JP5672097B2 - Filtration membrane plate inspection device by image processing - Google Patents

Description

  The present invention relates to an apparatus for inspecting a filtration membrane plate for filtering water by image processing. Specifically, even in a situation where camera shake has occurred due to a disturbance factor, a crack or a film formation failure on the filtration membrane plate is correctly detected. The present invention relates to a filtration membrane plate inspection apparatus inspected by image processing.

  In water purification plants and the like, a process of removing impurities in the water is performed by filtering water using a filtration membrane plate. If the filtration membrane plate is damaged such as a crack, there is a possibility that impurities cannot be sufficiently removed. Therefore, it is necessary to test whether the filter membrane plate is damaged.

  For this reason, various devices for judging (inspecting) whether or not there is any damage in the filtration membrane plate have been proposed.

  For example, in Patent Document 1, a filter is placed on the filtered water side, a light emitting element is installed on the upstream side, and a light receiving element is installed on the downstream side, and the raw water flows into the filtered water side due to damage to the filtration membrane, and impurities in the raw water are removed. A technique for detecting breakage of a filtration membrane by utilizing the fact that the amount of light captured by a filter and detected by a light receiving element decreases is disclosed.

  Patent Document 2 discloses a method of detecting whether bubbles are generated by pumping air and immersing an object to be inspected in a liquid.

Dana H. Ballard, Christopher M. Brown, “Computer Vision”, first published on May 25, 1987, Translated by Ikuo Fukumura, publisher Japan Computer Association, Computer Science Laboratory Series 29, p. 128-131

JP 2008-49250 A (see, for example, paragraphs [0018] to [0021], [0027], [FIG. 1], [FIG. 2], etc.) JP-A-11-264781 (see, for example, paragraphs [0001], [0008], [0009] [0011] to [0014], [FIG. 1], [FIG. 2], [FIG. 5], etc.) JP 2003-209735 A (see, for example, paragraphs [0022] to [0039], [FIG. 1], [FIG. 10], etc.) JP 2009-65492 A (see, for example, paragraphs [0024] to [0030], [FIG. 2], etc.)

  By the way, the filtration apparatus using a filtration membrane plate becomes a structure by which the several filtration membrane plate was assembled | attached to the module stored in the housing | casing (unit), for example. The module and the housing are formed so that liquid can flow inside and outside. These filtration membrane plates have a rectangular shape, and a large number of membrane holes are formed on both sides, and holes extending from the upper end to the lower end are formed inside. The filtration membrane plate has a lower end portion attached with a sealing tool for sealing the lower end portion, and an upper end portion with a water collecting header. A pipe or the like is connected to the water collection header, and the liquid is filtered by a filtration membrane plate by suction. For this reason, in the filtration apparatus using a filtration membrane plate, it is important that there are no defects such as cracks and film formation defects in a plurality of filtration membrane plates mounted on the device. Moreover, since a defect inspection is performed in the process of manufacturing the filtration membrane plate, it is necessary to avoid impurities from adhering to the filtration membrane plate.

  Therefore, a filtration device using the above-described filtration membrane plate is disposed so as to be immersed in the liquid in the liquid tank, and before the gas is pumped to the filtration membrane plate (before the gas is pumped) and the gas is fed to the filtration membrane plate. The filtration device is photographed with a camera in a pumped state (at the time of gas pumping) to obtain a photographed image before gas pumping and at the time of gas pumping, for example, the difference between these photographed images and the difference obtained thereby Based on the image, it is conceivable to determine the presence or absence of a perforated plate having cracks or film formation defects. In this case, for example, the camera is suspended from the ceiling of a building in which a test device for inspecting the quality of the filtration device is installed. By the way, when a car or the like passes through the vicinity of the building, or when the weather is severe, the building itself vibrates, and the camera may also vibrate due to the vibration, and camera shake may occur in the captured image. . If there is a camera shake in the photographed image at the time of gas pumping, there is a possibility that the bubbles coincide with the filter membrane plate in the photographed image before the gas pumping even if bubbles are generated from the filter membrane plate at the time of gas pumping. there were. Therefore, if the difference between the captured image before gas pumping and the captured image during gas pumping is the difference image without the bubbles, it is determined that there is no defect in the filtration membrane plate even if there are cracks or film formation defects in the filtration membrane plate. There was a possibility. In addition, there is a possibility of erroneous recognition that bubbles are generated even though bubbles are not generated.

