Developing roller image defect detection method, electronic equipment and storage medium

Abstract

The invention relates to the technical field of developing roller defect detection, in particular to a developing roller image defect detection method, electronic equipment and a storage medium, and the method comprises the following steps: equally dividing the outer surface layer of a to-be-detected developing roller into N to-be-detected areas; respectively acquiring images of the N to-be-detected areas to obtain N first area images; sequentially preprocessing the N first area images to obtain N second area images; respectively calculating the similarity between the N second area images and a preset standard image to obtain N similarity coefficients; comparing the N similarity coefficients with a preset similarity threshold value; and if the N similarity coefficients are smaller than a first preset threshold value, determining that the to-be-detected developing roller is a defective developing roller. According to the scheme provided by the invention, the to-be-detected developing roller can be automatically detected and screened, and the detection efficiency and the detection accuracy of the developing roller are improved.

Description

Technical Field

[0001] The present application relates to the technical field of developing roller defect detection, and in particular to a developing roller image defect detection method, electronic equipment and storage medium. Background Art

[0002] The developing roller is a component in a laser printer or copier. Its main function is to transfer toner from the toner cartridge to the photosensitive drum, and transfer the toner to the photosensitive drum through static electricity to form a pattern of text or image. Finally, these patterns are transferred to paper through the photosensitive drum and fixed by heating to form a stable print output. Usually, the developing roller is often made of conductive rubber or metal, and its surface is coated with a layer of chemical coating material that can absorb toner. The material is usually resistant to high temperature (usually around 180 degrees). The coating part is generally black and shiny, and the two ends are relatively thin, which is used to fix and install the drive gear. The basic structure of the developing roller is that the outer side of the metal shaft core is coated with an elastomer. When the developing roller rotates, the chemical coating material will contact the toner and carry a thin layer of toner.

[0003] At present, most of the existing developing rollers are coated with a conductive elastomer on the outside of the metal shaft core, and the rubber as the conductive elastomer usually also contains a conductive agent. In the production and molding process of the developing roller, the metal core shaft needs to be degreased and derusted with water, and assembled on the preheated rubber roller molding mold after coating with an adhesion promoter, and then the rubber roller molding mold is placed in a heating furnace for preheating, and then the outer barrel of liquid polyurethane material (polyurethane prepolymer, polyurethane polyol, etc.) is preheated in an oven and poured into a casting machine, poured into the preheated rubber roller mold, and then heated and composited by a molding machine and demolded, and then heated and molded for the second time in a heating furnace, and the polyurethane rubber stick is cut and ground to produce a preliminary developing roller.

[0004] However, during the production process of the developing roller, after going through the above-mentioned multiple processing techniques, since each process adopts a different process and there is a connection gap between each process, various defects often appear on the outer layer of the developing roller, which affects the direct use effect of the developing roller; in addition, since the developing roller is mainly affected by the outer layer of the developing roller when in use, if there are defects in the outer layer, it will inevitably cause the printed or copied paper to be unclear or defective; therefore, after the developing roller is completed, it is necessary to inspect the developing roller and remove the defective developing rollers. In the current process of inspecting the outer layer of the developing roller, it is often inspected manually, which reduces the inspection efficiency. At the same time, long-term manual screening can easily lead to visual fatigue, which in turn causes unqualified developing rollers to flow into the finished product warehouse, causing defective products to be carried in the finished product developing rollers, thereby affecting the quality of acceptance.

[0005] Therefore, how to intelligently detect defects on the developing roller is a technical problem that technicians currently need to solve. Summary of the invention

[0006] In order to overcome the problems existing in the related art, the present application provides a developing roller image defect detection method, electronic equipment and storage medium, which can realize automatic detection and screening of developing rollers to be inspected, improve the detection efficiency and detection accuracy of developing rollers, prevent unqualified developing rollers from flowing into the finished product warehouse, and reduce labor costs and overcome the problem that manual screening can easily cause visual fatigue.

[0007] The first aspect of the present application provides a developing roller image defect detection method, comprising:

[0008] The outer surface of the developer roller to be inspected is equally divided into N inspected areas, wherein the inspected areas are strip-shaped areas, and the N inspected areas are rotationally symmetric about the axis of the developer roller to be inspected, and N is an integer greater than 1;

[0009] Respectively acquiring N images of the area to be inspected to obtain N first area images, wherein the first area image is a strip-shaped image corresponding to one side surface of the developing roller to be inspected;

[0010] Preprocessing the N first region images in sequence to obtain N second region images, wherein the second region images are elastic body images that only retain a portion of the metal shaft core and the strip region;

[0011] Calculating the similarities between the N second area images and a preset standard image respectively to obtain N similarity coefficients, wherein the preset standard image is a strip area image corresponding to any side surface of the non-defective developing roller;

[0012] Compare the N similarity coefficients with the preset similarity thresholds respectively;

[0013] If any of the N similarity coefficients is less than a first preset threshold, it is determined that the developing roller to be inspected is a defective developing roller;

[0014] If the N similarity coefficients are all greater than or equal to the first preset threshold, it is determined that the developing roller to be inspected is a non-defective developing roller.

