Optical fiber end face pollution identification method based on image processing

Abstract

The invention relates to the technical field of image data processing, in particular to an optical fiber end face pollution identification method based on image processing, which comprises the following steps: acquiring a circular region image of an optical fiber end face to be detected; equally dividing the circular area image of the end face of the optical fiber to be detected into a plurality of fan-shaped area images, and determining initial abnormal degrees of the circular area image and the fan-shaped area images; determining the abnormal degree of the fan-shaped region image; determining the pollution degree of the to-be-detected optical fiber end face; and according to the pollution degree, pollution identification of the optical fiber end face is realized. According to the invention, the circular region is subdivided into a plurality of fan-shaped regions, so that local analysis can be carried out for pollution characteristics of different regions, and the characteristics of various types of pollution can be captured more comprehensively; the gray value frequency is used to construct the sequence, and according to the correlation between the gray information and the sequence, the detection capability of the gray distribution abnormity is enhanced, and the defect that the edge detection has no obvious edge pollution is made up.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing. More specifically, the present invention relates to a method for identifying fiber optic end face contamination based on image processing. Background Art

[0002] As an important medium for information transmission, optical fibers play an important role in the communication field. The information transmission of optical fibers is achieved by converting information into optical signals and through total internal reflection of light. When there is contamination on the end face of the optical fiber (such as oil stains, dust), the contaminants will scatter or absorb the optical signal, thereby affecting information transmission.

[0003] The Chinese patent application document with the publication number CN117288764A discloses a method and system for detecting the quality of an optical fiber end face. The method includes: obtaining an image of the optical fiber end face, determining the center and radius of the optical fiber according to the image of the optical fiber end face; dividing the image of the optical fiber end face into regions based on the center and radius of the optical fiber to obtain detection regions, where the detection regions include a core detection region, a cladding detection region, and a ceramic detection region; performing masking processing on each detection region to obtain a corresponding masking pattern, and detecting the masking pattern based on the Canny edge detection method to determine the stain position information of each detection region; marking the image of the optical fiber end face according to the stain position information to determine the stain marking contour, and judging the quality of the optical fiber end face according to the stain marking contour.

[0004] However, when determining the stain position information of each detection region in the detection using the above document, since different types of stains on the optical fiber end face in the image have different image characteristics, some contaminations have obvious edges in the image, while some contaminations do not have obvious edges in the image. Therefore, the above document can identify the contaminations with obvious edges through edge detection, but cannot detect the contaminations without obvious edges, reducing the detection accuracy, and thus unable to effectively identify the contamination of the optical fiber end face. Summary of the Invention

[0005] In order to solve the problem that due to different types of stains on the optical fiber end face in the image having different image characteristics, some contaminations have obvious edges in the image, while some contaminations do not have obvious edges in the image, and traditional edge detection can identify the contaminations with obvious edges, but cannot detect the contaminations without obvious edges, resulting in a decrease in detection accuracy and thus unable to effectively identify the contamination of the optical fiber end face, the present invention proposes a method for identifying fiber optic end face contamination based on image processing, which includes the following steps: Divide the image of the circular area of the fiber end face to be detected into several sector area images; Denote the image of the circular area and any one of all the sector area images as the target image. According to the frequency of occurrence of each gray value in the target image, construct the gray frequency sequence of the target image. According to the magnitude of the frequency, construct the main gray frequency sequence of the target image, and obtain the main gray value of the target image; Determine the initial abnormal degree of the target image according to the frequency of occurrence of each main gray value in the target image, the number of pixel points in the target image, and the correlation between the gray frequency sequence and the main gray frequency sequence of the target image; For any sector area image, obtain the representative gray value of each radius line in the sector area image, where the representative gray value is the average value of the gray values of all pixel points on the corresponding radius line in the sector area. Determine the abnormal degree of the sector area image according to the difference between the initial abnormal degrees of the circular area image and the sector area image, and the range of the representative gray values of all radius lines in the sector area image; Determine the pollution degree of the fiber end face to be detected according to the initial abnormal degree of the circular area image and the maximum and minimum values among the abnormal degrees of all sector area images; Realize the pollution identification of the fiber end face according to the magnitude of the pollution degree.

