Valve plate polishing surface recognition and detection method and system based on machine vision

Abstract

The present invention discloses a valve plate polishing surface recognition and detection method and system based on machine vision, which relates to the field of polishing detection technology, including: obtaining a polished valve plate and an unpolished valve plate and performing polishing treatment, obtaining smooth water ripple features and rough water ripple features based on machine vision; obtaining multiple valve plate water ripple features based on feature screening processing; performing surface recognition and detection based on the valve plate water ripple features; the present invention is used to solve the problem that in the existing valve plate polishing surface recognition and detection method, the polished workpiece needs to be cleaned and dried for detection of the polishing process, and the polishing state and unpolished defects of the workpiece cannot be detected during the polishing process, resulting in the need to consume excess cleaning and drying resources to repeatedly process the workpiece, thereby causing excessive resource consumption.

Description

Technical Field

[0001] The present invention relates to the field of polishing detection technology, and in particular to a valve plate polishing surface recognition and detection method and system based on machine vision. Background Art

[0002] ‌Valve disc is a device for controlling fluid flow, usually used in hydraulic systems;‌ In hydraulic systems, valve disc controls and protects the system by controlling the flow and direction of the liquid; The valve consists of three parts: the main body, the valve seat and the valve disc. The main body is the body on which the valve disc is installed, the valve seat is used to guide, seal and support the valve disc, and the valve disc is a freely movable circular thin disc responsible for opening or closing the passage; ‌Polishing‌ is a common surface treatment method, which achieves grinding and polishing of the workpiece surface through the rotational motion of the abrasive on the workpiece surface and the linear motion relative to the workpiece; Polishing methods include plane polishing, cylindrical polishing and spherical polishing, etc., which are suitable for workpieces of different shapes.

[0003] Existing methods for surface recognition and detection of valve plate polishing are usually improvements in improving the polishing efficiency of workpieces. For example, an equipment unit for cleaning, drying and detecting the polished workpiece is added at the outlet of the polishing device to reduce the transportation of the workpiece during the polishing process, thereby improving the processing efficiency of the workpiece during polishing. Although this improved method can improve the polishing efficiency by reducing the transportation steps, the detection of the polishing process requires the cleaning and drying of the polished workpiece, and it is impossible to detect the polishing state and unpolished defects of the workpiece during the polishing process, resulting in the consumption of excess cleaning when the workpiece is not completely polished. The cleaning and drying resources are used to repeatedly process the workpiece, thus causing the problem of excessive resource consumption; other surface recognition detection methods for valve plate polishing are usually improvements in the transportation of workpieces, but still cannot solve the problem that the polished workpiece needs to be cleaned and dried for detection during the polishing process, and the polishing state and unpolished defects of the workpiece cannot be detected during the polishing process, resulting in the need to consume excess cleaning and drying resources to repeatedly process the workpiece when the workpiece is not completely polished, thus causing the problem of excessive resource consumption. In view of this, it is necessary to improve the existing surface recognition detection methods for valve plate polishing. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the prior art to at least a certain extent. By proposing a valve plate polishing surface recognition and detection method and system based on machine vision, it is used to solve the problem that the existing valve plate polishing surface recognition and detection method requires the polished workpiece to be cleaned and dried for detection of the polishing process, and is unable to detect the polishing state of the workpiece and unpolished defects during the polishing process, resulting in the need to consume excess cleaning and drying resources to repeatedly process the workpiece when the workpiece is not completely polished, thereby causing the problem of excessive resource consumption.

[0005] To achieve the above objectives, in a first aspect, the present application provides a valve plate polishing surface recognition and detection method based on machine vision, comprising the following steps:

[0006] Obtaining the polished valve disc and the unpolished valve disc respectively, and putting them into a polishing machine for polishing, wherein the polishing includes smooth polishing and defect polishing; analyzing the polishing process based on machine vision, and obtaining smooth water pattern characteristics and rough water pattern characteristics based on the analysis results;

[0007] Performing feature screening processing on rough water ripple features based on smooth water ripple features, and obtaining multiple valve sheet water ripple features based on the processing results;

[0008] Surface recognition and detection is performed during valve plate polishing based on the valve plate water pattern characteristics.

[0009] Further, obtaining the polished valve disc and the unpolished valve disc respectively and placing them in a polishing machine for polishing treatment respectively includes:

[0010] Obtain a polished valve disc and record it as a smooth valve disc; obtain multiple unpolished valve discs and record them as rough valve discs; obtain the defect types on the surfaces of all rough valve discs and record them as valve disc defect FX1 to valve disc defect FX2. c ;

[0011] For any rough valve plate, the valve plate defects FX1 to FX2 on the rough valve plate surface are obtained based on AI recognition. c , and record the valve sheet defect FX with the largest number as the representative defect of the rough valve sheet; obtain the representative defects of all rough valve sheets.

