the horizontal direction, the calculator 22 performs an inclination adjustment for the entire captured image G.

[0107] Next, the calculator 22 determines whether the quadrilateral feature region BC bounded by the two first feature lines CL and the two second feature lines CL2 is a rectangle. If the feature region BC is a quadrilateral that is not a rectangle, the calculator 22 executes a first conversion process to convert the feature region BC into a rectangle.

[0108] Next, the calculator 22 determines whether the aspect ratio of the rectangular feature region BC matches the aspect ratio of the reference rectangle BB. If the aspect ratio of the feature region BC does not match the aspect ratio of the reference rectangle BB, the calculator 22 executes a second conversion process to convert the aspect ratio of the feature region BC.

[0109] Next, the calculator 22 determines whether the length of the long or short side of the rectangular feature region BC matches the length of the long or short side of the reference rectangle BB. If the length of the long or short side of the feature region BC does not match the length of the long or short side of the reference rectangle BB, the calculator 22 performs magnification correction to change the magnification of the feature region BC.

[0110] In the above example, a determination is made as to whether or not each correction process should be performed. However, the calculator 22 may execute each correction process in the above order without making each of the above determinations.

[0111] Furthermore, when tilt adjustment is performed as the correction process, the calculator 22 may extract the first characteristic line CL1 by detecting lines throughout the entire captured image G, instead of performing the process of specifying the determination reference area BD. In this case, by utilizing the fact that a plurality of first boundary lines CL1 exist due to the formation of a plurality of cylinder rings 210 and ring lands 280, a plurality of lines that are parallel to each other or have a high degree of parallelism among the detected lines may be extracted as the first characteristic line CL1.

[0112] In the above embodiment, the appearance reference portion EB and its corresponding characteristic portion CH include points BP and CP, lines BL1, BL2, CL1, and CL2, and shapes BB and CP. However, the appearance reference portion EB and characteristic portion CH required for the correction process may be set appropriately. That is, the appearance reference portion EB and its corresponding characteristic portion CH may be any one of the above.

[0113] In the above embodiment, the process of identifying the determination reference area BD has been described in which the calculator 22 estimates the first characteristic line CL1 from the luminance distribution in the vertical direction of the captured image G and estimates the second characteristic line CL2 from the luminance difference distribution in the horizontal direction of the captured image G. However, this is not limiting. For example, the calculator 22 may estimate the first characteristic line CL1 from the luminance difference distribution in the vertical direction of the captured image G and estimate the second characteristic line CL2 from the luminance distribution in the horizontal direction of the captured image G. Regardless of whether the first characteristic line CL1 or the second characteristic line CL2 is estimated, the process of identifying the determination reference area BD may be performed using either the luminance distribution or the luminance difference distribution.

[0114] Furthermore, the process of identifying the reference area BD may be performed using a distribution other than the luminance distribution or luminance difference distribution. For example, the elements of the captured image G, i.e., the numerical values ​​to be compared, may include, in addition to the luminance value or luminance difference, the amount of error of the luminance of the pixel or unit area from the average luminance of the surroundings, the ratio of the luminance of the pixel or unit area to a predetermined reference luminance, or the degree of hue identity when the captured image G is converted to another hue system such as the HSV (Hue, Saturation, Value) color space.

[0115] As described above, in the process of identifying the determination reference area BD, the calculator 22 can estimate the first characteristic line CL1 from the vertical distribution of the first element of the captured image G, and estimate the second characteristic line CL2 from the horizontal distribution of the second element of the captured image G. The first element and the second element may be the same element or different elements.

[0116] Furthermore, in the above embodiment, an example has been given in which the captured image G captured by the photographing device 8 is transmitted to the image correction device 2 via the communication network 7, but an image correction program may be installed in a communication terminal equipped with the photographing device 8, and the image correction process for the captured image G may be performed in the communication terminal. In other words, the communication terminal equipped with the photographing device 8 may constitute the image correction device 2.

[0117] In this case, the storage device 21 storing the reference parameters and the like may be configured as a cloud server capable of communicating with the image capturing device 8. For example, the communication terminal may transmit engine-specific data of the engine 100 included in the image G captured by the image capturing device 8, and obtain data on the corresponding reference parameters of the engine 100 from the cloud server. The communication terminal may execute the image correction process using the image G captured by the image capturing device 8 and the obtained data on the reference parameters. Similarly, image diagnosis may be performed in a communication terminal equipped with the image capturing device 8.