[0015] In one implementation method, respectively acquiring N images of the area to be inspected includes:

[0016] The image sensor is directed toward the developing roller to be inspected at a first preset angle, and the image sensor is used to obtain an image of a part of the developing roller to be inspected in the inspected area. The first preset angle is the angle formed by the plane where the axis of the developing roller to be inspected is located and the axis of the developing roller to be inspected that the image sensor directly faces;

[0017] During a first time period, the developing roller to be inspected is rotated counterclockwise or clockwise at a second preset angle every second time period, and an image of the developing roller to be inspected is acquired during the second time period; wherein the second preset angle is the angle of rotation of the developing roller to be inspected, the ratio of the first time period to the second time period is equal to N, and the product of the second preset angle and N is equal to 360°.

[0018] In one implementation method, the preprocessing includes: binarization processing and segmentation processing;

[0019] The binarization process is specifically as follows: setting the grayscale value of the pixel point of the first area image to 0 or 255, wherein the image of the developing roller to be inspected in the non-first area image is in a blank state;

[0020] The segmentation process is specifically as follows: segmenting the blank part of the first region image after binarization to obtain an image containing only the appearance shape of the developing roller to be inspected.

[0021] In one implementation method, after determining that the developing roller to be inspected is a defective developing roller, the method includes:

[0022] Recording the number of N similarity coefficients less than a preset similarity threshold, and marking the second region images less than the preset similarity threshold, to obtain M third region images, the third region images being defective strip region images, where M is an integer greater than 0, and M is less than or equal to N;

[0023] Subtracting the M third region images from the preset standard image respectively to obtain M fourth region images, wherein the fourth region images are images containing only the defective shape of the developing roller;

[0024] When M is equal to 1, the fourth region image is paired with each element of the defect image set in the database one by one, and the defect type of the developing roller to be inspected is determined according to the similarity of the pairing results;

[0025] When M is greater than 1 and less than or equal to N, the defect type of the developing roller to be inspected is determined according to the defect distribution of the fourth region image.

[0026] In an implementation method, in the one-to-one pairing of the fourth region image with each element of the defect image set in the database, the defect image set includes at least one of the following defect images: bubble defect image, coil defect image, port defect image, adhesive surface defect image, chatter mark defect image, dot mark defect image, printed block defect image, line defect image, and yin-yang defect image;

[0027] Among them, each element in the defect image set is an image that only contains the defect shape of the corresponding developing roller, and each element in the defect image set is obtained by inputting the defective developing roller image into the convolutional neural network model for training.

[0028] In an implementation method, in the one-to-one pairing of the fourth region image with each element of the defect image set in the database, determining the defect type of the developing roller to be inspected according to the similarity of the pairing results includes:

[0029] Performing a similarity comparison between the fourth region image and each element of the defect image set in the database, wherein the similarity comparison includes image contour comparison and pixel number comparison;

[0030] If the similarity between the image contour of the fourth area image and the image contour of the defect image in the defect image set is greater than a second preset threshold, and the difference between the number of pixels of the fourth area image and the number of pixels of the defect image is within a preset range value, then the fourth area image is determined to be the defect type corresponding to the defect image.

[0031] In an implementation method, determining the defect type of the developing roller to be inspected according to the defect distribution of the fourth region image includes:

[0032] Taking one end point of the developing roller to be inspected as the origin, a plane rectangular coordinate system is established;

[0033] sequentially inputting the M fourth region images into the plane rectangular coordinate system, and generating a pixel coordinate based on each pixel in each of the fourth region images, thereby obtaining M pixel coordinate sets, wherein one of the pixel coordinate sets includes all pixel coordinates in one of the fourth region images;

[0034] Performing separate position distribution analysis on the M pixel coordinate sets respectively, and determining defect type information of the M fourth region images respectively, the defect type information including type A defect, type B defect, and type C defect;

[0035] If there are one or more images of type C defect in the M fourth area images, it is determined that the developing roller to be inspected is a developing roller with type C defects, and a type C defect detection report is generated;

[0036] If none of the M fourth area images has an image of defect type C, and the number of developing rollers determined to be defective type B is greater than or equal to the number of developing rollers determined to be defective type A, the developing roller to be inspected is determined to be a developing roller with defective type B, and a defect detection report for type B is generated;

[0037] If none of the M fourth area images contain images of defect type C, and the number of developing rollers determined to be defective type B is less than the number of developing rollers determined to be defective type A, then the developing roller to be inspected is determined to be a defective developing roller of type A, and a defect detection report of type A is generated.

[0038] In an implementation method, performing separate position distribution analysis on the M pixel coordinate sets to determine defect type information of the M fourth region images respectively includes:

[0039] If all pixel coordinates in the pixel coordinate set are gathered within the same first unit area, the fourth area image is determined to be a type A defect, and the type A defect includes at least one of the following defects: bubble defect, point mark defect, and block defect;

[0040] If the number of pixel coordinates in the pixel coordinate set at the same horizontal coordinate or vertical coordinate reaches a third preset threshold, the fourth region image is determined to be a type of defect of Class B, and the type of defect of Class B includes at least one of the following defects: a chatter mark defect, a line defect, a coil defect, and a yin-yang defect;

[0041] If all pixel coordinates in the pixel coordinate set are gathered within the same second unit area range, the fourth area image is determined to be a Class C defect type, and the Class C defect type includes at least one of the following defects: port defects and adhesive surface defects; wherein the second unit area is larger than the first unit area range.