[0012] Furthermore, the polishing process is analyzed based on machine vision, and the smooth water pattern characteristics and rough water pattern characteristics are obtained based on the analysis results, including:

[0013] The maximum speed of the polishing machine when polishing the valve plate is recorded as the maximum polishing speed; the smooth valve plate is polished and polished;

[0014] The smooth polishing process is as follows: put the smooth valve plate into a polishing machine with a maximum polishing speed, align the center of the camera from directly above the smooth valve plate to the center of the smooth valve plate based on machine vision and start recording; when the polishing time is a unit time, stop polishing and stop recording, and record the image taken at this time as a smooth polishing image, wherein the unit time is the time for polishing the valve plate once, and the center of the smooth valve plate coincides with the center of the sandpaper during polishing.

[0015] Furthermore, the smooth polishing process also includes:

[0016] For image α corresponding to any frame except the first frame in the smooth polishing image, the water ripple area in image α is obtained based on machine vision, wherein the water ripple area is the area where the cooling water is located in the area other than the smooth valve plate; a plane rectangular coordinate system is established, which is recorded as the water ripple analysis coordinate system, wherein the units of the X-axis and the Y-axis of the water ripple analysis coordinate system are both cm; the water ripple area of ​​image α is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve plate, the midpoint of the smooth valve plate is marked in the first quadrant of the water ripple analysis coordinate system, which is recorded as the analysis calibration point;

[0017] The image corresponding to the previous frame of the frame where image α is located is recorded as the previous frame image; the water ripple area of ​​the previous frame image is obtained, and the water ripple area of ​​the previous frame image is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve sheet, the midpoint of the smooth valve sheet corresponding to the water ripple area of ​​the previous frame image is made to coincide with the analysis calibration point.

[0018] Furthermore, the smooth polishing process also includes:

[0019] The water ripple area of ​​the previous frame image is rotated for one circle with the analysis calibration point as the center, and the coincidence parameter of the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α during the rotation is obtained in real time, wherein the coincidence parameter is the value of the area of ​​the coincidence area between the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α divided by the area of ​​the water ripple area of ​​the image α;

[0020] The maximum overlap parameter obtained is recorded as the watermark similarity parameter of image α.

[0021] Furthermore, the smooth polishing process also includes:

[0022] Obtaining water ripple similarity parameters of images corresponding to all frames except the first frame in the smooth polishing image, and recording the image corresponding to the largest water ripple similarity parameter as the smooth analysis image, wherein, when the largest water ripple similarity parameter corresponds to multiple images, the image corresponding to the smallest time in the smooth polishing image is recorded as the smooth analysis image;

[0023] The grayscale image of the water ripple area corresponding to the smooth analysis image is obtained and recorded as a smooth grayscale image; the image formed by the edge of the smooth grayscale image is recorded as a smooth grayscale graph, and the pixel points in the smooth grayscale graph are binarized; the lengths of all black line segments in the smooth grayscale graph are stored in an array γ, and the array γ is recorded as a smooth water ripple feature.

[0024] Furthermore, the polishing process is analyzed based on machine vision, and the smooth water pattern characteristics and the rough water pattern characteristics are obtained based on the analysis results, which also includes:

[0025] For any valve plate defect FX, the rough valve plate representing the defect of the valve plate defect FX is recorded as a rough-processed valve plate, and all rough-processed valve plates are subjected to defect polishing treatment;

[0026] The defect polishing process is as follows: for any rough-processed valve plate, based on the processing method of smooth polishing process, the rough-processed valve plate is placed in a polishing machine with a maximum polishing speed, and an image is captured based on machine vision, and the obtained image is recorded as a defect polishing image; based on a method of obtaining smooth water pattern features from smooth polishing images, the defect polishing image is analyzed, and the obtained array is recorded as a rough water pattern array of the rough-processed valve plate;

[0027] Obtain the rough water pattern arrays of all roughened valve sheets, store the modes of all rough water pattern arrays into array β, and record array β as the rough water pattern feature of valve sheet defect FX;

[0028] Get the rough water characteristics of all valve defect FX.

[0029] Furthermore, based on the smooth water pattern feature, the rough water pattern feature is subjected to feature screening processing, and based on the processing result, a plurality of valve sheet water pattern features are obtained, including:

[0030] For any rough water pattern feature of a valve defect FX, when the set corresponding to the valve defect FX is a subset of the set corresponding to the smooth water pattern feature, the rough water pattern feature of the valve defect FX is recorded as the valve water pattern feature;

[0031] When the set corresponding to the valve feature FX is not a subset of the set corresponding to the smooth water pattern feature, the set consisting of numbers in the set corresponding to the valve feature FX that are different from the set corresponding to the smooth water pattern feature is recorded as the valve water pattern feature of the valve feature FX;

[0032] Get the valve water pattern features corresponding to all valve defect FX.