[0006] In the present invention, by subdividing the circular area of the fiber end face into several sector areas, local analysis can be carried out for the pollution characteristics of different areas, so as to more comprehensively capture the image characteristics of various types of pollution; Use the frequency of the gray values of the target image to construct the gray frequency sequence and the main gray frequency sequence, and calculate the abnormal degree based on the main gray value and its frequency of the gray value, which enhances the detection ability of the abnormal gray distribution and makes up for the defect of traditional edge detection in the case of no obvious edge pollution; Through the gray mean value and its range of multiple radius lines in the sector area, the details of the local gray change can be reflected, and the degree of regional pollution can be further accurately characterized to realize the effective identification of fine pollution; Combining the overall abnormal degree of the circular area and the maximum and minimum values of the abnormal degrees in all sector areas can comprehensively judge the overall situation of the pollution, and improve the objectivity and accuracy of the pollution degree evaluation.

[0007] Further, the acquisition method of the circular area image is as follows: Obtain the original image of the fiber end face to be detected, use the Hough circle detection to obtain all circular areas in the original image, and take the circular area image with the largest radius after gray-scale processing as the circular area image of the fiber end face to be detected.

[0008] Further, constructing the gray frequency sequence of the target image includes: According to the gray value order, form the gray frequency sequence of the target image with the frequency of occurrence of each gray value in the target image.

[0009] Further, the construction of the main gray frequency sequence of the target image includes: sorting the frequencies of each gray value in the target image from largest to smallest, retaining the frequencies of the preset number of gray values in order, and assigning the frequencies of the remaining gray values to 0. The preset number is the same as the number of layers of the fiber end face to be detected; according to the gray value order, the frequencies of each gray value in the target image are used to form the main gray frequency sequence of the target image.

[0010] Further, the obtaining of the main gray value among the gray values in the target image includes: sorting the frequencies of each gray value in the target image from largest to smallest, and selecting the gray values corresponding to the frequencies of the preset number of gray values in order as the main gray values of the target image. The preset number is the same as the number of layers of the fiber end face to be detected.

[0011] Further, the initial degree of abnormality satisfies: ; where is the initial degree of abnormality of the target image, is the number of main gray values, is the th frequency of the main gray value appearing in the target image, is the number of pixel points in the target image, is the correlation between the gray frequency sequence and the main gray frequency sequence of the target image.

[0012] In the present invention, by uniformly quantifying the frequency of occurrence of the main gray value, the total number of pixels, and the correlation between the gray frequency sequence and the main gray frequency sequence in the target image, the initial degree of abnormality is calculated, avoiding the subjectivity of traditional empirical judgment; the weighted sum of the frequencies of occurrence of the main gray value reflects the proportion of the main gray value in the image. Combining the correlation between the gray frequency sequence and the main gray frequency sequence, not only the distribution of the main gray is concerned, but also the change of the gray pattern is taken into account, improving the detection ability for subtle abnormalities.

[0013] Further, the correlation is the cosine similarity.

[0014] Further, the degree of abnormality satisfies: ; where is the degree of abnormality of the th sector region image, is the initial degree of abnormality of the circular region image, is the th initial degree of abnormality of the sector region image, is the range difference of the representative gray values of all radius lines in the th sector region image, is the absolute value symbol.

[0015] The degree of abnormality of the present invention accurately reflects the abnormal deviation of the local area from the whole by comparing the difference in the initial degree of abnormality between the circular area (whole) and the fan-shaped area (local), which is helpful to locate the specific pollution position; the representative gray value extreme difference of all radius lines in the fan-shaped area is used to reflect the gray fluctuation range of the area, and the difference in the degree of abnormality is used as a weight, so that the area with drastic gray change (which may correspond to obvious pollution) has a higher degree of abnormality, thereby improving the recognition ability of different pollution types; the gray value extreme difference is divided by the maximum possible gray value of 255 to achieve normalization, so that the degree of abnormality of different images or optical fiber end faces under different acquisition conditions is comparable, which improves the practicality and stability of the method.

