CN117934461A - Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber - Google Patents
Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber Download PDFInfo
- Publication number
- CN117934461A CN117934461A CN202410324286.2A CN202410324286A CN117934461A CN 117934461 A CN117934461 A CN 117934461A CN 202410324286 A CN202410324286 A CN 202410324286A CN 117934461 A CN117934461 A CN 117934461A
- Authority
- CN
- China
- Prior art keywords
- optical fiber
- gray level
- polished
- occurrence matrix
- polishing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000013307 optical fiber Substances 0.000 title claims abstract description 127
- 238000005498 polishing Methods 0.000 title claims abstract description 120
- 238000000034 method Methods 0.000 title claims abstract description 61
- 230000003746 surface roughness Effects 0.000 title claims abstract description 38
- 239000011159 matrix material Substances 0.000 claims abstract description 62
- 230000003287 optical effect Effects 0.000 claims abstract description 35
- 238000004458 analytical method Methods 0.000 claims abstract description 19
- 238000007637 random forest analysis Methods 0.000 claims description 30
- 239000000835 fiber Substances 0.000 claims description 14
- 238000003860 storage Methods 0.000 claims description 14
- 238000002360 preparation method Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 abstract description 7
- 238000012545 processing Methods 0.000 description 9
- 238000003066 decision tree Methods 0.000 description 8
- 244000137852 Petrea volubilis Species 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 238000000691 measurement method Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000000227 grinding Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 101000805921 Strongylocentrotus purpuratus Upstream stimulatory factor Proteins 0.000 description 1
- 241001422033 Thestylus Species 0.000 description 1
- 101000671634 Xenopus borealis Upstream stimulatory factor 1 Proteins 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000005253 cladding Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 239000005447 environmental material Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000012466 permeate Substances 0.000 description 1
- 230000003094 perturbing effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Abstract
The present disclosure relates to a method, a system and a device for analyzing the polishing surface roughness of a side polished optical fiber, wherein the method comprises the following steps: s01, acquiring an optical microscopic image containing the polished surface reflection light of the target optical fiber as an original image; s02, constructing a gray level co-occurrence matrix, and extracting texture features in the original image through the gray level co-occurrence matrix; s03, classifying and identifying the polished surface of the target optical fiber based on the extracted texture features to obtain the roughness grade of the target optical fiber. The system and apparatus are for performing the above method. The method for detecting and analyzing the polishing surface roughness of the side polishing optical fiber based on the machine vision method has the advantages of being low in implementation cost, simple to operate, efficient, accurate in analysis and capable of supporting online measurement.
Description
Technical Field
The disclosure relates to the technical field of optical fiber sensing, in particular to a method, a system and equipment for analyzing polishing surface roughness of a side polishing optical fiber.
Background
The optical fiber evanescent field sensor, also called an optical fiber evanescent wave sensor, can realize the sensing detection of environmental materials with high sensitivity by utilizing the evanescent field which permeates into surrounding media with different refractive indexes from the optical fiber core. The Side polished optical fiber (Side-polished fiber, SPF) is a special optical fiber manufactured by removing part of the cladding on the standard optical fiber by utilizing an optical micro-processing technology, and after the Side polishing, an evanescent field leaks from the fiber core, so that the optical fiber polished area is very sensitive to the covering material. By combining the side polished optical fiber with an excellent optical material, a high performance side polished optical fiber sensor can be obtained.
In recent years, researchers continuously improve the manufacturing process of the side polished optical fiber, including a V-shaped groove base block polishing technology, a wheel polishing technology, a femtosecond laser processing technology and the like, but various processing technologies inevitably cause the side polished optical fiber to generate a polished surface with a certain roughness.
The texture characteristics of the side polishing optical fiber polishing surface are important components of the surface roughness, the texture characteristics of the polishing surface influence the scattering characteristics of the optical fiber polishing surface, and have great influence on the optical performance, the sensing sensitivity, the reliability and the biocompatibility of the side polishing optical fiber sensing device, but no related research on a side polishing optical fiber polishing surface texture characteristic analysis method and an edge vision implementation system is reported at present.
The traditional contact type surface roughness detection method represented by the stylus measurement method has systematic errors on the side polished optical fiber with smaller surface roughness, and meanwhile, the measurement efficiency is low, the online measurement cannot be realized, and the real rough texture characteristics of the two-dimensional surface cannot be represented. The optical measurement method is used as a typical non-contact measurement method, so that deformation and errors caused by contact measurement on polished optical fibers are avoided, but the method still has the defects of time consumption in scanning, difficulty in real-time online measurement, high price and the like, and cannot be applied to the preparation process of side polished optical fibers.
In summary, in the prior art, there is no method for detecting and analyzing the roughness of the surface of the side polished optical fiber, and the conventional method for detecting the surface roughness has the defects of large error, low efficiency and high cost, and cannot be well applied to the process of detecting the roughness of the surface of the side polished optical fiber, so a more effective method for analyzing the roughness of the surface of the side polished optical fiber is provided, which is a technical problem to be solved by those skilled in the art.