[0033] Furthermore, the surface recognition detection based on the valve plate water pattern characteristics during valve plate polishing includes:

[0034] When the valve plate is polished, the polished valve plate is recorded as a real-time processed valve plate; based on the defect polishing process, a rough water pattern feature corresponding to the real-time processed valve plate is obtained and recorded as a real-time water pattern feature;

[0035] When the set corresponding to the real-time water pattern feature is a subset of the set corresponding to the smooth water pattern feature, the valve plate defect FX corresponding to the valve plate water pattern feature that is the same as the real-time water pattern feature is recorded as the key defect of the real-time processed valve plate;

[0036] When the set corresponding to the real-time water pattern feature is not a subset of the set corresponding to the smooth water pattern feature, the valve sheet defects FX corresponding to the valve sheet water pattern features with the largest number of the same numbers as those in the set of real-time water pattern features are recorded as the key defects of the valve sheet to be processed in real time;

[0037] The valve plate is polished again based on the key defects.

[0038] In a second aspect, the present application also provides a valve plate polishing surface recognition and detection system based on machine vision, including a polishing water pattern analysis module, a water pattern feature acquisition module and a polishing detection module;

[0039] The polishing water pattern analysis module is used to obtain the polished valve disc and the unpolished valve disc respectively, and put them into the polishing machine for polishing treatment respectively, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and the smooth water pattern characteristics and the rough water pattern characteristics are obtained based on the analysis results;

[0040] The water pattern feature acquisition module is used to perform feature screening processing on the rough water pattern feature based on the smooth water pattern feature, and acquire multiple valve plate water pattern features based on the processing result;

[0041] The polishing detection module is used to perform surface recognition detection when the valve plate is polished based on the water pattern characteristics of the valve plate.

[0042] Beneficial effects of the present invention: the present invention first obtains a polished valve disc and an unpolished valve disc respectively, and puts them into a polishing machine for polishing treatment respectively, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and smooth water pattern characteristics and rough water pattern characteristics are obtained based on the analysis result. The advantage of this is that by obtaining the smooth water pattern characteristics and the rough water pattern characteristics based on machine vision, the water pattern characteristics of the polished valve disc and the valve disc with different defects when being polished can be obtained more accurately, which is helpful for effective comparison in subsequent analysis, thereby identifying the polishing process of the valve disc with different defects, so as to complete the surface recognition detection of the valve disc when being polished;

[0043] The present application also performs feature screening processing on rough water ripple features based on smooth water ripple features, and obtains multiple valve plate water ripple features based on the processing results; finally, surface identification detection is performed based on the valve plate water ripple features when the valve plate is polished. The advantage of this is that by performing surface identification detection based on multiple valve plate water ripple features when the valve plate is polished, it is possible to obtain a large number of defect types that still exist in the valve plate during the valve plate polishing process, so that the staff can adjust the polishing sandpaper or rotation speed based on the defect types that still exist in the valve plate, thereby realizing the detection of the polishing state of the valve plate and the unpolished defects during the polishing process, thereby preventing the problem of excessive resource consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a functional block diagram of the system of the present invention;

[0045] Figure 2 is a flow chart of the steps of the method of the present invention;

[0046] Figure 3 A schematic diagram of a water pattern analysis coordinate system of the present invention;

[0047] Figure 4 Schematic diagram of a smooth grayscale image and a smooth grayscale graph of the present invention;

[0048] Figure 5 It is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Example 1, please refer to Figure 1 As shown, the present application provides a valve plate polishing surface recognition and detection system based on machine vision, including a polishing water pattern analysis module, a water pattern feature acquisition module and a polishing detection module;

[0051] The polishing water pattern analysis module is used to obtain the polished valve disc and the unpolished valve disc respectively, and put them into the polishing machine for polishing treatment respectively, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and the smooth water pattern characteristics and the rough water pattern characteristics are obtained based on the analysis results;

[0052] The polishing watermark analysis module includes a polishing watermark analysis unit. The polishing watermark analysis unit is configured with a polishing watermark analysis strategy. The polishing watermark analysis strategy includes:

[0053] Obtain a polished valve disc and record it as a smooth valve disc; obtain multiple unpolished valve discs and record them as rough valve discs; obtain the defect types on the surfaces of all rough valve discs and record them as valve disc defect FX1 to valve disc defect FX2. c ;

[0054] In the specific implementation process, in this embodiment, the valve plate defect FX may include defects that may exist on the valve plate surface, such as deformation, cracks, corrosion, pores, and cracks. In actual application, the valve plate defect FX may be defined according to the types of defects that may exist on the valve plate surface to be detected, so as to avoid the problem of increasing invalid data analysis by defining defects that will not appear;

[0055] For any rough valve plate, the valve plate defects FX1 to FX2 on the rough valve plate surface are obtained based on AI recognition. c and record the valve sheet defect FX with the largest number as the representative defect of the rough valve sheet; obtain the representative defects of all rough valve sheets;

[0056] In the specific implementation process, for example, during a data analysis, based on AI recognition, it is found that there are 10 pores and 2 cracks on the surface of a rough valve disc, which means that the surface of the rough valve disc has a large number of pores, that is, it is greatly affected by the pores. Therefore, during polishing, the pores should be polished mainly, and the pores can be recorded as representative defects of the rough valve disc;

[0057] The maximum speed of the polishing machine when polishing the valve plate is recorded as the maximum polishing speed; the smooth valve plate is polished and polished;