[0118] Summary of this disclosure An image correction device according to one aspect of the present disclosure is an image correction device that corrects a photographed image of a piston of an engine under inspection or a structure surrounding the piston in order to diagnose the engine using the photographed image, and includes: a memory that stores reference parameters related to a predetermined appearance reference portion of the piston that is predetermined depending on the engine; and a calculator that performs image correction calculations for the photographed image, wherein the calculator acquires the photographed image, acquires the reference parameters of the engine that correspond to the photographed image, extracts feature portions that correspond to the appearance reference portion from the photographed image, generates feature parameters from the feature portions that correspond to the reference parameters, and corrects the photographed image based on a comparison between the feature parameters and the reference parameters.

[0119] According to the above configuration, characteristic features extracted from a photographed image of a piston of an engine under inspection are compared with the appearance reference features of the piston stored in advance in a memory, and the photographed image is corrected so that the characteristic features in the photographed image appropriately correspond to the appearance reference features. Therefore, even if the photographed image varies depending on the photographer, the photographing conditions, etc., the photographed image can be corrected to an image suitable for image diagnosis of the engine.

[0120] The reference parameters may be associated with engine identification data for identifying the engine. This allows the reference parameters serving as the basis for image correction to be acquired from engine model information. Therefore, the captured image and the engine model information contained in the captured image are all that is required as input information to correct the captured image into an image suitable for engine image diagnosis.

[0121] The engine may include a cylinder having an oval scavenging port having a straight portion in a longitudinal direction, and a piston having a piston ring assembled on a circumferential surface thereof and sliding inside the cylinder, wherein the appearance reference portion includes a first boundary line extending horizontally between the piston ring and a ring land which is an exposed portion on a circumferential surface of the piston when the piston is viewed through the scavenging port of the cylinder, a second boundary line extending vertically between the piston ring or the ring land and the scavenging port, or an intersection of the first boundary line and the second boundary line, and the calculation unit may extract, as the characteristic portion, a first characteristic line corresponding to the first boundary line, a second characteristic line corresponding to the second boundary line, or an intersection of the first characteristic line and the second characteristic line from the photographed image.

[0122] According to the above configuration, the appearance reference portion is set based on a first boundary line between the piston ring and the ring land when the piston in the cylinder is viewed through the scavenging port, or based on a second boundary line between the piston ring or the ring land and the scavenging port. By setting a characteristic portion of the piston's appearance structure as the appearance reference portion in this way, it is possible to easily extract a characteristic portion corresponding to the appearance reference portion in the captured image and reduce the frequency of erroneous detection or detection failure.

[0123] The computing unit may estimate the first characteristic line from a vertical distribution of a first element of the photographed image, and estimate the second characteristic line from a horizontal distribution of a second element of the photographed image. By estimating the presence areas of characteristic portions by utilizing the fact that the piston rings and the ring lands are alternately arranged in the vertical direction and the difference in image elements between the piston rings and the ring lands, the positions and number of the presence areas of characteristic portions in the photographed image can be easily identified.

[0124] The computing unit may correct the captured image by performing a first conversion process that converts a quadrangle defined by two of the first characteristic lines and two of the second characteristic lines into a rectangle, thereby correcting a perspective difference contained in the captured image and making it suitable for image diagnosis.

[0125] The computing unit may correct the captured image by performing a second conversion process that converts the ratio of the lengths of two intersecting sides of a rectangle defined by the two first characteristic lines and the two second characteristic lines, thereby converting the aspect ratio of the feature in the captured image to an appropriate actual size.

[0126] The computing unit may correct the captured image so that the first characteristic line becomes a straight line, thereby correcting distortion in the captured image and making it suitable for image diagnosis.

[0127] The computing unit may adjust the tilt of the entire captured image so that the first feature line is aligned horizontally or the second feature line is aligned vertically, thereby correcting the captured image to an image suitable for image diagnosis regardless of the photographer's shooting posture.

[0128] The storage device may be configured as a cloud server. This allows reference parameters corresponding to a large number of engines to be stored in the cloud server, and allows multiple terminals to access the cloud server and acquire the reference parameters corresponding to the engine being inspected. This eliminates the need for terminals performing image correction and image diagnosis to individually store data on the reference parameters. This eliminates the need to change the reference parameters by simply updating the data on the cloud server, eliminating the need for the effort required for the change process.

[0129] An image correction method according to another aspect of the present disclosure is an image correction method that corrects a photographed image of a piston of an engine under inspection or a surrounding structure of the piston in order to diagnose the engine using the photographed image, the image correction method comprising the steps of: acquiring the photographed image; acquiring reference parameters related to a predetermined appearance reference portion of the piston that is predetermined for the engine corresponding to the photographed image; extracting a feature portion that corresponds to the appearance reference portion from the photographed image; generating a feature parameter equivalent to the reference parameter from the feature portion; and correcting the photographed image based on a comparison between the feature parameter and the reference parameter.