Disclosure of Invention
In order to solve the problems in the prior art, the disclosure aims to provide a method, a system and equipment for analyzing polishing surface roughness of a side polishing optical fiber. The method for detecting and analyzing the polishing surface roughness of the side polishing optical fiber based on the machine vision method has the advantages of being low in implementation cost, simple to operate, efficient, accurate in analysis and capable of supporting online measurement.
The method for analyzing the polishing surface roughness of the side polishing optical fiber comprises the following steps:
S01, acquiring an optical microscopic image containing the polished surface reflection light of the target optical fiber as an original image;
S02, constructing a gray level co-occurrence matrix, and extracting texture features in the original image through the gray level co-occurrence matrix;
S03, classifying and identifying the polished surface of the target optical fiber based on the extracted texture features to obtain the roughness grade of the target optical fiber.
Preferably, in step S02, constructing the gray level co-occurrence matrix includes: setting gray levelStep size/>And the direction of generation/>。
Preferably, the gray level is setThe method comprises the following steps:
Pre-selection of With individual gray levels as alternative gray levels,/>Select/>Each of the candidate gray levels is constructed separatelyA first candidate gray level co-occurrence matrix;
calculating contrast ratio of each first alternative gray level co-occurrence matrix to texture characteristic parameters extracted from optical microscopic images of polished surfaces of polished optical fibers on the same side ，/>；
Acquiring operation time consumption of extracting texture features of optical microscopic images of polished surfaces of polished optical fibers on the same side by using each first alternative gray level co-occurrence matrix，/>；
The selection index of each candidate gray level is calculated according to the following formula：
；
Wherein,And/>Coefficients representing contrast and run time, respectively, satisfy/>，/>，
Selecting the indexThe candidate gray level with the largest value is taken as the gray level of the gray level co-occurrence matrix。
Preferably, the step size is setThe method comprises the following steps:
Pre-selecting a plurality of Step sizes as alternative step sizes,/>Select/>Constructing/>, respectively, the alternative step sizesA second alternative gray level co-occurrence matrix;
Calculating entropy value of texture characteristic parameters extracted from optical microscopic images of polished surfaces of polished optical fibers on the same side by using each second alternative gray level co-occurrence matrix ，/>；
Presetting a value of entropyReasonable interval/>And a recommendation step set, wherein the reasonable interval/>The recommended step length set comprises a plurality of polishing paper parameters adopted in the preparation of the side polishing optical fiber, and each polishing paper parameter has a corresponding recommended step length;
determining whether at least one entropy value exists Is located in the reasonable interval/>If yes, selecting a reasonable interval/>Inner and closest to/>Step length corresponding to entropy value of the gray level co-occurrence matrix is used as step length/>If not, acquiring polishing paper parameters adopted in preparation of the target optical fiber, and selecting a corresponding recommended step length as the step length/>, of the gray level co-occurrence matrix, according to the recommended step length set。
Preferably, the generation direction is setThe method comprises the following steps:
Presetting the The generated direction angles are used as alternative angles, and/>Each of the alternative angles is constructed separately/>A third alternative gray level co-occurrence matrix;
Calculating the correlation of each third alternative gray level co-occurrence matrix to the texture characteristic parameters extracted from the optical microscopic image of the polished surface of the polished optical fiber on the same side ，/>；
Selecting the correlation thereofThe largest alternative angle is taken as the generation direction/>, of the gray level co-occurrence matrix。
Preferably, in step S02, extracting texture features in the original image through the gray level co-occurrence matrix includes:
Calculating ranking indexes of the gray level co-occurrence matrix on the texture characteristic parameters extracted from the polished surface image, wherein the ranking indexes comprise second moment, contrast, entropy, uniformity, correlation and variance, importance ranking is carried out on each ranking index by using at least two different ranking methods to obtain importance ranking results, and overlapping parts in each importance ranking result are selected as key indexes;
and extracting various key indexes in the original image, and constructing a texture feature subset of the polished surface of the target optical fiber.
Preferably, a random forest feature importance ranking method based on a coefficient of kunning and a random forest feature importance ranking method based on an OOB error are used for ranking the importance of the ranking indexes;
The key indexes comprise variance, second moment, entropy and contrast which are sequenced in sequence from big to small according to importance.
The utility model discloses a side polishing optic fibre grinds surface roughness analysis system, include:
the acquisition module is used for acquiring an optical microscopic image of the polished surface of the target optical fiber as an original image;
The extraction module is used for constructing a gray level co-occurrence matrix and extracting texture features in the original image through the gray level co-occurrence matrix;
And the classification module is used for classifying and identifying the target optical fiber based on the extracted texture characteristics to obtain the roughness grade of the target optical fiber.
A computer device of the present disclosure includes a signal-connected processor and a memory having stored therein at least one instruction or at least one program that, when loaded by the processor, performs a side polished fiber polished surface roughness analysis method as described above.
A computer readable storage medium of the present disclosure has stored thereon at least one instruction or at least one program that, when loaded by a processor, performs a side polished fiber polished surface roughness analysis method as described above.