[0058] The smooth polishing process is as follows: put the smooth valve plate into a polishing machine with a maximum polishing speed, align the center of the camera from directly above the smooth valve plate to the center of the smooth valve plate based on machine vision and start recording; when the polishing time is a unit time, stop polishing and stop recording, and record the image taken at this time as a smooth polishing image, wherein the unit time is the time for polishing the valve plate once, and the center of the smooth valve plate coincides with the center of the sandpaper during polishing;

[0059] In the specific implementation process, in this embodiment, the unit time is set to 5s. In actual processing, the unit time can be adjusted according to the actual sandpaper grit of the polishing machine and the polishing requirements of the valve plate to facilitate the simulation of the actual polishing process; the purpose of aligning the center of the smooth valve plate with the center of the sandpaper is to facilitate subsequent image analysis;

[0060] For image α corresponding to any frame except the first frame in the smooth polishing image, the water ripple area in image α is obtained based on machine vision, wherein the water ripple area is the area where the cooling water is located in the area other than the smooth valve plate; a plane rectangular coordinate system is established, which is recorded as the water ripple analysis coordinate system, wherein the units of the X-axis and the Y-axis of the water ripple analysis coordinate system are both cm; the water ripple area of ​​image α is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve plate, the midpoint of the smooth valve plate is marked in the first quadrant of the water ripple analysis coordinate system, which is recorded as the analysis calibration point;

[0061] In the specific implementation process, for example, during a data processing, the water pattern analysis coordinate system is obtained as follows: Figure 3 As shown, point JZ is the analysis calibration point, the area where circle ZZ1 is located is the area where the smooth valve disc is located, and the area in circle ZZ2 but not in circle ZZ1 is the water pattern area; because the cooling water on the surface of the sandpaper will produce water patterns due to the influence of the valve disc during polishing, by analyzing the water patterns when the smooth valve disc is polished and the water patterns when the rough valve disc is polished, different characteristics of water patterns when different defects are polished in the valve disc can be obtained, which is helpful to detect defects on the surface of the valve disc during the polishing process, so as to perform targeted polishing treatment, so as to solve the problem that the polishing state of the valve disc and the unpolished defects cannot be detected during the polishing process, resulting in the need to consume excess cleaning and drying resources to repeatedly process the workpiece when the workpiece is not completely polished, thereby causing excessive resource consumption;

[0062] The image corresponding to the previous frame of the frame where the image α is located is recorded as the previous frame image; the water ripple area of ​​the previous frame image is obtained, and the water ripple area of ​​the previous frame image is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve sheet, the midpoint of the smooth valve sheet corresponding to the water ripple area of ​​the previous frame image is made to coincide with the analysis calibration point;

[0063] The water ripple area of ​​the previous frame image is rotated for one circle with the analysis calibration point as the center, and the coincidence parameter of the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α during the rotation is obtained in real time, wherein the coincidence parameter is the value of the area of ​​the coincidence area between the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α divided by the area of ​​the water ripple area of ​​the image α;

[0064] In the specific implementation process, for example, during a data processing, when the water ripple area of ​​the previous frame image rotates around the analysis calibration point, the maximum overlap parameter is 90%, which means that the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α have a maximum overlap area of ​​90%. The larger the overlap area, the less the valve plate is affected by the polishing at this time, that is, the valve plate has been polished.

[0065] The maximum coincidence parameter obtained is recorded as the watermark similarity parameter of image α;

[0066] Obtaining water ripple similarity parameters of images corresponding to all frames except the first frame in the smooth polishing image, and recording the image corresponding to the largest water ripple similarity parameter as the smooth analysis image, wherein, when the largest water ripple similarity parameter corresponds to multiple images, the image corresponding to the smallest time in the smooth polishing image is recorded as the smooth analysis image;

[0067] In the specific implementation process, by obtaining the smooth analysis image, the image of the valve plate after polishing in the polishing machine can be obtained, which is helpful to provide reference data for the subsequent acquisition of the water pattern characteristics of the valve plate;

[0068] Obtain the grayscale image of the water ripple area corresponding to the smooth analysis image, which is recorded as a smooth grayscale image; record the image formed by the edge of the smooth grayscale image as a smooth grayscale graph, and perform binary processing on the pixels in the smooth grayscale graph; store the lengths of all black line segments in the smooth grayscale graph into an array γ, and record the array γ as a smooth water ripple feature;

[0069] In the specific implementation process, for example, during a data processing, the smooth grayscale image and smooth grayscale graph are obtained as follows: Figure 4 As shown in GH1 and GH2 in GH2, the lengths of line segments XD1 to XD4 in GH2 can be stored in array γ, and the smooth water ripple features can be obtained.