  In order to solve this, it is necessary to correct camera shake. Various methods for correcting camera shake have been developed (see, for example, Patent Documents 3 and 4). In the method described in Patent Document 3, first, the amount and direction of camera shake movement are calculated using an optical flow. The result of the optical flow is regarded as camera shake, and the camera shake is corrected by reconstructing the image. In the method described in Patent Document 4, first, a reference portion is installed in the background portion. The reference part is registered as a template. After that, pattern matching is performed, the reference portion is detected, and camera shake is calculated by comparing with the previous detection position. Finally, the camera shake is corrected by calculating the camera shake and reconstructing the image.

  In the techniques described in these Patent Documents 3 and 4, “(1) camera shake is calculated by image processing”, “(2) image reconstruction” is performed, and then “(3) background difference etc. By “image processing”, it is necessary to take four steps of “(4) Judging presence / absence of bubbles”, which takes time.

  In patent document 1 mentioned above, since the filtration membrane is contacting the raw | natural water containing an impurity, it could not be used for the defect inspection in the filtration membrane board manufacturing process made into object.

  From the above, the present invention has been proposed in view of the above-described problems, and even when camera shake occurs, the presence or absence of a filtration membrane plate having cracks or film formation defects can be detected in a short time. It aims at providing the filtration membrane plate test | inspection apparatus by the image processing which can be determined.

The filtration membrane plate inspection apparatus by image processing according to the first invention that solves the above-described problem is a filtration membrane plate inspection device that inspects a plurality of filtration membrane plates provided in the filtration device by image processing, An imaging unit that is attached to the upper ceiling and is immersed in the liquid in the liquid tank and shoots the plurality of filtration membrane plates, and the imaging unit while pumping gas to the plurality of filtration membrane plates An input image creating means for creating an input image from image signals obtained by photographing a plurality of filtration membrane plates, and an optical flow by calculating an optical flow based on the two input images created by the input image creating means The optical flow calculation processing means for outputting the calculation result of the above, and the presence or absence of a damaged filtration membrane plate based on the calculation result of the optical flow is determined, and the damaged filtration membrane plate for outputting the result data And a non-determination processing unit, presence determination processing means of the broken filter membrane plate, in the calculation result of the optical flow, the optical flow in the small area within a predetermined range compared to the input image the entire optical flow When it is “radial” or “irregular”, it is determined that the plurality of filtration membrane plates are damaged .

The filtration membrane plate inspection device by image processing according to the second invention that solves the above-mentioned problem is a filtration membrane plate inspection device by image processing according to the first invention, and the presence / absence judgment processing means of the damaged filtration membrane plate However, in the optical flow calculation result, when all the optical flows within a predetermined range of the input image are in the same direction, it is determined that the plurality of filtration membrane plates are not damaged.

A filtration membrane plate inspection apparatus by image processing according to a third invention that solves the above-mentioned problem is a filtration membrane plate inspection device by image processing according to the first or second invention, and the presence or absence of the damaged filtration membrane plate The determination processing means, in the optical flow calculation result, the optical flow in a small area within the predetermined range of the input image is “different from the direction of the optical flow of the entire captured image” and “constant direction” In this case, it is determined that the plurality of filtration membrane plates are not damaged.

A filtration membrane plate inspection apparatus using image processing according to a fourth invention that solves the above-described problem is a filtration membrane plate inspection device using image processing according to any one of the first to third inventions, wherein the damaged filtration is performed. The film plate presence / absence determination processing means is such that, in the optical flow calculation result, the optical flow in a small region within a predetermined range of the input image is partially different in direction compared to the optical flow of the entire captured image. In this case, it is determined that the plurality of filtration membrane plates are not damaged.

According to the filtration membrane plate inspection apparatus using image processing according to the first invention, even if camera shake occurs, the presence or absence of a filtration membrane plate having a crack or a film formation defect is determined in a short time. be able to. Moreover, the presence or absence of the filtration membrane plate with a crack, a film formation defect, etc. can be determined efficiently.

According to the filtration membrane plate inspection apparatus using image processing according to the second to fourth invention, on the produce operational effects similar to the filtration membrane plate inspection apparatus using image processing according to the first invention, cracks and film defects It is possible to efficiently determine the presence or absence of a filtration membrane plate without any such.