The method, the system and the equipment for analyzing the roughness of the side polishing optical fiber polishing surface have the advantages that the method, the system and the equipment are based on edge vision, the reflection direction of light rays can be changed by combining with rough textures of the polishing surface, so that when uniform light rays are irradiated to the optical fiber polishing surface, the intensity and the distribution of the light rays entering into a camera sensor of an optical microscope can carry information of the roughness of the polishing surface, namely, an acquired optical microscopic image of the polishing surface can carry the information of the roughness of the polishing surface, and a random forest classification algorithm is combined, the correlation between various indexes of the image and the roughness of the polishing surface of the optical fiber is constructed by utilizing the texture feature importance sorting, and the characterization of the roughness information of the polishing surface is realized by acquiring the texture features of the side polishing optical fiber polishing surface, so that the roughness analysis of the polishing surface can be realized by adopting the optical microscopic image based on the reflection light rays of the polishing surface.
Drawings
FIG. 1 is a flow chart showing the steps of a method for analyzing the polishing surface roughness of a side polished optical fiber according to the present embodiment;
FIG. 2 is a schematic diagram of the structure of the polishing surface roughness analysis system of the side polishing optical fiber according to the present embodiment;
Fig. 3 is a schematic structural diagram of the computer device according to the present embodiment.
Reference numerals illustrate: the device comprises a 1-clamp table, a 2-light source, a 3-acquisition unit, a 4-processing unit, a 5A-first connection control unit, a 5B-second connection control unit, a 6-upper computer and a 7-target optical fiber.
101-Processor, 102-memory.
Detailed Description
As shown in fig. 1, a method for analyzing polishing surface roughness of a side polished optical fiber according to the present disclosure includes the following steps:
s01, acquiring an optical microscopic image containing the polished surface reflection light of the target optical fiber as an original image; specifically, uniform light is emitted by a light source to irradiate the polished surface of the target optical fiber, and an optical microscopic image including light reflected by the polished surface of the target optical fiber can be acquired by disposing an optical microscopic CCD vertically downward above the target optical fiber. Because the rough texture of the polished surface can change the reflection direction of light, the intensity and distribution of the light entering the optical microscope CCD can carry the information of the roughness of the polished surface.
S02, constructing a gray level co-occurrence matrix (GLCM), and extracting texture features in the original image through the gray level co-occurrence matrix;
further, in step S02, the construction of the gray level co-occurrence matrix includes calculating and setting three construction parameters, which are respectively: gray scale level Step size/>And the direction of generation/>。
Setting gray levelThe method comprises the following steps:
Pre-selection of With individual gray levels as alternative gray levels,/>Select/>Each of the candidate gray levels is constructed separatelyA first candidate gray level co-occurrence matrix;
calculating the Contrast ratio (Contrast) of each first alternative gray level co-occurrence matrix to the texture characteristic parameters extracted from the optical microscopic image of the polished optical fiber polished surface on the same side according to the following formula ，/>；
，/>；
Wherein,Representing the gray level in the image as/>Pixel to gray level of/>Probability of pixel of/>Representing gray levels;
Acquiring operation time consumption of extracting texture features of optical microscopic images of polished surfaces of polished optical fibers on the same side by using each first alternative gray level co-occurrence matrix ，/>；
The selection index of each candidate gray level is calculated according to the following formula：
；
Wherein,And/>Coefficients representing contrast and run time, respectively, satisfy/>，/>，
Selecting the indexThe candidate gray level with the largest value is taken as the gray level of the gray level co-occurrence matrixThus, the contrast/>, can be comprehensively consideredAnd run time, select the appropriate gray level.
The gray level reflects the quality of an image and the definition degree of textures, the higher the gray level setting is, the larger the contrast mean value is, the better the texture details of the image can be kept, but the too high gray level setting can cause the generated GLCM dimension to be too large to influence the operation speed, but when the parameter gray level is too small to be selected, the texture details can be lost after the gray level is compressed, and the real fiber polished surface texture distribution and roughness are not reflected.
For example, selecting gray levels of 16, 64, 128 and 256 as candidate gray levels, respectively, constructing four first candidate gray co-occurrence matrices, and calculating the contrast ratio of the four first candidate gray co-occurrence matrices by the above formulaRun time consuming/>And further calculate the selection index/>According to the selection index/>The most suitable gray level is selected.
The step length determines the distance between two pixels in the image and the point, if the texture of the image is thinner, the step length is not suitable to be set too large, otherwise, detail information is lost, and the extracted characteristic parameters cannot accurately describe the texture characteristics of the image; the step size setting is too small and the description of the coarser texture image falls into local uniformity.
Setting step lengthThe method comprises the following steps:
Pre-selecting a plurality of Step sizes as alternative step sizes,/>Select/>Constructing/>, respectively, the alternative step sizesA second alternative gray level co-occurrence matrix;
The entropy value is used for measuring the information quantity or complexity of the image, and if the image does not have any texture, namely the gray level co-occurrence matrix is almost zero, the entropy value is close to 0; if the image is distributed with less textures, the entropy value is smaller; if the image contains complex and rich textures, the entropy value of the image is larger and the information quantity is larger.