[0016] ;in, is the contamination degree of the optical fiber end face, is the initial abnormality degree of the circular area image, is the maximum value of the abnormality degree of all fan-shaped area images, It is the minimum value of the abnormality degree of all fan-shaped area images.

[0017] The present invention nonlinearly adjusts the cardinality through the power exponential term, so that the pollution level not only depends on the overall and maximum abnormal value, but also can be dynamically adjusted according to the distribution difference of the abnormal level, thereby improving the detailed description of the pollution state. When the range of the abnormal level in the fan-shaped area is large, the index will decrease, thereby suppressing the value of the pollution level, reflecting the situation that the pollution distribution is uneven, the local difference is large, but the overall pollution may not be serious; on the contrary, when the range is small, the index is close to 1, highlighting the overall pollution level, and improving the scientific nature of the judgment.

[0018] Furthermore, the method for realizing contamination identification of the optical fiber end face includes: in response to the contamination degree being greater than a preset contamination threshold, determining that the optical fiber end face to be inspected is contaminated, and issuing a warning prompt, completing the contamination identification of the optical fiber end face based on image processing, wherein the contamination threshold is the average of the contamination degrees of multiple uncontaminated optical fiber end faces.

[0019] By comparing the pollution degree with a preset pollution threshold, the system of the present invention can automatically identify the pollution state of the optical fiber end face without manual intervention, thereby improving detection efficiency and accuracy; the pollution threshold is set based on the average pollution degree of multiple pollution-free optical fiber end faces, ensuring that the threshold is representative and objective, avoiding misjudgment or missed judgment caused by a fixed threshold, and helping to adapt to detection needs under different equipment and environments; once it is detected that the pollution degree exceeds the threshold, the system immediately issues a warning prompt to achieve early detection and early processing, reduce the impact of pollution on optical fiber transmission performance, and improve the reliability of the communication system.

[0020] The present invention has the following beneficial effects: (1) Traditional methods rely on edge detection and can only identify contaminants with obvious edges. In contrast, the present invention can capture contaminants without obvious edges (such as uniformly attached stains, diffuse reflection impurities, etc.) through the analysis of the gray-scale frequency sequence and the main gray-scale frequency sequence. For example, when the fiber end face is evenly covered with oil stains, although there is no clear edge, the gray-scale distribution will deviate from the normal state, and accurate identification can be achieved through the frequency and correlation analysis of the main gray-scale values. By combining the initial degree of abnormality (based on the gray-scale distribution) and the degree of abnormality in the fan-shaped area (based on the gray-scale range of the radius line), contamination is judged from both a global (circular area) and a local (fan-shaped area) perspective, avoiding misjudgment in a single dimension. For example, when there are local stains in a certain fan-shaped area, the gray-scale range of its radius line will increase significantly, forming a difference from other areas, and thus being accurately detected.

[0021] (2) The present invention processes data through methods such as the mean value of gray-scale values (representing gray-scale values) and frequency statistics, which can weaken the influence of random noise. For example, when there are a small number of noise points in the image, the frequency of their gray-scale values appears low and will not significantly interfere with the construction of the main gray-scale frequency sequence. However, the gray-scale changes in the contaminated area, due to their regularity (such as the gray-scale offset of consecutive pixels), can be effectively captured. Whether it is a dot-shaped, strip-shaped, or large-area contamination, positioning can be achieved through the subdivision of the fan-shaped area and the analysis of the radius line. For example, for a scratch distributed along the radius direction, the mean value of the gray-scale of the radius line in the fan-shaped area where it is located will show regular fluctuations and the range will increase, thus being identified as an abnormal area.