[0070] The polishing water mark analysis strategy also includes: for any valve plate defect FX, the rough valve plate representing the defect as the valve plate defect FX is recorded as a rough-processed valve plate, and all rough-processed valve plates are subjected to defect polishing;

[0071] The defect polishing process is as follows: for any rough-processed valve plate, based on the processing method of smooth polishing process, the rough-processed valve plate is placed in a polishing machine with a maximum polishing speed, and an image is captured based on machine vision, and the obtained image is recorded as a defect polishing image; based on a method of obtaining smooth water pattern features from smooth polishing images, the defect polishing image is analyzed, and the obtained array is recorded as a rough water pattern array of the rough-processed valve plate;

[0072] Obtain the rough water pattern arrays of all roughened valve sheets, store the modes of all rough water pattern arrays into array β, and record array β as the rough water pattern feature of valve sheet defect FX;

[0073] In the specific implementation process, in this embodiment, the method of obtaining rough water pattern characteristics by roughening the valve plate is consistent with the method of obtaining smooth water pattern characteristics by smoothing the valve plate;

[0074] Get the rough water characteristics of all valve defect FX.

[0075] The water pattern feature acquisition module is used to perform feature screening processing on rough water pattern features based on smooth water pattern features, and acquire multiple valve plate water pattern features based on the processing results; the water pattern feature acquisition module includes a water pattern feature acquisition unit, and the water pattern feature acquisition unit is configured with a water pattern feature acquisition strategy, and the water pattern feature acquisition strategy includes:

[0076] For any rough water pattern feature of a valve defect FX, when the set corresponding to the valve defect FX is a subset of the set corresponding to the smooth water pattern feature, the rough water pattern feature of the valve defect FX is recorded as the valve water pattern feature;

[0077] In the specific implementation process, for example, in a data processing, the sets corresponding to the valve defect FX are 3, 5 and 20, and the sets corresponding to the smooth water pattern features are 6, 2, 3, 5 and 20. Then, through analysis, it can be obtained that the set corresponding to the valve defect FX is a subset of the set corresponding to the smooth water pattern features. Then, the rough water pattern features of the valve defect FX can be recorded as the valve water pattern features, so as to facilitate the subsequent identification of the valve defect FX.

[0078] When the set corresponding to the valve feature FX is not a subset of the set corresponding to the smooth water pattern feature, the set consisting of numbers in the set corresponding to the valve feature FX that are different from the set corresponding to the smooth water pattern feature is recorded as the valve water pattern feature of the valve feature FX;

[0079] Get the valve water pattern features corresponding to all valve defect FX.

[0080] The polishing detection module is used to perform surface recognition detection when the valve plate is polished based on the water pattern characteristics of the valve plate; the polishing detection module includes a polishing detection unit, and the polishing detection unit is configured with a polishing detection strategy, and the polishing detection strategy includes:

[0081] When the valve plate is polished, the polished valve plate is recorded as a real-time processed valve plate; based on the defect polishing process, a rough water pattern feature corresponding to the real-time processed valve plate is obtained and recorded as a real-time water pattern feature;

[0082] When the set corresponding to the real-time water pattern feature is a subset of the set corresponding to the smooth water pattern feature, the valve plate defect FX corresponding to the valve plate water pattern feature that is the same as the real-time water pattern feature is recorded as the key defect of the real-time processed valve plate;

[0083] In the specific implementation process, by obtaining key defects, the types of defects that still exist in large numbers in the valve plate can be obtained during the polishing process of the valve plate, so that the staff can adjust the polishing sandpaper or rotation speed based on the types of defects that still exist in the valve plate, thereby realizing the detection of the polishing state of the valve plate and the unpolished defects during the polishing process, and preventing the problem of excessive resource consumption;

[0084] When the set corresponding to the real-time water pattern feature is not a subset of the set corresponding to the smooth water pattern feature, the valve sheet defects FX corresponding to the valve sheet water pattern features with the largest number of the same numbers as those in the set of real-time water pattern features are recorded as the key defects of the valve sheet to be processed in real time;

[0085] The valve plate is polished again based on the key defects.

[0086] Example 2, please refer to Figure 2 As shown, the present application also provides a valve plate polishing surface recognition and detection method based on machine vision, comprising the following steps:

[0087] Step S1, respectively obtaining a polished valve disc and an unpolished valve disc, and respectively placing them in a polishing machine for polishing, wherein the polishing includes smooth polishing and defect polishing; analyzing the polishing process based on machine vision, and obtaining smooth water pattern features and rough water pattern features based on the analysis results;

[0088] Step S1 includes: Step S101, obtaining a polished valve disc and recording it as a smooth valve disc; obtaining multiple unpolished valve discs and recording them as rough valve discs respectively; obtaining the defect types on the surfaces of all rough valve discs and recording them as valve disc defects FX1 to valve disc defects FX2 respectively. c ;

[0089] Step S102: for any rough valve plate, obtain valve plate defects FX1 to valve plate defects FX2 on the rough valve plate surface based on AI recognition. c and record the valve sheet defect FX with the largest number as the representative defect of the rough valve sheet; obtain the representative defects of all rough valve sheets;

[0090] Step S103, recording the maximum speed of the polishing machine when polishing the valve plate as the maximum polishing speed; performing a smooth polishing process on the smooth valve plate;