Therefore, in this embodiment, the entropy of the texture feature parameters extracted from the optical microscopic image of the polished fiber polished surface on the same side by each second candidate gray level co-occurrence matrix is calculated (Entropy)，/>；
，/>；
Wherein,Representing the gray level in the image as/>Pixel to gray level of/>Probability of pixel of/>Representing gray levels;
Presetting a value of entropy Reasonable interval/>And a recommendation step set, wherein the reasonable interval/>The recommended step length set comprises a plurality of polishing paper parameters adopted in the preparation of the side polishing optical fiber, and each polishing paper parameter has a corresponding recommended step length;
specifically, entropy value Reasonable interval/>The method can be obtained by combining multiple test experiments according to the characteristics of the polished surface optical microscopic image, and usually takes a central interval to avoid the condition that the entropy value is 0 or close to 0, and the characteristic parameters cannot describe the texture characteristics of the image, and similarly, the condition that the entropy value is too large is avoided, so that the method can be used for obtaining the polishing surface optical microscopic image through the reasonable interval/>And carrying out numerical screening on the entropy value to enable the entropy value to be in a reasonable range.
The polishing paper parameters adopted in the preparation of the side polishing optical fiber generally refer to the number of polishing sand paper, generally speaking, the larger the number of sand paper is, the smaller the grinding rate is, and the larger the grinding rate is, namely, the finer the texture of the polishing surface is, the larger the number of sand paper is, the larger the entropy value of the corresponding polishing surface image is, based on the relation, in the recommended step size set of the embodiment, the number of polishing sand paper and the recommended step size integrally show a tendency of negative correlation, in a specific embodiment, a plurality of number intervals from small to large can be set, and the recommended step size set from large to small is correspondingly formed.
Determining whether at least one entropy value existsIs located in the reasonable interval/>If yes, selecting a reasonable interval/>Inner and closest to/>Step length corresponding to entropy value of the gray level co-occurrence matrix is used as step length/>Specifically, if there is only one entropy value/>Located in reasonable interval/>In, directly select the entropy value/>The corresponding alternative step length is used as the step length/>, of the gray level co-occurrence matrixIf more than two entropy values exist/>Located in reasonable interval/>In, the respective entropy value is compared with a smaller boundary value/>The difference between the two is selected to be smaller, namely the nearest/>Step length corresponding to entropy value of the gray level co-occurrence matrix is used as step length/>。
If all entropy values are not in a reasonable intervalIn the process, parameters of polishing paper adopted in preparation of the target optical fiber, namely the mesh number of the polishing sand paper are obtained, a mesh number interval in which the mesh number of the polishing paper is positioned is determined according to the mesh number of the polishing sand paper, a recommended step length corresponding to the mesh number interval is further selected, and the corresponding recommended step length is selected as the step length of the gray level co-occurrence matrix。
Exemplary, selecting step sizes of 1,2, 4, 6 and 8 respectively to calculate entropy values of gray level co-occurrence matrix of side polished fiber polished surface imageAnd obtaining parameters, namely the mesh number, of the polishing paper prepared by wheel type polishing of the side polishing optical fiber.
When the average value of the entropy does not belong to the preset reasonable intervalIn the process, polishing paper parameters adopted in preparation of side polishing optical fibers, specifically the mesh number of sand paper, are obtained, and the step length/>, of the gray level co-occurrence matrix is selected。
Setting a generation directionThe method comprises the following steps:
Presetting the The generation direction angles are selected as alternative angles, and the generation direction angles are respectively 0 degree, 45 degrees, 90 degrees and 135 degrees, and/>, which are exemplaryEach of the alternative angles is constructed separately/>A third alternative gray level co-occurrence matrix;
Calculating the Correlation (Correlation) of each third alternative gray level co-occurrence matrix to the texture characteristic parameters extracted from the optical microscopic image of the polished surface of the polished optical fiber on the same side ，/>；
Selecting the correlation thereofThe largest alternative angle is taken as the generation direction/>, of the gray level co-occurrence matrix。
Further, in step S02, extracting texture features in the original image through the gray level co-occurrence matrix includes:
And calculating the ranking indexes of the gray level co-occurrence matrix on the texture characteristic parameters extracted from the polished surface image, wherein the ranking indexes comprise a second moment (Angluar Second Moment), a Contrast (Contrast), an entropy value (Entropy), uniformity (Homogeneity), correlation and variance (variance).
Specifically, the gray level of the gray level co-occurrence matrix is determined as described aboveStep size/>And the direction of generation/>The calculation formula for calculating the above indexes based on the three parameters belongs to common general knowledge, and will not be described here.
Performing importance ranking on each ranking index by using at least two different ranking methods to obtain importance ranking results, and selecting coincident parts in each importance ranking result as key indexes;
And extracting various key indexes in the original image, and constructing a texture feature subset of the polished surface of the target optical fiber.
In the step, firstly, indexes with larger influence on classification and identification of polishing surface roughness in six polishing surface images are listed, wherein the indexes are second moment, contrast, entropy, uniformity, correlation and variance respectively. The step aims at selecting a plurality of key indexes, such as four key indexes, which are required to meet the requirement of basically reflecting the roughness information of the polished surface, so that the parameters participating in calculation can be reduced, and the data operand is further reduced. When the target optical fiber is judged later, only the key indexes in the original image of the target optical fiber are required to be extracted, and all six indexes are not required to be extracted, so that the calculation amount in actual classification judgment can be reduced, and the response speed is improved.