[0022] (3) The present invention mainly involves basic operations such as gray-scale statistics, mean value calculation, and correlation analysis, without the need for complex deep learning models or high-computing-power hardware support. It can be quickly deployed in embedded systems or industrial cameras and is suitable for real-time online detection scenarios. Parameters such as the number of fan-shaped areas and the screening threshold of the main gray-scale frequency sequence can be adjusted according to the material, light reflection characteristics, etc. of different types of fiber end faces, improving the versatility and adaptability of the algorithm. Description of the Drawings

[0023] Figure 1 is a schematic diagram of fiber end face contamination in the fiber end face contamination recognition method based on image processing according to an embodiment of the present invention.

[0024] Figure 2 is a schematic diagram of edge detection of the fiber end face in the fiber end face contamination recognition method based on image processing according to an embodiment of the present invention.

[0025] Figure 3 is a flowchart of the steps of the fiber end face contamination recognition method based on image processing according to an embodiment of the present invention.

[0026] Figure 4It is a schematic diagram for dividing the image of a fan-shaped area in the fiber end face pollution recognition method based on image processing according to an embodiment of the present invention.

[0027] Figure 5 It is a schematic diagram of multiple radius lines of any fan-shaped area image in the fiber end face pollution recognition method based on image processing according to an embodiment of the present invention. Detailed implementation manners

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described. The described embodiments are part of the embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative efforts based on the embodiments of the present invention belong to the protection scope of the present invention.

[0029] Next, the specific implementation manners of the present invention will be described in detail with reference to the accompanying drawings.

[0030] As Figure 1 and Figure 2 shown, since edge detection cannot detect the pollution on the left side in Figure 1 that has no obvious edge, resulting in the inability to effectively identify the pollution of the fiber end face. Therefore, the present invention proposes a fiber end face pollution recognition method based on image processing.

[0031] Please refer to Figure 3 , which shows a flowchart of the steps of a fiber end face pollution recognition method based on image processing provided by an embodiment of the present invention. The method includes the following steps: S01: Obtain a circular area image of the fiber end face to be detected.

[0032] Specifically, obtain the original image of the fiber end face to be detected, use the Hough circle detection to obtain all circular areas in the original image, and use the circular area image with the largest radius after grayscale processing as the circular area image of the fiber end face to be detected.

[0033] S02: Divide the circular area image of the fiber end face to be detected into several fan-shaped area images, and determine the initial abnormal degree of the circular area image and the fan-shaped area images.

[0034] It should be noted that due to the use characteristics of the optical fiber, even if there are only fine pollutants on the optical fiber, it may have a great impact on the use of the optical fiber. And the various structures of the fiber end face have a fixed shape, and the fiber end face is centrosymmetric. When there are pollutants on the fiber end face, it will change the image characteristics of some areas in the image. In order to refine the fiber end face pollution recognition and enable fine pollutants to be recognized. Therefore, in this step, the circular area image is divided to obtain several fan-shaped area images; and the gray frequency sequence, the main gray frequency sequence and the main gray value are determined according to the gray information.

[0035] Divide the circular area image of the fiber optic end face to be detected into several sector area images. Denote the circular area image and any one of all the sector area images as the target image. According to the frequency of occurrence of each gray value in the target image, construct the gray frequency sequence of the target image. According to the magnitude of the frequency, construct the main gray frequency sequence of the target image, and obtain the main gray value of the target image.

[0036] As Figure 4 shown, the implementer can set the number of sector area images according to the specific implementation situation. For example, 8.

[0037] Specifically, the construction of the gray frequency sequence of the target image includes: In the order of gray values, form the frequency of occurrence of each gray value in the target image into the gray frequency sequence of the target image (serialized representation of the gray histogram of the target image).

[0038] Specifically, the construction of the main gray frequency sequence of the target image includes: Sort the frequency of occurrence of each gray value in the target image from large to small, retain the frequency of occurrence of the preset number of gray values in order, and assign the frequency of occurrence of the remaining gray values to 0. The preset number is the same as the number of layers of the fiber optic end face to be detected (4 in this example, and the implementer can set it according to the actual number of layers of the fiber optic to be detected); In the order of gray values, form the frequency of occurrence of each gray value in the target image into the main gray frequency sequence of the target image.