[0042] A second aspect of the present application provides an electronic device, including:

[0043] Processor; and

[0044] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method as described above.

[0045] A third aspect of the present application provides a non-transitory machine-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method as described above.

[0046] The technical solution provided by this application may have the following beneficial effects:

[0047] In the technical solution, the outer surface of the developer roller to be inspected is equally divided into N inspected areas, so that the N inspected areas are all rotationally symmetrical with the axis of the developer roller to be inspected, and then the images of the N inspected areas are respectively acquired to obtain N first area images, and the N first area images are pre-processed in turn to remove image contents irrelevant to the developer roller to be inspected, and N second area images are obtained, so that the second area images become elastic body images that only retain part of the metal shaft core and the strip area, and then the similarities of the N second area images and the preset standard image are respectively calculated to obtain N similarity coefficients, and the N first area images are respectively compared to obtain N second area images. The size of the similarity coefficient is equal to the preset similarity threshold. When there is a situation that one of the N similarity coefficients is less than the first preset threshold, the developer roller to be inspected is determined to be a defective developer roller, and the developer roller to be inspected is marked as an unqualified product; when the N similarity coefficients are all greater than or equal to the first preset threshold, the developer roller to be inspected is determined to be a non-defective developer roller, and the developer roller to be inspected is marked as a qualified product; thereby realizing automated detection and screening of the developer roller to be inspected, improving the detection efficiency and detection accuracy of the developer roller, preventing unqualified developer rollers from flowing into the finished product warehouse, reducing labor costs, and overcoming the problem that manual screening is also prone to visual fatigue.

[0048] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The above and other objects, features and advantages of the present application will become more apparent through a more detailed description of exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0050] Figure 1 is a schematic flow chart of a developing roller image defect detection method shown in an embodiment of the present application;

[0051] Figure 2 is a schematic diagram of a developing roller shown in an embodiment of the present application;

[0052] Figure 3 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The preferred embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0054] The technical solution of the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0055] Embodiment 1

[0056] See also Figure 1 and Figure 2 , Figure 1 FIG. 1 is a flow chart of a developing roller image defect detection method. Figure 2 It is a schematic diagram of the developing roller. In this example, since the developing roller is made of conductive rubber or metal and its surface is coated with a chemical coating material that can absorb carbon powder, the coating part is generally black and shiny, and the two ends are relatively thin, which are used to fix and install the driving gear; at the same time, the basic structure of the developing roller is that the outside of the metal shaft core is coated with an elastomer. Therefore, when the developing roller is used, it is necessary to ensure that the surface of the developing roller has no defects (that is, the surface of the coating part has no defects) to ensure that the paper printed by the printer or copier is complete and not missing.

[0057] However, after the development roller is produced, it still needs to be inspected and screened to ensure the quality of the finished product of the development roller and prevent unqualified products from flowing into the customer; and for the existing development roller detection method, the traditional method of manually inspecting and screening the development roller cannot improve efficiency, and is also very prone to errors. For this reason, this example provides a method for detecting image defects of a development roller to improve the efficiency and accuracy of detection; specifically, based on the product structural characteristics of the development roller (its appearance is cylindrical, and there is metal at both ends), this example divides the outer layer of the development roller to be inspected into N areas to be inspected, wherein the areas to be inspected are strip areas, and by dividing the development roller into detection areas, the N areas to be inspected are rotationally symmetrical with the axis of the development roller to be inspected, so that each position of the entire development roller can be inspected, thereby obtaining an image of the entire development roller, and realizing comprehensive detection of each position of the development roller.

[0058] After the outer layer of the developer roller to be inspected is divided into N areas to be inspected, this example obtains images of the N areas to be inspected respectively, and arranges the N areas to be inspected in turn to obtain N first area images, wherein the N first area images are arranged and combined to obtain a plane image of the entire developer roller to be inspected, thereby obtaining a plane image of the entire developer roller to be inspected, and the first area image refers to a strip-shaped image corresponding to one side of the developer roller to be inspected, for example: the developer roller to be inspected is divided into 4 areas to be inspected, each area to be inspected obtained is a strip-shaped area corresponding to the developer roller to be inspected, and the images obtained from the 4 areas to be inspected are combined to obtain a plane image of the entire developer roller to be inspected; however, this example can directly determine whether the developer roller to be inspected is a defective developer roller by separately judging whether the 4 areas to be inspected have defects, thereby improving the detection efficiency.

[0059] Among them, since the developing roller to be inspected has a cylindrical shape, in this example, the outer layer of the developing roller to be inspected can also be equally divided into 5, 6, 9 and 12 areas to be inspected, wherein the higher the score of the equal division, the clearer the image obtained for inspection. Since the outer layer of the developing roller to be inspected is cylindrical, when obtaining the image by plane photography, the fewer the number of equal divisions, the more likely it is that the obtained image will be incomplete, especially when photographing the positions on both sides and both ends of the area to be inspected. Due to the black appearance of the developing roller and the light being easily absorbed by the black outer layer of the developing roller, the image of this part is easily unclear, and then the subsequent defect judgment will make it impossible to comprehensively and accurately identify and judge the defects of the developing roller to be inspected.

[0060] In this example, after obtaining the images of various parts of the developer roller to be inspected, the N first area images are analyzed and judged separately, thereby replacing the existing method of "inspecting the developer roller one by one by workers individually", which can greatly improve the inspection efficiency of the developer roller and improve the inspection accuracy; for example, before the N first area images are inspected, this example uses image processing to pre-process the N first area images in turn, remove image content irrelevant to the developer roller body, and obtain N second area images, wherein the second area image is an elastic body image that only retains part of the metal shaft core and the strip area, and eliminates the influence of other irrelevant image content, thereby improving accuracy; then calculate the similarity between the N second area images and the preset standard image, and then obtain N similarity coefficients, wherein the preset standard image uses any of the defective developer rollers. The image of the strip area corresponding to the side surface is, for example, randomly selecting an area to be inspected with the same number of equal divisions as in this example on a defect-free developing roller as a preset standard image, and clearing the image content irrelevant to the developing roller body, and calculating the similarity coefficient with the second area image, and then setting a preset similarity threshold, and comparing the above N similarity coefficients with the preset similarity threshold respectively. When the N similarity coefficients are less than the first preset threshold, it means that the developing roller to be inspected has defects, thereby directly determining that the developing roller to be inspected is a defective developing roller, realizing rapid detection and improving the detection efficiency of the developing roller; at the same time, when the N similarity coefficients are all greater than or equal to the first preset threshold, it means that each area to be inspected of the developing roller to be inspected has no defects, thereby directly determining that the developing roller to be inspected is a defect-free developing roller, thereby realizing automated detection of the developing roller and improving the detection efficiency of the developing roller; such as:

[0061] In this example, the outer layer of the developer roller to be inspected is equally divided into 6 inspected areas, and the 6 inspected areas are rotationally symmetrical with the axis of the developer roller to be inspected. Then, images of the 6 inspected areas are acquired respectively to obtain 6 first area images. The 6 first area images are preprocessed in turn to remove image contents irrelevant to the developer roller to be inspected, and 6 second area images are obtained, so that the 6 second area images are all elastic body images with only a part of the metal shaft core and the strip area retained. Then, the similarity between the 6 second area images and the preset standard image is calculated respectively to obtain 6 similarity coefficients. The 6 similarity coefficients are compared with the preset similarity thresholds respectively. When one of the 6 similarity coefficients is negative, the similarity coefficient is calculated. When the similarity coefficients are less than the first preset threshold, it means that the developing roller to be inspected corresponding to the second area image has defects, and then the developing roller to be inspected can be determined to be a defective developing roller, and the developing roller to be inspected is marked as an unqualified product; and when the 6 similarity coefficients are all greater than or equal to the first preset threshold, it means that the developing rollers to be inspected corresponding to the 6 second area images have no defects, and then the developing roller to be inspected is determined to be a non-defective developing roller, and the developing roller to be inspected is marked as a qualified product; thereby realizing automated detection and screening of the developing rollers to be inspected, improving the detection efficiency and detection accuracy of the developing rollers, preventing unqualified developing rollers from flowing into the finished product warehouse, reducing labor costs, and overcoming the problem that manual screening is also prone to visual fatigue.

[0062] Embodiment 2

[0063] In this embodiment, after the outer layer of the developer roller to be inspected is equally divided into N areas to be inspected, since the developer roller has a cylindrical shape, after the equal division, it is necessary to accurately photograph the N areas to be inspected to obtain images corresponding to the N areas to be inspected (i.e., first area images). To this end, when acquiring the image of the developer roller to be inspected in this example, a first preset angle is set, and the first preset angle is used as a reference, and the shooting surface of the image sensor is facing the developer roller to be inspected, and the image of the area to be inspected of a part of the developer roller to be inspected is obtained through the image sensor, wherein the first preset angle is actually the angle formed by the plane where the axis of the developer roller to be inspected is located and the axis of the developer roller to be inspected directly facing the image sensor; under the design of this scheme, the image sensor actually takes the first preset angle as a constant angle, and always corresponds to the developer roller to be inspected, to ensure that the image sensor does not change position when shooting the image, and to ensure that the image captured each time is of the same size, so as to facilitate subsequent judgment and recognition.

[0064] It is worth noting that, in order to improve the efficiency of detection and ensure that each equally divided area to be inspected of the developer roller to be inspected can be accurately obtained each time, this example sets a constant time length (i.e., the first time length) to ensure that the image acquisition time for each developer roller to be inspected is consistent, so as to improve the efficiency of image acquisition, and within the first time length, the developer roller to be inspected is rotated counterclockwise or clockwise at a second preset angle every second time length, and within the second time length, the image of the developer roller to be inspected is acquired by the image sensor, thereby acquiring images of N areas to be inspected; it is worth noting that, in order to ensure that the images acquired each time do not overlap or are not missed, The second preset angle in this example is actually the angle of rotation of the developing roller to be inspected. After the outer layer of the developing roller to be inspected is equally divided into N areas to be inspected, the size of the second preset angle can be directly obtained by calculation. It should be noted that since the developing roller is cylindrical, when the developing roller is rotated, the size of its second preset angle is the fraction of the outer layer of the developing roller to be inspected. For example, the ratio of the first time length to the second time length is equal to N, and the product of the second preset angle and N is equal to 360°. If the outer layer of the developing roller to be inspected is equally divided into 6 parts, the degree of each rotation of the second preset angle is 60°.