In a specific embodiment, two different ranking methods are used for importance ranking, one is a random forest feature importance ranking method based on a coefficient of kunning and the other is a random forest feature importance ranking method based on an OOB error.
Specifically, a Random Forest (RF) feature importance ranking based on a coefficient of Kernine is adopted to obtain a feature importance ranking 1 of GLCM texture feature parameters for classifying and identifying the roughness of the polished surface of the optical fiber.
In the random forest algorithm, bootstrap self-sampling technology is adopted, namely, replaced sampling is carried out and repeated N times, so that N training sample sets are formed. And constructing decision trees one by one for the N training sample sets, and selecting branch characteristics of sample data based on the coefficient of the radix to realize sample classification. The final forest is composed of N CART decision trees, and after each tree is separately classified, the classification result is obtained by adopting a voting mode according to the prediction result of the N decision trees in the set of the new sample category.
And (3) adopting the Random Forest (RF) feature importance ranking based on the OOB error to obtain feature importance ranking 2 of the GLCM texture feature parameters for classifying and identifying the roughness degree of the polished surface of the optical fiber.
Feature importance ordering based on RF out-of-bag error estimation (OOB): the training set of each decision tree in the random forest is generated by bootstrap self-sampling, some samples may be sampled multiple times, some samples may not be sampled, the sampled samples are called in-bag samples accounting for about 2/3 of the total number, and the samples not sampled are called out-of-bag samples (OOB) accounting for about 1/3 of the total number, for calculating the importance of the feature. If there are k decision trees in RF, the importance of the feature s can be derived from the following procedure:
1) Constructing k decision trees;
2) Let k=1, obtain the corresponding out-of-bag sample OOB1;
3) Calculating a prediction error err of the current decision tree Tk to the OOB 1;
4) Randomly perturbing the value of the feature s in the OOB1 to be OOB1', and calculating a prediction error err ' of the current decision tree Tk on the OOB1 ';
5) Let k=2, 3,., K, repeat steps 1) to 4);
6) The importance of the feature s is calculated by the following formula:
If the change of the classification accuracy before and after the random disturbance of the feature s is not obvious, the feature s plays a small role in classification, and the importance is low; on the contrary, the importance is high.
The overlapping parts of the elements in the importance sequence 1 and the importance sequence 2 are selected as key indexes, the specific selection method can be designed according to the number of key indexes to be selected, if four key indexes are required to be selected, the first five elements in the importance sequence 1 and the importance sequence 2 can be selected respectively, the elements overlapping in the two groups are taken as key indexes, and in the embodiment, the key indexes comprise variance, second moment, entropy and contrast which are sequentially ordered from big importance to small.
In order to verify the validity of the four key indexes, the following comparative tests were performed in this embodiment:
Respectively constructing two types of index sets, and forming an index set 1 by variance, second moment, entropy and contrast; the index set 2 is composed of a second moment, entropy, contrast and correlation.
RF classification experiments were performed using 2 index sets for the roughness of the side polished fiber polished surface (sample No.1, spf1#) respectively, and the experimental results are shown in table 1 below:
TABLE 1 SPF1# polishing surface roughness RF classification experiment
As can be seen from the analysis of the above Table 1, the RF classification accuracy of the index set 1 is improved by 13.05% compared with that of the index set 2, so that the variance, the second moment, the entropy and the contrast in the GLCM texture characteristic parameters are more sensitive to classification and identification of different roughness of the polished surface of the optical fiber.
RF classification experiments were performed using 2 index sets for the roughness of the side polished fiber polished surface (sample No.2, spf2#) respectively, and the experimental results are shown in table 2 below:
TABLE 2 SPF2# polished surface roughness RF Classification verification experiment
Analyzing the above table 2, the RF classification accuracy of the index set 1 is improved by 8.69% compared with the index set 2, and higher classification recognition accuracy is shown. Compared with the index set 2, the index set 1 shows higher classification precision of the polishing surface roughness of the side polishing optical fiber for the samples 1 and 2, thereby proving the effectiveness and rationality of the selected key indexes.
Based on the determined key indexes, extracting specific numerical values of variance, second moment, entropy and contrast in an original image when the polishing surface roughness classification of the target optical fiber is carried out, and constructing a target optical fiber polishing surface texture feature subset.
S03, based on the extracted texture feature subset, carrying out random forest classification and identification on the polished surface of the target optical fiber to obtain the roughness grade of the target optical fiber, and calculating classification accuracy.
Specifically, in this embodiment, random Forest (RF) classification is adopted, when a random forest model is set, the output result of the random forest is set to be roughness grades with different sizes, which may be classified into one to five grades, one grade is the coarsest grade, and the five grades are the smoothest grade, and after the key index of the target optical fiber polishing surface image is extracted, the key index is input into the random forest model, and the output of the random forest model is the roughness grade classification result of the target optical fiber polishing surface.