[0039] Specifically, the obtaining of the main gray value among the gray values in the target image includes: Sort the frequency of occurrence of each gray value in the target image from large to small, and select the gray values corresponding to the frequency of occurrence of the preset number of gray values in order as the main gray value of the target image. The preset number is the same as the number of layers of the fiber optic end face to be detected.

[0040] It should be noted that an optical fiber usually consists of four parts: the core, the inner cladding, the outer cladding, and the protective layer. Since different components of the optical fiber are made of different materials, the core, the cladding, and the protective layer in the optical fiber end face image have different color characteristics. When there are areas in the image that deviate from the image features of the four components of the optical fiber, there are contaminants on the optical fiber end face. Therefore, the contamination of the optical fiber end face can be identified by the color distribution of the optical fiber end face image. Therefore, this step obtains the initial degree of abnormality of the optical fiber end face image according to the gray value distribution of the image.

[0041] Determine the initial degree of abnormality of the target image based on the frequency of occurrence of each main gray value in the target image, the number of pixel points in the target image, and the correlation between the gray frequency sequence and the main gray frequency sequence of the target image.

[0042] Specifically, the initial degree of abnormality satisfies: ; In the formula, is the initial degree of abnormality of the target image, is the number of main gray values, is the frequency of occurrence of the th main gray value in the target image, is the number of pixel points in the target image, is the correlation between the gray frequency sequence and the main gray frequency sequence of the target image.

[0043] Specifically, the correlation is the cosine similarity.

[0044] Among them, since the fiber end face structure has 4 layers in this example, corresponding to 4 regions, and different regions have different color characteristics, when the fiber end face is pollution-free, the sum of the number of pixel points corresponding to the top 4 gray values (main gray values) with the highest frequencies in the target image, should be close to the number of pixel points in the target image. When there is pollution on the fiber end face, the polluted part will cover the fiber, so that there will be a certain change in the gray value distribution of the target image, making deviate from ; Therefore the larger, the less likely it is that the target image contains pollution, and the smaller the initial degree of abnormality of the target image; the smaller, the more likely it is that the target image contains pollution, and the larger the initial degree of abnormality of the target image. The larger , it means that the top 4 main gray values with the highest frequencies in the target image occupy more positions in the target image, so the target image is less likely to have pollutants, and the initial degree of abnormality of the target image is smaller; The smaller , it means that the top 4 main gray values with the highest frequencies in the target image occupy fewer positions in the target image, so the target image is more likely to have pollutants, and the initial degree of abnormality of the target image is larger.

[0045] S03: Determine the degree of abnormality of the fan-shaped region image.

[0046] It should be noted that the contamination on the optical fiber end face may exhibit local and directional characteristics (such as scratches distributed along the radial direction and local stains within a fan-shaped area). It is difficult to accurately locate the contaminated area only through the global initial abnormal degree. Therefore, in this step, by analyzing the gray-scale distribution characteristics of the radial lines within the fan-shaped area and combining the global and local abnormal differences, refined positioning of the contamination is achieved.

[0047] For any fan-shaped area image, such as Figure 5 shown, obtain the representative gray-scale value of each radial line within the fan-shaped area image (the implementer can set the number of radial lines according to the specific implementation situation. For example, set one every 1°). The representative gray-scale value is the average of the gray-scale values of all pixel points on the corresponding radial line within the fan-shaped area. Determine the abnormal degree of the fan-shaped area image based on the difference between the initial abnormal degree of the circular area image and the fan-shaped area image, and the range of the representative gray-scale values of all radial lines within the fan-shaped area image.

[0048] Specifically, the abnormal degree satisfies: ; In the formula, is the abnormal degree of the th fan-shaped area image, is the initial abnormal degree of the circular area image, is the initial abnormal degree of the th fan-shaped area image, is the range of the representative gray-scale values of all radial lines within the th fan-shaped area image, is the absolute value symbol.