[0091] Step S104, smooth polishing processing is as follows: Step S1041, put the smooth valve plate into a polishing machine with a maximum polishing speed, align the center of the camera from directly above the smooth valve plate to the center of the smooth valve plate based on machine vision and start recording; when the polishing time is a unit time, stop polishing and stop recording, and record the image taken at this time as a smooth polishing image, wherein the unit time is the time for polishing the valve plate once, and the center of the smooth valve plate coincides with the center of the sandpaper during polishing;

[0092] Step S1042, for an image α corresponding to any frame except the first frame in the smooth polishing image, a water ripple area in the image α is obtained based on machine vision, wherein the water ripple area is the area where the cooling water is located in the area other than the smooth valve plate; a plane rectangular coordinate system is established, which is recorded as a water ripple analysis coordinate system, wherein the units of the X-axis and the Y-axis of the water ripple analysis coordinate system are both cm; the water ripple area of ​​the image α is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve plate, the midpoint of the smooth valve plate is marked in the first quadrant of the water ripple analysis coordinate system, which is recorded as an analysis calibration point;

[0093] Step S1043, record the image corresponding to the previous frame of the frame where the image α is located as the previous frame image; obtain the water ripple area of ​​the previous frame image, and place the water ripple area of ​​the previous frame image in the first quadrant of the water ripple analysis coordinate system after grayscale processing, and based on the positional relationship between the water ripple area and the smooth valve sheet, make the midpoint of the smooth valve sheet corresponding to the water ripple area of ​​the previous frame image coincide with the analysis calibration point;

[0094] Step S1044, rotating the water ripple area of ​​the previous frame image for one circle with the analysis calibration point as the center, and obtaining in real time the overlap parameter between the water ripple area of ​​the previous frame image and the water ripple area of ​​image α during the rotation, wherein the overlap parameter is the value of the area of ​​the overlap between the water ripple area of ​​the previous frame image and the water ripple area of ​​image α divided by the area of ​​the water ripple area of ​​image α;

[0095] Step S1045, recording the obtained maximum coincidence parameter as the watermark similarity parameter of image α;

[0096] Step S1046, obtaining water ripple similarity parameters of images corresponding to all frames except the first frame in the smooth polishing image, and recording the image corresponding to the largest water ripple similarity parameter as the smooth analysis image, wherein, when the largest water ripple similarity parameter corresponds to multiple images, the image corresponding to the smallest time in the smooth polishing image is recorded as the smooth analysis image;

[0097] Step S1047, obtain the grayscale image of the water ripple area corresponding to the smooth analysis image, and record it as a smooth grayscale image; record the image formed by the edge of the smooth grayscale image as a smooth grayscale graphic, and perform binarization on the pixels in the smooth grayscale graphic; store the lengths of all black line segments in the smooth grayscale graphic into an array γ, and record the array γ as a smooth water ripple feature.

[0098] Step S1 further includes: step S105, for any valve sheet defect FX, recording the rough valve sheet representing the defect as the valve sheet defect FX as a roughened valve sheet, and performing defect polishing treatment on all roughened valve sheets;

[0099] Step S106, defect polishing treatment is as follows: Step S1061, for any rough-processed valve plate, based on the processing method of smooth polishing treatment, the rough-processed valve plate is placed in a polishing machine with a maximum polishing speed, and an image is captured based on machine vision, and the obtained image is recorded as a defect polishing image; based on a method of obtaining smooth water pattern features from smooth polishing images, the defect polishing image is analyzed, and the obtained array is recorded as a rough water pattern array of the rough-processed valve plate;

[0100] Step S1062, obtaining rough water pattern arrays of all roughened valve sheets, storing the modes of all rough water pattern arrays into array β, and recording array β as the rough water pattern feature of valve sheet defect FX;

[0101] Step S1063, obtaining rough water pattern features of all valve sheet defects FX.

[0102] Step S2, performing feature screening processing on the rough water pattern feature based on the smooth water pattern feature, and obtaining a plurality of valve plate water pattern features based on the processing result;

[0103] Step S2 comprises: step S201, for any rough water mark feature of the valve defect FX, when the set corresponding to the valve defect FX is a subset of the set corresponding to the smooth water mark feature, the rough water mark feature of the valve defect FX is recorded as the valve water mark feature;

[0104] Step S202, when the set corresponding to the valve feature FX is not a subset of the set corresponding to the smooth water pattern feature, a set consisting of numbers in the set corresponding to the valve feature FX that are different from the set corresponding to the smooth water pattern feature is recorded as the valve water pattern feature of the valve feature FX;

[0105] Step S203, obtaining valve sheet water pattern features corresponding to all valve sheet defects FX.