After the roughness class classification result is obtained, the classification result can be counted, the classification precision is calculated, the accuracy of the classification result is verified, and the random forest model is adjusted, so that the classification precision is further improved. The specific calculation method of the classification precision is as follows:
TP (true) is defined, true is predicted as true number, TN (true negative), false is predicted as false number, FP (false positive), false is predicted as true number, FN (false negative), true is predicted as false number.
The classification accuracy ACC is:
Acc= (tp+tn)/(tp+tn+fp+fn), i.e.: the correct total number/total number of samples is classified.
The method is based on edge vision, the reflection direction of light can be changed by combining with rough textures of the polishing surface, so that when uniform light irradiates the optical fiber polishing surface, the intensity and distribution of the light entering the optical microscope camera sensor can carry information of the roughness of the polishing surface, namely, an acquired optical microscopic image of the polishing surface can carry information of the roughness of the polishing surface, and a random forest classification algorithm is combined, and the correlation between various indexes of the image and the roughness of the polishing surface of the optical fiber is constructed by utilizing the texture feature importance sorting, so that the texture features of the image of the polishing surface of the side polishing optical fiber can realize the representation of the information of the roughness of the polishing surface, and therefore, the method can realize the analysis of the roughness of the polishing surface based on the optical microscopic image of the light reflected by the polishing surface and has the advantages of low realization cost, simplicity in operation, high efficiency, accuracy in analysis and support of online measurement.
The embodiment also provides a side polishing optical fiber polishing surface roughness analysis system, which comprises:
the acquisition module is used for acquiring an optical microscopic image of the polished surface of the target optical fiber as an original image;
The extraction module is used for constructing a gray level co-occurrence matrix and extracting texture features in the original image through the gray level co-occurrence matrix;
And the classification module is used for classifying and identifying the target optical fiber based on the extracted texture characteristics to obtain the roughness grade of the target optical fiber.
Illustratively, as shown in fig. 2, the system includes a jig stage 1 for fixing and carrying an optical fiber, a light source 2, an acquisition unit 3, a processing unit 4, a first connection control unit 5A, a second connection control unit 5B, and an upper computer 6.
The detachable optical fiber clamp is arranged on the clamp table 1 for fixing and bearing the optical fibers and is used for clamping the optical fibers. The optical fiber clamp of the clamp table 1 can be replaced by an optical fiber clamp used for preparing the wheel type side polishing optical fiber, and the wheel type side polishing optical fiber device can be used for finishing side polishing of the optical fiber after replacement.
The light source 2 is arranged below the side polishing optical fiber polishing section and can move along the direction parallel to the side polishing optical fiber core so as to be convenient to adjust.
The acquisition unit 3 is a Fang Xianwei-mirror CCD arranged on the side polishing optical fiber polishing section and can move along the direction parallel to the side polishing optical fiber core so as to be convenient to adjust.
The processing unit 4 may be configured to implement the foregoing algorithm for analyzing the polishing surface roughness of the side polished optical fiber, including: and (3) image processing, extracting rough texture features of the side polishing optical fiber polishing surface, and sorting and optimizing different roughness classification identification of the side polishing optical fiber polishing surface based on the subset of the rough texture features of the side polishing optical fiber polishing surface. The second connection control unit 5B is connected with the upper computer 6 and used for storing the analysis result of the processing unit, and can be further fed back to the wheel type side polishing optical fiber device.
The first connection control unit 5A may be used to search and select the microscope CCD in the acquisition unit 3 connected to the current processing unit 4 through the Camera interface, and control the on and off of the CCD Camera.
The analysis system of the polishing surface roughness of the side polishing optical fiber in this embodiment and the foregoing analysis method belong to the same inventive concept, and can be understood with reference to the foregoing description, and will not be repeated here.
As shown in fig. 3, this embodiment further provides a computer device, which includes a processor 101 and a memory 102 connected by a bus signal, where at least one instruction or at least one program is stored in the memory 102, and the at least one instruction or the at least one program performs the method for analyzing the polishing surface roughness of the side polished optical fiber as described above when the at least one instruction or the at least one program is loaded by the processor 101. The memory 102 may be used to store software programs and modules, and the processor 101 executes various functional applications by running the software programs and modules stored in the memory 102. The memory 102 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, application programs required for functions, and the like; the storage data area may store data created according to the use of the device, etc. In addition, memory 102 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 102 may also include a memory controller to provide access to the memory 102 by the processor 101.
The method embodiments provided by the embodiments of the present disclosure may be performed in a computer terminal, a server, or a similar computing device, i.e., the above-described computer apparatus may include a computer terminal, a server, or a similar computing device. The internal structure of the computer device may include, but is not limited to: processor, network interface and memory. Wherein the processor, network interface, and memory within the computer device may be connected by a bus or other means.