[0049] Among them, since the optical fiber end face is centrosymmetric, when there is no pollutant on the optical fiber end face, the gray-scale value distribution of the pixel points within the circular area image should be similar to the gray-scale value distribution of the fan-shaped area image. Therefore, when there is no pollutant on the optical fiber end face, the initial abnormal degree between the circular area image and the fan-shaped area image should be similar. Therefore, the smaller it is, the less likely there is a pollutant in the th fan-shaped area image, and the smaller the abnormal degree of the th fan-shaped area image; the larger it is, the more likely there is a pollutant in the th fan-shaped area image, and the The greater the degree of abnormality of the image of a sector area. Similarly, since the fiber end face is centrosymmetric, when there is no contaminant on the fiber end face, the mean value (representing the gray value) of the pixel points on the radius line corresponding to the image of each sector area should be the same or similar. When there is a contaminant on the fiber end face, there will be a certain difference between the representative gray values corresponding to the sector area images. Therefore, The greater, the more likely there is a contaminant in the image of the th sector area, and the greater the degree of abnormality of the image of the th sector area; The smaller, the less likely there is no contaminant in the image of the th sector area, and the smaller the degree of abnormality of the image of the

[0050] th sector area. For the convenience of calculation, is normalized by dividing it by 255 (the maximum gray value).

[0050] S04: Determine the pollution degree of the fiber end face to be detected.

[0051] It should be noted that when there is pollution on the fiber end face, there will also be corresponding changes in the image features of each arc area of the fiber end face diagram. Therefore, in this step, according to the initial degree of abnormality of the circular area image of the fiber end face and the degree of abnormality of the sector area image, the pollution degree of the fiber end face is obtained.

[0052] According to the initial degree of abnormality of the circular area image and the maximum and minimum values among the degrees of abnormality of all sector area images, determine the pollution degree of the fiber end face to be detected.

[0053] Specifically, the pollution degree satisfies: ; where is the pollution degree of the fiber end face, is the initial degree of abnormality of the circular area image, is the maximum value among the degrees of abnormality of all sector area images, is the minimum value among the degrees of abnormality of all sector area images.

[0054] where The greater it is, the more likely there is a contaminant on the fiber end face that changes the gray value distribution of the fiber end face image, and the greater the pollution degree of the fiber end face; The smaller it is, the more likely there is no contaminant on the fiber end face, and the relatively smaller the degree of abnormality of the fiber end face. The larger it is, the greater the difference in the degree of abnormality between the images of each sector region. When there is no defect on the fiber end face, the degrees of abnormality between the images of each sector region should be similar. To make the pollution degree of the defective fiber end face more prominent, by to perform upward correction, when it is larger, the smaller it is, then to the greater the correction degree, making the fiber end face with pollution have a greater pollution degree; when it is smaller, the larger it is, then the closer it is to 1, to the smaller the correction degree. But there are two situations at this time. One is that when the pollution of the images of each sector region is relatively uniform, although is smaller, but because and themselves have relatively large values, it will still reflect the overall pollution degree, and the pollution degree will be relatively large at this time; the other situation is that when there is no pollution on the fiber end face, the degrees of abnormality of the images of each sector region all approach 0, so approaches 0, and approaches 1, and the pollution degree will be relatively small at this time to avoid misidentifying a fiber end face without pollution as a polluted state.

[0055] S05: According to the magnitude of the pollution degree, to achieve the pollution identification of the fiber end face.

[0056] Specifically, the achieving of the pollution identification of the fiber end face includes: In response to the pollution degree being greater than a preset pollution threshold, it is determined that there is pollution on the fiber end face to be detected, and a warning prompt is issued to complete the pollution identification of the fiber end face based on image processing. The pollution threshold is the average value of the pollution degrees of multiple fiber end faces without pollution (implementers can also set it according to specific implementation situations, and multiple thresholds can also be set to refine the pollution state).