[0106] Step S3, performing surface recognition detection when the valve plate is polished based on the valve plate water pattern characteristics; Step S3 includes:

[0107] Step S301, when the valve plate is polished, the polished valve plate is recorded as a real-time processed valve plate; based on the defect polishing process, a rough water mark feature corresponding to the real-time processed valve plate is obtained and recorded as a real-time water mark feature;

[0108] Step S302, when the set corresponding to the real-time water pattern feature is a subset of the set corresponding to the smooth water pattern feature, the valve sheet defect FX corresponding to the valve sheet water pattern feature identical to the real-time water pattern feature is recorded as a key defect of the real-time processed valve sheet;

[0109] Step S303, when the set corresponding to the real-time water pattern feature is not a subset of the set corresponding to the smooth water pattern feature, the valve sheet defects FX corresponding to the valve sheet water pattern features having the largest number of the same numbers as those in the set of the real-time water pattern features are recorded as the key defects of the valve sheet to be processed in real time;

[0110] Step S304: polishing the valve plate again based on the key defects.

[0111] Example 3, please refer to Figure 5 As shown, Figure 5 The structural schematic diagram of an electronic device is illustrated, and the electronic device may include: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor may call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in the valve plate polishing surface recognition and detection method based on machine vision are executed to achieve the following functions: first, the polished valve plate and the unpolished valve plate are obtained respectively, and are put into a polishing machine for polishing treatment respectively, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and the smooth water pattern feature and the rough water pattern feature are obtained based on the analysis result; then, the rough water pattern feature is subjected to feature screening processing based on the smooth water pattern feature, and multiple valve plate water pattern features are obtained based on the processing result; finally, the surface recognition and detection is performed when the valve plate is polished based on the valve plate water pattern feature.

[0112] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0113] Embodiment 4, the present application also provides a computer-readable storage medium, and the present application provides a storage medium on which a computer program is stored. When the computer program is executed by the processor, the steps in the valve plate polishing surface recognition and detection method based on machine vision are executed to achieve the following functions: first, the polished valve plate and the unpolished valve plate are respectively obtained, and are placed in a polishing machine for polishing treatment, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and smooth water ripple features and rough water ripple features are obtained based on the analysis results; then, the rough water ripple features are feature screened based on the smooth water ripple features, and multiple valve plate water ripple features are obtained based on the processing results; finally, surface recognition and detection are performed based on the valve plate water ripple features during valve plate polishing.

[0114] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0115] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A valve plate polishing surface recognition and detection method based on machine vision, characterized in that: The steps include: Obtaining the polished valve disc and the unpolished valve disc respectively, and putting them into a polishing machine for polishing, wherein the polishing includes smooth polishing and defect polishing; analyzing the polishing process based on machine vision, and obtaining smooth water pattern characteristics and rough water pattern characteristics based on the analysis results; Performing feature screening processing on rough water ripple features based on smooth water ripple features, and obtaining multiple valve sheet water ripple features based on the processing results; Surface recognition and detection are performed based on the valve plate water pattern characteristics during valve plate polishing; Obtaining the polished valve disc and the unpolished valve disc respectively, and placing them in a polishing machine for polishing treatment respectively includes: Obtain a polished valve disc and record it as a smooth valve disc; obtain multiple unpolished valve discs and record them as rough valve discs; obtain the defect types on the surfaces of all rough valve discs and record them as valve disc defect FX1 to valve disc defect FX2. c ; For any rough valve plate, the valve plate defects FX1 to FX2 on the rough valve plate surface are obtained based on AI recognition. c and record the valve sheet defect FX with the largest number as the representative defect of the rough valve sheet; obtain the representative defects of all rough valve sheets; The polishing process is analyzed based on machine vision, and the smooth water pattern characteristics and rough water pattern characteristics are obtained based on the analysis results, including: The maximum speed of the polishing machine when polishing the valve plate is recorded as the maximum polishing speed; the smooth valve plate is polished and polished; The smooth polishing process is as follows: put the smooth valve plate into a polishing machine with a maximum polishing speed, align the center of the camera from directly above the smooth valve plate to the center of the smooth valve plate based on machine vision and start recording; when the polishing time is a unit time, stop polishing and stop recording, and record the image taken at this time as a smooth polishing image, wherein the unit time is the time for polishing the valve plate once, and the center of the smooth valve plate coincides with the center of the sandpaper during polishing; Smooth polishing also includes: For image α corresponding to any frame except the first frame in the smooth polishing image, the water ripple area in image α is obtained based on machine vision, wherein the water ripple area is the area where the cooling water is located in the area other than the smooth valve plate; a plane rectangular coordinate system is established, which is recorded as the water ripple analysis coordinate system, wherein the units of the X-axis and the Y-axis of the water ripple analysis coordinate system are both cm; the water ripple area of ​​image α is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve plate, the midpoint of the smooth valve plate is marked in the first quadrant of the water ripple analysis coordinate system, which is recorded as the analysis calibration point; The image corresponding to the previous frame of the frame where the image α is located is recorded as the previous frame image; the water ripple area of ​​the previous frame image is obtained, and the water ripple area of ​​the previous frame image is grayed and placed in the first quadrant of the water ripple analysis coordinate system, and based on the positional relationship between the water ripple area and the smooth valve sheet, the midpoint of the smooth valve sheet corresponding to the water ripple area of ​​the previous frame image is made to coincide with the analysis calibration point; Smooth polishing also includes: The water ripple area of ​​the previous frame image is rotated for one circle with the analysis calibration point as the center, and the coincidence parameter of the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α during the rotation is obtained in real time, wherein the coincidence parameter is the value of the area of ​​the coincidence area between the water ripple area of ​​the previous frame image and the water ripple area of ​​the image α divided by the area of ​​the water ripple area of ​​the image α; The maximum overlap parameter obtained is recorded as the watermark similarity parameter of image α.