[0130] An image correction program according to another aspect of the present disclosure is an image correction program that corrects a photographed image of a piston or a surrounding structure of an engine to be inspected in order to diagnose the engine using the photographed image, and causes the computer to acquire the photographed image, acquire reference parameters related to a predetermined appearance reference portion of the piston that is predetermined for the engine corresponding to the photographed image, extract a feature portion that corresponds to the appearance reference portion from the photographed image, generate a feature parameter equivalent to the reference parameter from the feature portion, and correct the photographed image based on a comparison between the feature parameter and the reference parameter. [Explanation of symbols]

[0131] 2. Image correction device 21 Memory 22 Arithmetic unit 100 Engine 120 Scavenging port 140 piston 210 Piston ring 280 Ringland BL1 First boundary line BL2 2nd boundary line BP Intersection of the first and second boundary lines CH Features CL1 First characteristic line CL2 Second characteristic line CP Intersection of the first and second feature lines EB Appearance Standards G Captured image

Claims

1. 1. An image correction device that corrects a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine using the photographed image, a memory that stores reference parameters related to a predetermined appearance reference portion of the piston that is predetermined according to the engine; a computing unit that performs an image correction calculation for the captured image, The computing unit Acquire the captured image; acquiring reference parameters related to a predetermined appearance reference portion of the piston of the engine corresponding to the photographed image; extracting a feature portion corresponding to a predetermined appearance reference portion of the piston from the photographed image, and generating, from the feature portion, feature parameters corresponding to reference parameters related to the predetermined appearance reference portion of the piston; an image correction device that corrects the captured image based on a comparison between the feature parameter and the reference parameter so that the value of the generated feature parameter becomes the value of the corresponding reference parameter.

2. 2. The image correction device according to claim 1, wherein the reference parameters are associated with engine identification data for identifying the engine.

3. The engine includes a cylinder having an oval scavenging port having a straight portion in a longitudinal direction, and a piston having a piston ring assembled to a circumferential surface thereof and sliding within the cylinder, the appearance reference portion includes a first boundary line extending in a horizontal direction between the piston ring and a ring land that is an exposed portion on a peripheral surface of the piston when the piston is viewed through the scavenging port of the cylinder, a second boundary line extending in a vertical direction between the piston ring or the ring land and the scavenging port, or an intersection of the first boundary line and the second boundary line, 3. The image correction device according to claim 1, wherein the computing unit extracts, from the captured image, a first feature line corresponding to the first boundary line, a second feature line corresponding to the second boundary line, or an intersection of the first feature line and the second feature line as the feature portion.

4. 4. The image correction device according to claim 3, wherein the computing unit estimates the first characteristic line from a luminance distribution in a vertical direction of the captured image, and estimates the second characteristic line from a luminance difference distribution in a horizontal direction of the captured image.

5. 5. The image correction device according to claim 3, wherein the computing unit corrects the captured image by performing a first conversion process that converts a quadrangle defined by two of the first characteristic lines and two of the second characteristic lines into a rectangle.

6. 6. The image correction device according to claim 3, wherein the computing unit corrects the captured image by performing a second conversion process such that a ratio of lengths of two intersecting sides in a rectangle defined by two of the first characteristic lines and two of the second characteristic lines matches a ratio in the corresponding reference parameter.

7. 7. The image correction device according to claim 3, wherein the computing unit corrects the captured image so that the first characteristic line becomes a straight line.

8. 8. The image correction device according to claim 3, wherein the calculator adjusts the tilt of the entire captured image so that the first feature line is aligned horizontally or so that the second feature line is aligned vertically.

9. The image correction device according to claim 1 , wherein the storage device is configured as a cloud server.

10. 1. An image correction method for correcting a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine, the method comprising: Acquire the captured image; acquiring reference parameters related to predetermined appearance reference portions of the piston that are predetermined for the engine corresponding to the photographed image; extracting a feature portion corresponding to a predetermined appearance reference portion of the piston from the photographed image, and generating, from the feature portion, feature parameters corresponding to reference parameters related to the predetermined appearance reference portion of the piston; an image correction method for correcting the captured image based on a comparison between the feature parameter and the reference parameter, so that the value of the generated feature parameter becomes the value of the corresponding reference parameter;

11. An image correction program for correcting a photographed image of a piston of an engine to be inspected or a peripheral structure of the piston in order to diagnose the engine using the photographed image, Computer, Acquire the captured image; acquiring reference parameters related to predetermined appearance reference portions of the piston that are predetermined for the engine corresponding to the photographed image; extracting a feature portion corresponding to a predetermined appearance reference portion of the piston from the photographed image, and generating, from the feature portion, feature parameters corresponding to reference parameters related to the predetermined appearance reference portion of the piston; an image correction program that causes the captured image to be corrected based on a comparison between the feature parameter and the reference parameter so that the value of the generated feature parameter becomes the value of the corresponding reference parameter.

Images