[0057] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for identifying fiber end face contamination based on image processing, characterized in that, Including: Dividing the circular region image of the fiber end face to be detected into several sector region images evenly; Denoting the circular region image and any one of all the sector region images as the target image. According to the frequency of occurrence of each gray value in the target image, constructing the gray frequency sequence of the target image. According to the magnitude of the frequency, constructing the main gray frequency sequence of the target image, and obtaining the main gray value of the target image. Determining the initial abnormal degree of the target image according to the frequency of occurrence of each main gray value in the target image, the number of pixel points in the target image, and the correlation between the gray frequency sequence and the main gray frequency sequence of the target image; For any sector region image, obtaining the representative gray value of each radius line in the sector region image, where the representative gray value is the average of the gray values of all pixel points on the corresponding radius line in the sector region. Determining the abnormal degree of the sector region image according to the difference between the initial abnormal degrees of the circular region image and the sector region image, and the range of the representative gray values of all radius lines in the sector region image; Determining the pollution degree of the fiber end face to be detected according to the initial abnormal degree of the circular region image and the maximum and minimum values among the abnormal degrees of all sector region images; Realizing the pollution identification of the fiber end face according to the magnitude of the pollution degree.

2. The method for identifying the contamination of the optical fiber end face based on image processing according to claim 1, wherein The obtaining method of the circular region image is as follows: Obtaining the original image of the fiber end face to be detected, using the Hough circle detection to obtain all circular regions in the original image, and taking the circular region with the largest radius after gray-scale processing as the circular region image of the fiber end face to be detected.

3. The method for identifying fiber end face contamination based on image processing according to claim 1, wherein The constructing of the gray frequency sequence of the target image includes: According to the gray value order, forming the gray frequency sequence of the target image with the frequency of occurrence of each gray value in the target image.

4. The method for identifying the contamination of the optical fiber end face based on image processing according to claim 1, wherein The constructing of the main gray frequency sequence of the target image includes: Sorting the frequency of occurrence of each gray value in the target image from large to small, retaining the frequency of occurrence of the preset number of gray values in order, and assigning the frequency of occurrence of the remaining gray values to 0, where the preset number is the same as the number of layers of the fiber end face to be detected; According to the gray value order, forming the main gray frequency sequence of the target image with the frequency of occurrence of each gray value in the target image.

5. The method for identifying the contamination of the optical fiber end face based on image processing according to claim 1, wherein, The obtaining of the main gray value among the gray values in the target image includes: Sorting the frequency of occurrence of each gray value in the target image from large to small, and selecting the gray values corresponding to the frequency of occurrence of the preset number of gray values in order as the main gray value of the target image, where the preset number is the same as the number of layers of the fiber end face to be detected.

6. The method for identifying the contamination of the optical fiber end face based on image processing according to claim 1, wherein The initial abnormal degree satisfies: ； In the formula, is the initial abnormal degree of the target image, is the number of main gray values, is the frequency of occurrence of the th main gray value in the target image, is the number of pixel points in the target image, is the correlation between the gray frequency sequence and the main gray frequency sequence of the target image.

7. The method for identifying the contamination of the optical fiber end face based on image processing according to claim 1 or 6, characterized in that, The correlation is the cosine similarity.

8. The method for identifying fiber end face contamination based on image processing according to claim 1, wherein The abnormal degree satisfies: ； Wherein, is the abnormality degree of the th sector area image, is the initial abnormality degree of the circular area image, is the initial abnormality degree of the th sector area image, is the range of the representative gray values of all the radius lines within the th sector area image, is the absolute value symbol.

9. The method for identifying fiber end face contamination based on image processing according to claim 1, wherein The pollution degree satisfies: ； Among them, is the contamination degree of the fiber end face, is the initial abnormal degree of the circular region image, is the maximum value among the abnormal degrees of all sector region images, is the minimum value among the abnormal degrees of all sector region images.

10. The method for identifying fiber end face contamination based on image processing according to claim 1, characterized in that, The realizing of the pollution identification of the fiber end face includes: In response to the pollution degree being greater than the preset pollution threshold, determining that the fiber end face to be detected is polluted, and sending out a warning prompt to complete the pollution identification of the fiber end face based on image processing, where the pollution threshold is the average value of the pollution degrees of multiple fiber end faces without pollution.

Images