2. The valve plate polishing surface recognition and detection method based on machine vision according to claim 1 is characterized in that: Smooth polishing also includes: Obtaining water ripple similarity parameters of images corresponding to all frames except the first frame in the smooth polishing image, and recording the image corresponding to the largest water ripple similarity parameter as the smooth analysis image, wherein, when the largest water ripple similarity parameter corresponds to multiple images, the image corresponding to the smallest time in the smooth polishing image is recorded as the smooth analysis image; The grayscale image of the water ripple area corresponding to the smooth analysis image is obtained and recorded as a smooth grayscale image; the image formed by the edge of the smooth grayscale image is recorded as a smooth grayscale graph, and the pixel points in the smooth grayscale graph are binarized; the lengths of all black line segments in the smooth grayscale graph are stored in an array γ, and the array γ is recorded as a smooth water ripple feature.

3. The method for identifying and detecting the polished surface of a valve plate based on machine vision according to claim 2 is characterized in that: The polishing process is analyzed based on machine vision, and the smooth water pattern characteristics and rough water pattern characteristics are obtained based on the analysis results, including: For any valve plate defect FX, the rough valve plate representing the defect of the valve plate defect FX is recorded as a rough-processed valve plate, and all rough-processed valve plates are subjected to defect polishing treatment; The defect polishing process is as follows: for any rough-processed valve plate, based on the processing method of smooth polishing process, the rough-processed valve plate is placed in a polishing machine with a maximum polishing speed, and an image is captured based on machine vision, and the obtained image is recorded as a defect polishing image; based on a method of obtaining smooth water pattern features from smooth polishing images, the defect polishing image is analyzed, and the obtained array is recorded as a rough water pattern array of the rough-processed valve plate; Obtain the rough water pattern arrays of all roughened valve sheets, store the modes of all rough water pattern arrays into array β, and record array β as the rough water pattern feature of valve sheet defect FX; Get the rough water characteristics of all valve defect FX.

4. The method for identifying and detecting the polished surface of a valve plate based on machine vision according to claim 1, characterized in that: Based on the smooth water pattern feature, the rough water pattern feature is subjected to feature screening and processing, and multiple valve sheet water pattern features are obtained based on the processing result, including: For any rough water pattern feature of a valve defect FX, when the set corresponding to the valve defect FX is a subset of the set corresponding to the smooth water pattern feature, the rough water pattern feature of the valve defect FX is recorded as the valve water pattern feature; When the set corresponding to the valve feature FX is not a subset of the set corresponding to the smooth water pattern feature, the set consisting of numbers in the set corresponding to the valve feature FX that are different from the set corresponding to the smooth water pattern feature is recorded as the valve water pattern feature of the valve feature FX; Get the valve water pattern features corresponding to all valve defect FX.

5. The valve plate polishing surface recognition and detection method based on machine vision according to claim 1 is characterized in that: Surface recognition and detection based on valve plate water pattern characteristics during valve plate polishing include: When the valve plate is polished, the polished valve plate is recorded as a real-time processed valve plate; based on the defect polishing process, a rough water pattern feature corresponding to the real-time processed valve plate is obtained and recorded as a real-time water pattern feature; When the set corresponding to the real-time water pattern feature is a subset of the set corresponding to the smooth water pattern feature, the valve plate defect FX corresponding to the valve plate water pattern feature that is the same as the real-time water pattern feature is recorded as the key defect of the real-time processed valve plate; When the set corresponding to the real-time water pattern feature is not a subset of the set corresponding to the smooth water pattern feature, the valve sheet defects FX corresponding to the valve sheet water pattern features with the largest number of the same numbers as those in the set of real-time water pattern features are recorded as the key defects of the valve sheet to be processed in real time; The valve plate is polished again based on the key defects.

6. A valve plate polishing surface recognition and detection system based on machine vision, used to implement the valve plate polishing surface recognition and detection method based on machine vision according to any one of claims 1 to 5, characterized in that: It includes a polishing water pattern analysis module, a water pattern feature acquisition module and a polishing detection module; The polishing water pattern analysis module is used to obtain the polished valve disc and the unpolished valve disc respectively, and put them into the polishing machine for polishing treatment respectively, wherein the polishing treatment includes smooth polishing treatment and defect polishing treatment; the polishing treatment process is analyzed based on machine vision, and the smooth water pattern characteristics and the rough water pattern characteristics are obtained based on the analysis results; The water pattern feature acquisition module is used to perform feature screening processing on the rough water pattern feature based on the smooth water pattern feature, and acquire multiple valve plate water pattern features based on the processing result; The polishing detection module is used to perform surface recognition detection when the valve plate is polished based on the water pattern characteristics of the valve plate.

Images