The processor 101 (or CPU) is a computing core and a control core of a computer device. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WI-FI, mobile communication interface, etc.). Memory 102 (Memory) is a Memory device in a computer device for storing programs and data. It is understood that the memory 102 herein may be a high-speed RAM memory device or a non-volatile memory device (non-volatile memory), such as at least one magnetic disk memory device; optionally, at least one memory device located remotely from the aforementioned processor 101. The memory 102 provides storage space that stores an operating system of the electronic device, which may include, but is not limited to: windows (an operating system), linux (an operating system), android (an Android, a mobile operating system) system, IOS (a mobile operating system) system, etc., which are not limiting of the present disclosure; also stored in this memory space are one or more instructions, which may be one or more computer programs (including program code), adapted to be loaded and executed by the processor 101. In the embodiment of the present disclosure, the processor 101 loads and executes one or more instructions stored in the memory 102 to implement the method for analyzing the polishing surface roughness of the side polished optical fiber according to the embodiment of the method described above.
Embodiments of the present disclosure also provide a computer-readable storage medium having stored thereon at least one instruction or at least one program that, when loaded by the processor 101, performs a side polished fiber polished surface roughness analysis method as described above. The computer-readable storage medium carries one or more programs which, when executed, implement methods in accordance with embodiments of the present disclosure.
According to embodiments of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. Examples may include, but are not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
In the description of the present disclosure, it should be understood that the azimuth or positional relationships indicated by the azimuth terms such as "front, rear, upper, lower, left, right", "lateral, vertical, horizontal", and "top, bottom", etc., are generally based on the azimuth or positional relationships shown in the drawings, merely to facilitate description of the present disclosure and simplify the description, and without being otherwise described, these azimuth terms do not indicate and imply that the apparatus or elements referred to must have a specific azimuth or be configured and operated in a specific azimuth, and thus should not be construed as limiting the scope of protection of the present disclosure.
It will be apparent to those skilled in the art from this disclosure that various other changes and modifications can be made which are within the scope of the invention as defined in the claims.
Claims (10)
1. The method for analyzing the polishing surface roughness of the side polished optical fiber is characterized by comprising the following steps of:
S01, acquiring an optical microscopic image containing the polished surface reflection light of the target optical fiber as an original image;
S02, constructing a gray level co-occurrence matrix, and extracting texture features in the original image through the gray level co-occurrence matrix;
S03, classifying and identifying the polished surface of the target optical fiber based on the extracted texture features to obtain the roughness grade of the target optical fiber.
2. The method for analyzing the polishing surface roughness of the side polished optical fiber according to claim 1, wherein in step S02, constructing the gray level co-occurrence matrix includes: setting gray levelStep size/>And the direction of generation/>。
3. The method for analyzing the polishing surface roughness of a side polished optical fiber according to claim 2, wherein a gray level is setThe method comprises the following steps:
Pre-selection of With individual gray levels as alternative gray levels,/>Select/>Each of the candidate gray levels is constructed/>, respectivelyA first candidate gray level co-occurrence matrix;
calculating contrast ratio of each first alternative gray level co-occurrence matrix to texture characteristic parameters extracted from optical microscopic images of polished surfaces of polished optical fibers on the same side ，/>；
Acquiring operation time consumption of extracting texture features of optical microscopic images of polished surfaces of polished optical fibers on the same side by using each first alternative gray level co-occurrence matrix，/>；
The selection index of each candidate gray level is calculated according to the following formula：
；
Wherein,And/>Coefficients representing contrast and run time, respectively, satisfy/>，/>，
Selecting the indexThe candidate gray level with the largest value is taken as the gray level/>, of the gray level co-occurrence matrix。
4. The method for analyzing the polishing surface roughness of a side polished optical fiber according to claim 2, wherein a step size is setThe method comprises the following steps:
Pre-selecting a plurality of Step sizes as alternative step sizes,/>Select/>Constructing/>, respectively, the alternative step sizesA second alternative gray level co-occurrence matrix;
Calculating entropy value of texture characteristic parameters extracted from optical microscopic images of polished surfaces of polished optical fibers on the same side by using each second alternative gray level co-occurrence matrix ，/>；
Presetting a value of entropyReasonable interval/>And a recommendation step set, wherein the reasonable interval/>The recommended step length set comprises a plurality of polishing paper parameters adopted in the preparation of the side polishing optical fiber, and each polishing paper parameter has a corresponding recommended step length;
determining whether at least one entropy value exists Is located in the reasonable interval/>If yes, selecting a reasonable interval/>Inner and closest to/>Step length corresponding to entropy value of the gray level co-occurrence matrix is used as step length/>If not, acquiring polishing paper parameters adopted in preparation of the target optical fiber, and selecting a corresponding recommended step length as the step length/>, of the gray level co-occurrence matrix, according to the recommended step length set。
5. The method for analyzing the polishing surface roughness of a side polished optical fiber according to claim 2, wherein a generation direction is setThe method comprises the following steps:
Presetting the The generated direction angles are used as alternative angles, and/>Each of the alternative angles is constructed separately/>A third alternative gray level co-occurrence matrix;
Calculating the correlation of each third alternative gray level co-occurrence matrix to the texture characteristic parameters extracted from the optical microscopic image of the polished surface of the polished optical fiber on the same side ，/>；
Selecting the correlation thereofThe largest alternative angle is taken as the generation direction/>, of the gray level co-occurrence matrix。
6. The method for analyzing the roughness of a side polished optical fiber according to any one of claims 2 to 5, wherein in step S02, extracting texture features in the original image by the gray level co-occurrence matrix includes:
Calculating ranking indexes of the gray level co-occurrence matrix on texture feature parameters extracted from the polished surface image, wherein the ranking indexes comprise second moment, contrast, entropy, uniformity, correlation and variance, importance ranking is carried out on each ranking index by using at least two different ranking methods, an importance ranking result is obtained, and the overlapped part of each importance ranking result is selected as a key index;
and extracting various key indexes in the original image, and constructing a texture feature subset of the polished surface of the target optical fiber.