[0065] Therefore, through the above-mentioned technical scheme, this example can effectively divide the developing roller to be inspected into N parts, and when obtaining images of the N areas to be inspected respectively, the developing roller to be inspected can be accurately divided into equal parts, so that the images of the areas to be inspected obtained are accurate, and the images obtained are effectively prevented from being overlapping or missing, thereby improving the detection accuracy of the developing roller to be inspected and greatly improving the detection efficiency.

[0066] Embodiment 3

[0067] In this example, after obtaining the first area image, in order to eliminate other irrelevant image contents in the first area image (i.e., only retain the developing roller), this example also requires preprocessing of the first area image to ensure that the image of the area to be inspected is not affected by other pixels, thereby ensuring image accuracy; wherein, the preprocessing includes: binarization processing and segmentation processing, wherein the binarization processing is specifically: setting the grayscale value of the pixel points of the first area image to 0 or 255, eliminating the pixels of the non-developing roller body in the first area image, so that the image of the developing roller to be inspected in the non-first area image is blank, and performing segmentation processing at the same time, wherein the segmentation processing is specifically: segmenting the blank part of the first area image after binarization to obtain an image containing only the appearance shape of the developing roller to be inspected, thereby obtaining an image containing only the developing roller body.

[0068] Among them, binarization processing refers to dividing the pixels in the image into two categories by setting a threshold: one category is the pixels with grayscale values ​​greater than or equal to the threshold, these points are set to white (grayscale value is 255); the other category is the pixels with grayscale values ​​less than the threshold, these points are set to black (grayscale value is 0). In this way, the image is simplified into a binary image with only black and white colors, so that the entire image presents an obvious black and white effect; and the segmentation processing is to divide the image into several specific areas with unique properties; the image segmentation method can be processed using any of the following methods: threshold-based segmentation method, region-based segmentation method, edge-based segmentation method, etc.

[0069] Embodiment 4

[0070] In this example, after determining that the developing roller to be inspected is a defective developing roller, in order to find out the cause of the defect in the developing roller and determine the defect type of the developing roller to be inspected, so as to facilitate the subsequent classification of unqualified developing rollers, this example needs to further analyze the defective developing roller, by determining the defect type of the developing roller, and then finding out the cause of the defect based on the defect type of the developing roller, thereby improving the production method of the developing roller and reducing the defective rate of the developing roller production; in addition, due to the particularity of the developing roller, when identifying the defect type, it is necessary to specifically determine the defect shape and type of the developing roller, so as to ensure that the same defect will not occur in the subsequent production of developing rollers.

[0071] Specifically, after determining that the developing roller to be inspected is a defective developing roller, this example records the number of N similarity coefficients that are less than a preset similarity threshold, marks the second area images that are less than the preset similarity threshold, and selects the marked second area images, and rearranges and measures them to obtain M third area images; wherein the third area image is a defective strip area image, M is an integer greater than 0, and M is less than or equal to N; the defect type of the developing roller is determined by extracting the defective second area image and then performing separate defect analysis on the second area image.

[0072] After obtaining M third area images, this example obtains M fourth area images by subtracting the M third area images from the preset standard image, wherein the fourth area image is an image containing only the defect shape of the developing roller; then, according to the number of fourth area images, different defect recognition methods are selected to determine the defect type of the developing roller as quickly as possible: for example, when M is equal to 1, it indicates that after the developing roller is equally divided, there is only one defect, and there are no defects in other positions; therefore, the type of the developing roller can be quickly determined by comparison, and the cause of the corresponding defect can also be easily determined by technicians. To improve efficiency, this example compares the fourth area image with each of the defect image sets in the database. The elements are matched one by one, and the defect type of the developing roller to be inspected is determined according to the similarity of the matching results, so as to directly determine the defect type of the developing roller, and then the developing roller is classified according to the defect type of the developing roller, and the existing production process is improved according to the defect type of the developing roller; in addition, when the M is greater than 1 and less than or equal to the N, it means that the developing roller has multiple defects after being equally divided, which means that there are huge hidden dangers in the existing developing roller production process, and it is necessary to perform separate defect identification on each fourth area image respectively, so as to determine the position and shape of the developing roller defect; for this reason, for this case, this example determines the defect type of the developing roller to be inspected according to the defect distribution of the fourth area image.

[0073] It should be noted that, in this example, by matching the fourth area image with each element of the defect image set in the database one by one, the defect image set includes at least one of the following defect images: bubble defect image, coil defect image, port defect image, rubber surface defect image, chatter mark defect image, dot mark defect image, block defect image, line defect image and yin-yang defect image; in the defect image set, the defect images included are defect shapes that are often encountered in the production of the developing roller, and are also common defects in the production process of the developing roller. By directly comparing the above-mentioned defect images, the type of the defect can be quickly determined; wherein ... Each element of is an image that only contains the defect shape of the corresponding developing roller, and each element in the defect image set is obtained by inputting the defective developing roller image into the convolutional neural network model for training; since the defect images in the defect image set are defect images that often appear in the production process of the developing roller, in the previous production process, if it is determined that the developing roller has a defect, the staff will take a picture of the defective position of the developing roller, obtain the defective image of the developing roller, and input the defective image into the convolutional neural network model. With multiple training and accumulation, a comparison database of common defects is obtained, and the defect type of the developing roller can be determined by direct comparison.

[0074] For example, in this example, when the fourth region image is paired one by one with each element of the defect image set in the database, the defect type of the developing roller to be inspected is determined according to the similarity of the pairing results, specifically, the steps are as follows:

[0075] In this example, the fourth area image is compared with each element of the defect image set in the database for similarity, wherein the similarity comparison includes image contour comparison and pixel number comparison; since common developing roller defects all have commonalities, and after training with the convolutional neural network model, the compared standard defect image can be well matched with the current developing roller defect, therefore, in this example, when the similarity between the image contour of the fourth area image and the image contour of the defect image in the defect image set is greater than the second preset threshold, and the difference between the number of pixels of the fourth area image and the number of pixels of the defect image is also within the preset range value, then the fourth area image can be directly determined to be the defect type corresponding to the defect image; for example, assuming that the similarity between the image contour of the fourth area image and the spot defect image in the defect image set is greater than the second preset threshold, and the difference between the number of pixels of the fourth area image and the number of pixels of the spot defect image is also within the preset range value, then the developing roller corresponding to the fourth area image can be directly determined to be a spot defect developing roller, and similarly, the comparison with other defect images in the defect image set is also the same.

[0076] It is worth noting that if the similarity between the image contour of the fourth area image and the image contour of the defective image in the defective image set is not greater than the second preset threshold, even if the difference between the number of pixels of the fourth area image and the number of pixels of the defective image is within the preset range, the defect type of the developing roller cannot be directly determined; similarly, even if the similarity between the image contour of the fourth area image and the image contour of the defective image in the defective image set is greater than the second preset threshold, but the difference between the number of pixels of the fourth area image and the number of pixels of the defective image is not within the preset range, the defect type of the developing roller cannot be directly determined; in addition, when the above situation occurs, it means that the existing production process of the developing roller is likely to have a new defect type, and staff should be arranged to investigate to avoid equipment failure.

[0077] In addition, after comparing and determining the defects of the developing roller to be inspected, for some defect types in the defect image set, the defective developing roller can be turned into a non-defective developing roller through manual repair. Please refer to Example 5 for specific technical content.

[0078] Embodiment 5

[0079] In this example, when M is greater than 1 and less than or equal to N, it means that the developing roller has multiple defects after being equally divided, and there are multiple possibilities of defects. Therefore, it is necessary to perform separate defect identification on each fourth region image respectively, so as to find out the location of the developing roller where the defect is located, and generate a defect detection report, so that the technical personnel can check the production equipment and repair or scrap the defective developing roller. Specifically, determining the defect type of the developing roller to be inspected according to the defect distribution of the fourth region image includes the following steps:

[0080] When determining that there are defects in multiple positions of the developer roller to be inspected, this example establishes a plane rectangular coordinate system by taking one end point of the developer roller to be inspected as the origin; M fourth area images are sequentially input into the plane rectangular coordinate system, and a pixel coordinate is generated based on each pixel in each fourth area image, and M pixel coordinate sets are obtained accordingly, wherein one pixel coordinate set corresponds to all pixel coordinates in one fourth area image; so that each fourth area image can specifically determine the position of the defect, and analyze the type and shape of the defect at the same time, such as: by performing separate position distribution analysis on the M pixel coordinate sets, the defect type information of the M fourth area images is determined respectively, and the defects are classified to facilitate staff to check. In this example, the defect type information is divided into Class A defect type, Class B defect type and Class C defect type, wherein Class C defect type is a defect type with a relatively large defect area, which represents a type with high difficulty in repair; Class B defect type is second, and Class A defect type is a small area defect, which represents a relatively simple repair.

[0081] For example, when there are one or more images of type C defect in the M fourth area images, it indicates that the developing roller has large-area defects and cannot be repaired, and can only be treated as scrapped. Therefore, it is determined that the developing roller to be inspected is a type C defective developing roller, and a type C defect detection report is generated;

[0082] When none of the M fourth area images has an image of defect type C, and the number of developing rollers determined to be defective type B is greater than or equal to the number of developing rollers determined to be defective type A, it indicates that the defect of the developing roller is a local area defect, and it is necessary for workers to evaluate it before determining whether it can be repaired. Therefore, it is determined that the developing roller to be inspected is a developing roller with defect type B, and a defect detection report of type B is generated to facilitate the staff to repair it;

[0083] When there are no images of Class C defect type in the M fourth area images, and the number of developing rollers determined to be Class B defects is less than the number of developing rollers determined to be Class A defects, it means that the defects of the developing roller are scattered dots, depressions, convex spots and the like, which are not obvious and are easy to repair, and can become a qualified product through reprocessing. Therefore, the developing roller to be inspected is determined to be a Class A defective developing roller, and a Class A defect detection report is generated.

[0084] It is worth noting that in this example, when performing separate position distribution analysis on the M pixel coordinate sets and determining the defect type information of the M fourth region images respectively, in order to determine the type of defect of the developing roller, this example further analyzes and judges, such as:

[0085] When all pixel coordinates in the pixel coordinate set are gathered within the same first unit area, it is determined that the fourth area image is of type A defect, and the type A defect includes at least one of the following defects: bubble defect, dot defect and block defect, wherein the type A defect can be repaired manually to make the developing roller a qualified product;

[0086] When the number of pixel coordinates in the pixel coordinate set at the same horizontal coordinate or vertical coordinate reaches a third preset threshold, it is determined that the fourth region image is a type of defect of Class B, and the type of defect of Class B includes at least one of the following defects: chattering defect, line defect, coil defect and yin-yang defect, wherein the type of defect of Class B requires workers to analyze and evaluate, and after evaluation, they can be modified to make the developing roller a qualified product;

[0087] When all pixel coordinates in the pixel coordinate set are gathered within the same second unit area range, the fourth area image is determined to be a Class C defect type, and the Class C defect type includes at least one of the following defects: port defects and glue surface defects; wherein, the second unit area is larger than the first unit area range, and the Class C defect type is a defect that cannot be modified and can only be treated as scrap.

[0088] Embodiment 6

[0089] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides an electronic device and corresponding embodiments.

[0090] See also Figure 3 , the electronic device 1000 includes a memory 1010 and a processor 1020 .

[0091] The processor 1020 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0092] The memory 1010 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage devices. Among them, ROM can store static data or instructions required by the processor 1020 or other modules of the computer. The permanent storage device may be a readable and writable storage device. The permanent storage device may be a non-volatile storage device that does not lose the stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory may store some or all instructions and data required by the processor at run time. In addition, the memory 1010 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 1010 may include a readable and / or writable removable storage device, such as a laser disc (CD), a read-only digital versatile disc (such as a DVD-ROM, a double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a mini SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer-readable storage medium does not include carrier waves and transient electronic signals transmitted wirelessly or wired.

[0093] The memory 1010 stores executable codes, and when the executable codes are processed by the processor 1020 , the processor 1020 can execute part or all of the methods described above.

[0094] The scheme of the present application has been described in detail above with reference to the accompanying drawings. In the above embodiments, the description of each embodiment has its own emphasis. For the part that is not described in detail in a certain embodiment, refer to the relevant description of other embodiments. Those skilled in the art should also know that the actions and modules involved in the description are not necessarily required for the present application. In addition, it can be understood that the steps in the method of the embodiment of the present application can be adjusted in order, merged and deleted according to actual needs, and the modules in the device of the embodiment of the present application can be merged, divided and deleted according to actual needs.

[0095] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0096] Alternatively, the present application can also be implemented as a non-temporary machine-readable storage medium (or computer-readable storage medium, or machine-readable storage medium) on which executable code (or computer program, or computer instruction code) is stored. When the executable code (or computer program, or computer instruction code) is executed by a processor of an electronic device (or electronic device, server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0097] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the application herein may be implemented as electronic hardware, computer software, or combinations of both.

[0098] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system and method according to multiple embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a part of a module, a program segment or a code, and the part of the module, the program segment or the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or the flow chart, and the combination of the square boxes in the block diagram and / or the flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0099] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A developing roller image defect detection method, characterized in that: include: The outer surface of the developer roller to be inspected is equally divided into N areas to be inspected, wherein the areas to be inspected are strip-shaped areas, and the N areas to be inspected are rotationally symmetric about the axis of the developer roller to be inspected, and N is an integer greater than 1; Respectively acquiring images of N areas to be inspected to obtain N first area images, wherein the first area image is a strip-shaped image corresponding to one side surface of the developing roller to be inspected; Preprocessing the N first region images in sequence to obtain N second region images, where the second region images are images of the elastic body that only retain a portion of the metal shaft core and the strip region; Calculating the similarities between N second region images and a preset standard image respectively to obtain N similarity coefficients, wherein the preset standard image is a strip region image corresponding to any side surface of a non-defective developing roller; Comparing the N similarity coefficients with the preset similarity thresholds respectively; If any of the N similarity coefficients is less than a first preset threshold, it is determined that the developing roller to be inspected is a defective developing roller; If the N similarity coefficients are all greater than or equal to the first preset threshold, it is determined that the developing roller to be inspected is a non-defective developing roller.

2. The developing roller image defect detection method according to claim 1, characterized in that: The step of respectively acquiring images of the N areas to be inspected comprises: The image sensor is placed at a first preset angle facing the developing roller to be inspected, and the image sensor is used to obtain an image of a part of the developing roller to be inspected in the inspected area. The first preset angle is the angle formed by the plane where the axis of the developing roller to be inspected is located and the axis of the developing roller to be inspected that the image sensor directly faces; During a first time period, the developing roller to be inspected is rotated counterclockwise or clockwise at a second preset angle every second time period, and an image of the developing roller to be inspected is acquired during the second time period; wherein, the second preset angle is the angle of rotation of the developing roller to be inspected, the ratio of the first time period to the second time period is equal to N, and the product of the second preset angle and N is equal to 360°.

3. The developing roller image defect detection method according to claim 1, characterized in that: The preprocessing includes: binarization processing and segmentation processing; The binarization process specifically includes: setting the grayscale value of the pixel points of the first area image to 0 or 255, wherein the image of the developing roller to be inspected in the image other than the first area image is in a blank state; The segmentation process specifically includes segmenting the blank part of the first region image after binarization to obtain an image containing only the appearance shape of the developing roller to be inspected.

4. The developing roller image defect detection method according to claim 1, characterized in that: After determining that the developing roller to be inspected is a defective developing roller, the method comprises: Recording the number of N similarity coefficients less than a preset similarity threshold, and marking the second region images less than the preset similarity threshold, to obtain M third region images, wherein the third region images are defective strip region images, wherein M is an integer greater than 0, and M is less than or equal to N; Subtracting the M third region images from the preset standard image respectively to obtain M fourth region images, wherein the fourth region images are images containing only the defective shape of the developing roller; Wherein, when the M is equal to 1, the fourth region image is paired with each element of the defect image set in the database one by one, and the defect type of the developing roller to be inspected is determined according to the similarity of the pairing results; When M is greater than 1 and less than or equal to N, the defect type of the developing roller to be inspected is determined according to the defect distribution of the fourth region image.

5. The developing roller image defect detection method according to claim 4, characterized in that: In the one-to-one pairing of the fourth region image with each element of a defect image set in a database, the defect image set includes at least one of the following defect images: a bubble defect image, a coil defect image, a port defect image, a rubber surface defect image, a chatter mark defect image, a dot mark defect image, a block defect image, a line defect image, and a yin-yang defect image; Wherein, each element in the defect image set is an image containing only the corresponding defect shape of the developing roller, and each element in the defect image set is obtained by inputting the defective developing roller image into the convolutional neural network model for training.

6. The developing roller image defect detection method according to claim 5, characterized in that: In the one-to-one pairing of the fourth region image with each element of the defect image set in the database, determining the defect type of the developing roller to be inspected according to the similarity of the pairing results includes: Performing a similarity comparison between the fourth region image and each element of the defect image set in the database, wherein the similarity comparison includes image contour comparison and pixel number comparison; If the similarity between the image contour of the fourth area image and the image contour of the defect image in the defect image set is greater than a second preset threshold, and the difference between the number of pixels of the fourth area image and the number of pixels of the defect image is within a preset range value, then the fourth area image is determined to be the defect type corresponding to the defect image.

7. The developing roller image defect detection method according to claim 4, characterized in that: The step of determining the defect type of the developing roller to be inspected according to the defect distribution of the fourth region image includes: Taking one end point of the developing roller to be inspected as the origin, a plane rectangular coordinate system is established; sequentially inputting the M fourth region images into the plane rectangular coordinate system, and generating a pixel coordinate based on each pixel in each of the fourth region images, thereby obtaining M pixel coordinate sets, wherein one of the pixel coordinate sets includes all pixel coordinates in one of the fourth region images; Performing separate position distribution analysis on the M pixel coordinate sets respectively, and determining defect type information of the M fourth region images respectively, wherein the defect type information includes a type A defect, a type B defect, and a type C defect; If there are one or more images of type C defect in the M fourth area images, it is determined that the developing roller to be inspected is a developing roller with type C defect, and a type C defect detection report is generated; If none of the M fourth area images has an image of defect type C, and the number of developing rollers determined to be defective type B is greater than or equal to the number of developing rollers determined to be defective type A, the developing roller to be inspected is determined to be a developing roller with defective type B, and a defect detection report for type B is generated; If none of the M fourth area images contain images of defect type C, and the number of developing rollers determined to be defective type B is less than the number of developing rollers determined to be defective type A, then the developing roller to be inspected is determined to be a defective developing roller of type A, and a defect detection report of type A is generated.

8. The developing roller image defect detection method according to claim 7, characterized in that: The performing separate position distribution analysis on the M pixel coordinate sets to respectively determine defect type information of the M fourth region images includes: If all pixel coordinates in the pixel coordinate set are gathered within the same first unit area, the fourth area image is determined to be a type of defect A, and the type of defect A includes at least one of the following defects: bubble defect, dot defect, and block defect; If the number of pixel coordinates in the pixel coordinate set at the same horizontal coordinate or vertical coordinate reaches a third preset threshold, the fourth region image is determined to be a type of defect of Class B, and the type of defect of Class B includes at least one of the following defects: chatter mark defect, line defect, coil defect, and yin-yang defect; If all pixel coordinates in the pixel coordinate set are gathered within the same second unit area range, the fourth area image is determined to be a Class C defect type, and the Class C defect type includes at least one of the following defects: port defects and adhesive surface defects; wherein the second unit area is larger than the first unit area range.

9. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 8.

10. A non-transitory machine-readable storage medium having executable codes stored thereon, which, when executed by a processor of an electronic device, causes the processor to execute the method according to any one of claims 1 to 8.

Images