It is a block diagram of the filtration membrane board test | inspection apparatus by the image process which concerns on 1st Example of this invention. It is a block diagram of the filtration membrane board test | inspection apparatus by the image process which concerns on 1st Example of this invention. It is a figure which shows the processing flow in the filtration membrane plate test | inspection apparatus by the image processing which concerns on 1st Example of this invention. It is a figure which shows an example of the picked-up image image | photographed with the camera with which the filtration membrane plate test | inspection apparatus by the image processing which concerns on 1st Example of this invention comprises. FIGS. 5A and 5B are diagrams for explaining a case where there is no bubble in the optical flow calculation result, FIG. 5A shows a captured image (input image) of the previous frame, and FIG. 5B shows a captured image of the current frame. (Input image). It is a figure for demonstrating the case where there exists a bubble by the calculation result of an optical flow. It is a figure for demonstrating the case where a foreign material is mixed in the calculation result of an optical flow. It is a figure for demonstrating the case of a noise by the calculation result of an optical flow.

  Embodiments of a filtration membrane plate inspection apparatus using image processing according to the present invention will be described in each of the examples.

  A filtration membrane plate inspection apparatus using image processing according to a first embodiment of the present invention will be specifically described with reference to FIGS.

  The filtration membrane plate inspection device by image processing according to the present embodiment is a device for identifying a filtration membrane plate having defects such as cracks and film formation defects by image processing in a filtration device equipped with a plurality of filtration membrane plates. .

  As shown in FIG. 1, for example, as shown in FIG. 1, a filtration device 10 using a filtration membrane plate includes a plurality of (five in the illustrated example) filtration membrane plates 11 to 15 through a module (not shown). (Not shown). The module is a framework for assembling a plurality of filtration membrane plates at a predetermined interval, and is formed so that liquid can flow inside and outside. The housing is a container for storing the module, and is formed so that liquid can flow inside and outside. The plurality of filtration membrane plates 11 to 15 are arranged upright in the housing. Further, a plurality of filtration membrane plates 11 to 15 are arranged adjacent to each other in the lateral direction (left and right direction in the figure) at intervals of a dozen millimeters. Each of the plurality of filtration membrane plates 11 to 15 has a rectangular shape, a plurality of membrane holes are formed on both sides, and a hole extending from the upper end to the lower end is formed inside. A lower end sealing tool (not shown) is attached to the lower end portion of the filtration membrane plate, and the lower end side of the filtration membrane plate is sealed by the lower end sealing tool. A water collection header (not shown) is attached to the upper end of the filtration membrane plate. Piping etc. are connected to the water collection header. During operation, water is sucked, and at the time of inspection, a gas such as air is pumped from the pipe to the filtration membrane plate via the water collection header.

  As shown in FIG. 2, the filtration membrane plate inspection apparatus according to the present embodiment includes one camera (upper camera) 2 as an imaging unit, an input image creation unit 21 a, an optical flow calculation processing unit 21 b, and a breakage. A filtration membrane plate presence / absence determination processing unit 21c, an inspection result output unit 21d, a memory (first storage unit) M1, and a memory (second storage unit) M2 are provided. In addition, the input image creation unit 21a, the optical flow calculation processing unit 21b, the presence / absence determination processing unit 21c of the damaged filtration membrane plate, the test result output unit 21d, the memory (first storage unit) M1, and the memory (second storage unit) M2 is configured as arithmetic processing means of a processing computer (not shown). The processing computer includes a monitor (not shown).

  As shown in FIG. 1, the camera 2 is attached to a ceiling 1 of a building where a device for inspecting a filtration device is installed, directly above a water tank (liquid tank) 5 in which a filtration device 10 to be inspected is disposed. It is configured to capture an area directly below the vertical. That is, the camera 2 is configured to photograph a plurality of filtration membrane plates 11 to 15 included in the filtration device 10 that is disposed by being immersed in water (liquid) accommodated in the water tank 5. The camera 2 is connected to the input image creation unit 21a so that an image signal obtained by photographing can be output to the input image creation unit 21a. Thereby, an apparatus can be comprised simply.

  The input image creation unit (input image creation means) 21 a creates an input image from an image signal input from the upper camera 2. When the above filtering device 10 is photographed by the upper camera 2, for example, as shown in FIG. 4, a photographed image 30 in a state where five filtration membrane plates 31 to 35 are arranged adjacent to each other as the input image. Create The input image creation unit 21a is connected to the memory M1 so that the input image can be output to the memory M1. The input image is sent to the optical flow calculation processing unit 21b as necessary through the memories M1 and M2.

  The optical flow calculation processing unit (optical flow calculation processing means) 21b has two input images, for example, a first input image and a second input image, a second input image and a third input image,. That is, the optical flow is calculated based on the (N−1) th input image and the Nth input image. However, the number N of the input image is a number assigned in time series and is a positive number larger than 2. The optical flow is a method for calculating a motion vector based on two input images. For example, the method described in Non-Patent Document 1 described above can be used. The optical flow result, which is the calculation result of the optical flow, is sent to the presence / absence determination processing unit 21b of the damaged filtration membrane plate through the memory M2 as necessary. For example, a captured image (input image) 30 of the previous frame as shown in FIG. 4 or FIG. 5A is obtained, and a predetermined direction, for example, FIG. 5B or FIG. As shown in FIG. 8, when a camera shake occurs in the direction of A1, a captured image (input image) of the current frame arranged at a position moved in the direction opposite to the camera shake direction A1 is obtained. (See, for example, FIG. 5B and FIGS. 6-8).

  Based on the optical flow result, “A. No bubbles”, “B. Bubbles”, “C. Foreign matter mixed” 21c is determined based on the optical flow result. , “D. Noise”, which state is determined, and whether or not the filtration membrane plate is damaged (presence / absence of a damaged filtration membrane plate) is determined (determined). The determination result is sent to the inspection result output processing unit 21d through the memory M2 as necessary.

<A. No air bubbles>
When there are no bubbles, basically only camera shake occurs. Therefore, in a situation where only camera shake occurs, for example, as shown in FIG. 5B, it seems that the filtration membrane plates 41 to 45 that are objects photographed by the camera have moved in the same direction. As indicated by arrows in FIG. 5B, optical flows in the same direction are all detected. Therefore, when all the optical flows in the same direction are detected, for example, the optical flows in the same direction are detected in the region (predetermined range) 48 surrounding the adjacent filtration membrane plates in FIG. In this case, it can be determined that “A. No bubbles”. Thereby, it can be judged (determination) that there are no cracks or film formation defects in the plurality of filtration membrane plates. In addition, the setting location of the said area | region 48 is arbitrary, and it is also possible to set it in the several location where a bubble may generate | occur | produce.

<B. With bubbles>
When there is a bubble, for example, when a single bubble is generated, the filtration membrane plate itself does not move, and the camera and the single bubble move. Therefore, as shown in FIG. While all show the same direction, in the single bubble 56, the optical flow which spreads radially is detected with time passage. Further, when bubbles (continuous bubbles) are continuously generated, the filtration membrane plate itself does not move, and the camera and the open bubbles move. Therefore, as shown in FIG. 6, the optical flows of the respective filtration membrane plates are all in the same direction. On the other hand, in the open cell 57, an irregular optical flow is detected in a small region within a predetermined range. Therefore, when the optical flow in the small region 58 or the small region 59 within the predetermined range shown in FIG. 6 is “radial” or “irregular” as compared with the optical flow of the entire input image, “B. It can be determined that there are bubbles. Thereby, it can be judged (determined) that a plurality of filtration membrane plates have cracks or film formation defects. The setting locations of the small regions 58 and 59 are arbitrary, and can be set at a plurality of locations where bubbles may be generated.

<C. Foreign matter contamination>
When foreign matter such as insects or dust is mixed, for example, when a situation occurs where the foreign matter drifts on the water surface, the filter membrane plate itself does not move and the camera moves and the foreign matter moves on the water surface. Therefore, as shown in FIG. 7, while the optical flows of the filtration membrane plates all indicate the same direction, in the foreign matter 66, an optical flow in a certain direction is detected in a small area. Therefore, compared with the optical flow of the entire input image, the optical flow in the small area 68 within the predetermined range shown in FIG. 7 is “different from the direction of the optical flow of the entire input image” and “constant direction”. In such a case, it can be determined that “C. foreign matter”. Thereby, it can be judged (determination) that there are no cracks or film formation defects in the plurality of filtration membrane plates. In addition, the setting location of the small area 68 is arbitrary, and can be set at a plurality of locations.

<D. Noise>
When noise is included in the calculation result of the optical flow, the filtration membrane plate itself does not move, and only the camera moves. Therefore, as shown in FIG. 8, while the optical flows of each filtration membrane plate all indicate the same direction, In the case of the noise 76, an optical flow in a certain direction is detected in a small area. Therefore, when the optical flow in the small area 78 within the predetermined range shown in FIG. 8 is “differing in the direction of only part” compared to the optical flow of the entire input image, compared to the optical flow of the entire input image. Can be determined as “D. noise”. Thereby, it can be judged (determination) that there are no cracks or film formation defects in the plurality of filtration membrane plates. In addition, the setting location of the said small area | region 78 is arbitrary, and it is also possible to set in several places.

  The inspection result output processing unit (inspection result output processing means) 21d creates an output result (inspection result) from the determination result, and outputs the output result to the monitor. Specifically, when the damaged filtration membrane plate presence / absence determination processing unit 21c determines that “B. Air bubbles are present”, the inspection result output processing unit 21d outputs a broken filtration membrane plate “Yes” to the monitor. To do. In cases other than “B. presence of air bubbles”, that is, if the damaged filter membrane presence / absence determination processing unit 21c determines that “A. no air bubbles”, “C. foreign matter”, “D. noise”, The result output processing unit 21d outputs to the monitor that the damaged filtration membrane plate is “No”. Thereby, an operator can confirm an output result.

  The memory M1 stores various data, for example, input images. The memory M2 stores various data such as an input image, an optical flow result, and a determination result.

  Here, based on FIG. 3, a procedure for inspecting a filtration device on which a plurality of filtration membrane plates are mounted using the above-described filtration membrane plate inspection device by image processing will be briefly described.

  First, before the inspection, the filtering device is arranged at a predetermined position, and the camera and the lighting device are arranged at a predetermined position. Specifically, as shown in FIG. 1, a plurality of filtration membrane plates 11 to 15 of the filtration device 10 are arranged so as to be immersed in water in a water tank 5 in which liquid such as pure water without impurities is stored, A camera 2 for photographing the filtration device 10 from above is arranged on the ceiling 1 above the filtration device 10 (water tank), and lighting devices 3 and 4 for illuminating the filtration device 10 from above are arranged adjacent to the camera 2. . The rotation about the optical axis of the camera 11 is adjusted so that the filter membrane plates 11 to 15 and the image axis of the camera 11 are parallel. However, these inspections are performed in a state where the entire filtration membrane plates 11 to 15 are immersed in water. Although details will be described later, air of a specified pressure is pumped, an input image is created from the image signal from the camera, and an optical flow is calculated from the two input images. The presence or absence of a filter membrane plate is judged.

  Subsequently, as shown in FIG. 3, in the above processing computer, the input image creation unit 21 a is obtained by photographing with the camera 2 in a state where air is fed to the filtration membrane plates 11 to 15 at a specified pressure. Based on the received image signal, an initial input image (first input image) is created (step P1). Subsequently, the input image creation unit 21a performs input image (second and subsequent images) based on the image signal obtained by photographing with the camera 2 in a state where air is fed to the filtration membrane plates 11 to 15 at a specified pressure. (Input image) is created (step P2). That is, two or more input images are obtained in time series by Step P1 and Step P2. In addition, when image | photographing with the camera 2, you may make it illuminate the filtration apparatus 10 with the illuminating devices 3 and 4 suitably.

  Subsequently, the optical flow calculation processing unit 21b calculates an optical flow based on the latest two input images among the input images obtained in Steps P1 and P2 (Step P3). Subsequently, the presence / absence judgment processing unit 21c of the damaged filtration membrane plate performs “A. No bubble”, “B. Bubble”, “C. Foreign matter” based on the optical calculation result obtained by the optical flow calculation processing unit 21b. It is determined which of the four types of “mixing” and “D. noise” is present, and it is determined whether or not there is a filtration membrane plate having a crack or a film formation defect (step P4).

  Subsequently, the arithmetic processing means of the processing computer determines whether or not the inspection of the filtering device has been completed (step P5). As a result of the determination, if the inspection of the filtration device is not completed, the process returns to Step P2. On the other hand, when the inspection of the filtration device is completed, the process proceeds to Step P6.

  Finally, the inspection result output processing unit 21d outputs the determination result obtained in step P4 (the presence or absence of a filtration membrane plate having cracks or film formation defects) to the monitor or the like as an output result (step P6). .

  Therefore, according to the filtration membrane plate inspection apparatus using image processing according to the present embodiment, even when camera shake occurs, the optical flow is calculated from two input images in time series, and the optical flow result is calculated. Based on these states, it is determined which of the four states of “A. No bubbles”, “B. Bubbles”, “C. Foreign matter mixed in”, and “D. Noise”. It is possible to determine the presence or absence of a filtration membrane plate with defects. That is, even when camera shake occurs, it is possible to accurately inspect a filtration membrane plate having cracks and film formation defects.

  In addition, even if camera shake occurs, it is possible to efficiently determine the presence or absence of a filtration membrane plate with cracks or film formation defects using only the optical flow results without correcting camera shake or reconstructing images. can do.

  In the above description, the case in which the inspection result is output in the last step P6 has been described. However, the inspection result can be output in real time between step P4 and step P5.

  In the first embodiment described above, the description has been made using the filtration membrane plate inspection apparatus based on the image processing for inspecting the five filtration membrane plates. It is also possible to provide a filtration membrane plate inspection apparatus by image processing for inspecting the above filtration membrane plate.

  In the first embodiment described above, the filtering device is arranged in the water tank, an input image is created from an image signal obtained by photographing this with a camera, an optical flow is calculated from the two input images, and this calculation is performed. Based on the results, we explained the filtration membrane plate inspection device based on image processing that determines the presence or absence of filtration membrane plates with cracks or membrane formation defects.However, a module with a plurality of filtration membrane plates assembled is placed in the water tank. Create an input image from the image signal obtained by photographing this with the camera, calculate the optical flow from the two input images, and based on this calculation result, whether there is a filtration membrane plate with cracks and film formation defects It is also possible to use a filtration membrane plate inspection apparatus based on image processing for determining the above.

  According to the filtration membrane plate inspection apparatus using image processing according to the present invention, even if camera shake occurs, it is possible to determine the presence or absence of a filtration membrane plate having cracks or film formation defects in a short time. Therefore, it can be used extremely beneficially in the water treatment industry for producing filtration membrane plates.

1 Ceiling 2 Camera (Upper camera)
3, 4 Lighting 5 Water tank 10 Filtration apparatus 11-15 Filtration membrane plate 21a Input image creation unit 21b Optical flow calculation processing unit 21c Presence judgment processing unit 21d of damaged filtration membrane plate Inspection result output processing unit 30 Photographed image of previous frame (input image)
31-35 Filtration membrane plate 40 in previous frame captured image (input image) Captured image (input image) of current frame
41-45 Filter membrane 48 in current frame captured image (input image) Region 50 Captured image (input image) of current frame
51-55 Filtration membrane 56 in the captured image (input image) of the current frame Single bubble 57 Continuous bubbles 58, 59 Small area 60 Captured image of the current frame (input image)
61 to 65 Filtration membrane 66 in the captured image (input image) of the current frame Foreign matter 68 Small region 70 Captured image of the current frame (input image)
71-75 Filtration membrane 76 in the captured image (input image) of the current frame Noise 78 Small area A1 Camera blur direction M1, M2 Memory

Claims (4)

A filtration membrane plate inspection device that inspects a plurality of filtration membrane plates provided in the filtration device by image processing,
An imaging unit that is attached to the ceiling above the liquid tank and shoots the plurality of filtration membrane plates disposed by being immersed in the liquid in the liquid tank;
An input image creating means for creating an input image from an image signal obtained by photographing the plurality of filtration membrane plates by the photographing means while pumping gas to the plurality of filtration membrane plates;
Optical flow calculation processing means for calculating an optical flow based on the two input images created by the input image creating means and outputting a calculation result of the optical flow;
Determining the presence or absence of a damaged filtration membrane plate based on the calculation result of the optical flow, comprising the presence or absence judgment processing means of a damaged filtration membrane plate for outputting the result data ;
In the calculation result of the optical flow, the optical flow in a small region within a predetermined range is “radial” or “irregular” in the calculation result of the optical flow. If it is, it is determined that the plurality of filtration membrane plates are damaged . A filtration membrane plate inspection apparatus based on image processing, wherein: A filtration membrane plate inspection apparatus using image processing according to claim 1 ,
The damaged filtration membrane plate presence / absence determination processing means has no damage to the plurality of filtration membrane plates when all the optical flows within a predetermined range of the input image are in the same direction in the calculation result of the optical flow. A filtration membrane plate inspection apparatus using image processing, characterized in that A filtration membrane plate inspection apparatus using image processing according to claim 1 or 2 ,
In the calculation result of the optical flow, the optical flow in the small area within the predetermined range of the input image is “different from the direction of the optical flow of the entire captured image” in the calculation result of the optical flow. And when it becomes "a fixed direction", it determines with there being no damage in these filtration membrane plates, The filtration membrane plate test | inspection apparatus by the image processing characterized by the above-mentioned. A filtration membrane plate inspection apparatus using image processing according to any one of claims 1 to 3 ,
In the optical flow calculation result, the optical filter in the small area within the predetermined range of the input image is only partially compared with the optical flow of the entire captured image. When the directions are different, it is determined that there is no breakage in the plurality of filtration membrane plates.

Images