7. The method for analyzing the polishing surface roughness of the side polishing optical fiber according to claim 6, wherein the importance ranking method based on the random forest feature importance ranking method based on the kenel coefficient and the random forest feature importance ranking method based on the OOB error are used for ranking the importance of the ranking indexes;
The key indexes comprise variance, second moment, entropy and contrast which are sequenced in sequence from big to small according to importance.
8. A side polished fiber polished surface roughness analysis system, comprising:
the acquisition module is used for acquiring an optical microscopic image of the polished surface of the target optical fiber as an original image;
The extraction module is used for constructing a gray level co-occurrence matrix and extracting texture features in the original image through the gray level co-occurrence matrix;
And the classification module is used for classifying and identifying the target optical fiber based on the extracted texture characteristics to obtain the roughness grade of the target optical fiber.
9. A computer device comprising a processor and a memory in signal connection, wherein the memory has stored therein at least one instruction or at least one program that, when loaded by the processor, performs the method of side polished fiber polished surface roughness analysis of any of claims 1-7.
10. A computer readable storage medium having stored thereon at least one instruction or at least one program, wherein the at least one instruction or the at least one program when loaded by a processor performs the method of side polished fiber polished surface roughness analysis as claimed in any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410324286.2A CN117934461A (en) | 2024-03-21 | 2024-03-21 | Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410324286.2A CN117934461A (en) | 2024-03-21 | 2024-03-21 | Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117934461A true CN117934461A (en) | 2024-04-26 |
Family
ID=90756040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410324286.2A Pending CN117934461A (en) | 2024-03-21 | 2024-03-21 | Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117934461A (en) |
-
2024
- 2024-03-21 CN CN202410324286.2A patent/CN117934461A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111814385B (en) | Method, device and computer equipment for predicting quality of machined part | |
| CN110402487B (en) | Method and apparatus for classifying defects detected from a wafer | |
| US20130077092A1 (en) | Substrate surface defect inspection method and inspection device | |
| CN116188475B (en) | Intelligent control method, system and medium for automatic optical detection of appearance defects | |
| JP6005498B2 (en) | Lens processing system, tool change time detection method, and spectacle lens manufacturing method | |
| CN113487533B (en) | Part assembly quality digital detection system and method based on machine learning | |
| CN115825103B (en) | End detection system for angular port optical fiber array based on image technology | |
| CN111598827A (en) | Appearance flaw detection method, electronic device and storage medium | |
| CN116342597B (en) | Method and system for detecting electroplating processing defects on surface of automobile part | |
| CN104318249A (en) | Grinding surface roughness grade identification method based on multi-classification support vector machine | |
| KR20220117328A (en) | A metal texture phase classification method, a metal texture phase classification apparatus, a metal texture phase learning method, a metal texture phase learning apparatus, a material property prediction method of a metal material, and a material property prediction apparatus of a metal material | |
| CN113034482A (en) | Surface roughness detection method based on machine vision and machine learning | |
| AU644164B2 (en) | Method for determining and controlling fiber luster properties | |
| CN117934461A (en) | Method, system and equipment for analyzing polishing surface roughness of side polishing optical fiber | |
| KR101886570B1 (en) | Apparatus and method for managing part database for supporting 3d printing-based maintenance, and 3d printing-based part maintenance system | |
| CN116071348B (en) | Workpiece surface detection method and related device based on visual detection | |
| US20220260489A1 (en) | System and method for analyzing molecular interactions on living cells using biosensor techniques | |
| CN116295141A (en) | Fuel cell bipolar plate flatness detection method and system | |
| CN115239663A (en) | Method and system for detecting defects of contact lens, electronic device and storage medium | |
| CN115171823A (en) | Material optimization method and device based on resin lens collision detection | |
| CN115398228A (en) | Method for determining imaging condition of metal structure, method for imaging metal structure, method for classifying phase of metal structure, device for determining imaging condition of metal structure, device for imaging metal structure, device for classifying phase of metal structure, method for predicting material property of metal material, and device for predicting material property of metal material | |
| TW202208839A (en) | Automated visual-inspection system | |
| CN117889774A (en) | Real point cloud noise reduction method and device, electronic equipment and medium | |
| JP4235074B2 (en) | Pass / fail judgment device, pass / fail judgment program, and pass / fail judgment method | |
| Wang et al. | Complicated intermittent scratches detection research on surface of optical components based on adaptive sector scanning algorithm cascading mean variance threshold algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |