CN116682072A - Bridge disease monitoring system - Google Patents
Bridge disease monitoring system Download PDFInfo
- Publication number
- CN116682072A CN116682072A CN202310977695.8A CN202310977695A CN116682072A CN 116682072 A CN116682072 A CN 116682072A CN 202310977695 A CN202310977695 A CN 202310977695A CN 116682072 A CN116682072 A CN 116682072A
- Authority
- CN
- China
- Prior art keywords
- layer
- input end
- convolution layer
- output
- output end
- 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.)
- Granted
Links
- 201000010099 disease Diseases 0.000 title claims abstract description 41
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 41
- 238000012544 monitoring process Methods 0.000 title claims abstract description 24
- 238000001914 filtration Methods 0.000 claims abstract description 41
- 230000032683 aging Effects 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 16
- 230000002159 abnormal effect Effects 0.000 claims description 41
- 238000012549 training Methods 0.000 claims description 20
- 238000012216 screening Methods 0.000 claims description 18
- 230000007547 defect Effects 0.000 claims description 16
- 238000012937 correction Methods 0.000 claims description 14
- 238000000034 method Methods 0.000 claims description 10
- 238000010586 diagram Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 description 8
- 238000006243 chemical reaction Methods 0.000 description 7
- 230000009286 beneficial effect Effects 0.000 description 6
- 230000002093 peripheral effect Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000035882 stress Effects 0.000 description 3
- 229910010293 ceramic material Inorganic materials 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000000547 structure data Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/36—Applying a local operator, i.e. means to operate on image points situated in the vicinity of a given point; Non-linear local filtering operations, e.g. median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/54—Extraction of image or video features relating to texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/176—Urban or other man-made structures
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Nonlinear Science (AREA)
- Medical Informatics (AREA)
- Image Processing (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a bridge disease monitoring system, which belongs to the technical field of image processing, wherein a filtering unit is adopted to filter noise data in bridge images, and the filtering images are aged and corrected according to aging coefficients of the bridges, so that the influence of service years and noise data is removed, and the recognition precision is improved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a bridge disease monitoring system.
Background
When the bridge is used, the road and the bridge can cause excessive bridge bearing due to factors such as over-high bearing capacity, over-high externally applied load force and the like, so that cracks are generated. Meanwhile, the season changes, the expansion caused by heat and contraction caused by cold causes the inside of the concrete to deform to different degrees, and certain stress appears in the internal structure of the concrete under the influence of deformation constraint force, and when the stress is larger than the stress born by the concrete, cracks can be caused.
According to the existing bridge disease monitoring scheme, the YOLO and VGGNet neural network are adopted to extract characteristics of bridge images and classify the bridge images to obtain bridge disease conditions, but the YOLO and VGGNet neural network both comprise a large number of convolution pooling layers, the calculation complexity is high, the memory consumption of a computer is high, the bridge images are directly processed, a large amount of data in the images are normal structure data, a small part of abnormal disease data cannot be fully expressed in the neural network, and therefore the accuracy of disease monitoring is low.
Disclosure of Invention
Aiming at the defects in the prior art, the bridge disease monitoring system provided by the invention solves the problems of higher computational complexity and low disease monitoring accuracy in the existing bridge disease monitoring scheme.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: a bridge defect monitoring system comprising: the device comprises a bridge image acquisition unit, a filtering unit, an ageing correction unit, an abnormal data screening unit and a disease area identification unit;
the bridge image acquisition unit is used for acquiring the bridge image to obtain a bridge image; the filtering unit is used for filtering the bridge image to obtain a filtered image; the aging correction unit is used for performing aging correction on the filtered image according to the aging coefficient to obtain a corrected image; the abnormal data screening unit is used for screening abnormal data from the corrected image; the disease area identification unit is used for carrying out target identification on the abnormal data to obtain the disease type.
Further, the specific filtering process of the filtering unit includes: setting upTaking the central pixel point under the filter window as the pixel point to be filtered, calculating the distance characteristic value of the pixel point to be filtered and other pixel points in the filter window, and carrying out filter processing on the pixel point to be filtered according to the distance characteristic value and the filter threshold value to obtain the filter pixel point.
Further, the expression of the distance characteristic value is:
,
wherein ,for distance characteristic value, ++>For the pixel value of the pixel to be filtered, < >>Dividing pixel values for a filter windowExternal->A pixel value.
Further, the expression of the filtering process is:
,
wherein ,for filtering the pixel values of the pixel points, < >>Is a filtering threshold.
The beneficial effects of the above further scheme are: according to the invention, the filtering threshold value is set, and when the difference value between the distance characteristic value and the filtering threshold value is larger, the difference between the pixel values of the pixel points to be filtered and the pixel values of the peripheral pixel points is larger, so that the pixel points are proved to be abnormal points, the pixel values of the peripheral pixel points are emphasized to be taken for optimizing the pixel values of the pixel points to be filtered, meanwhile, the filtering threshold value is set, and the texture characteristic can be still reserved for the texture pixel points, so that the characteristic of a lesion area can be still reserved and cannot be filtered in the filtering process.
Further, the expression of the aging correction unit is:
,
wherein ,to correct the%>Pixel value of each pixel, +.>For filtering the picture->Pixel value of each pixel, +.>Is the aging coefficient.
Further, the abnormal data screening unit includes: the device comprises a gray level processing module, a space transformation module, an abnormal frequency screening module and a reconstruction module;
the gray processing module is used for carrying out gray processing on the corrected image to obtain a gray image;
the space transformation module is used for converting the real space of the gray level diagram into the frequency space to obtain frequency space data;
the abnormal frequency screening module is used for subtracting the frequency space data from the stored standard frequency space data to obtain abnormal frequency space data;
the reconstruction module is used for reconstructing the abnormal frequency space data to obtain abnormal data.
The beneficial effects of the above further scheme are: according to the invention, a gray level diagram is obtained through gray level processing, so that real-frequency conversion is facilitated, real space to frequency space conversion is carried out on the gray level diagram, frequency space data is obtained, the frequency space data represents data components, the frequency space data is subtracted from stored standard frequency space data, abnormal component data is obtained, the abnormal component data is reconstructed, and the frequency space is converted into real space, so that abnormal data is obtained.
Further, the disease area identifying unit includes: the weight output module comprises a first convolution layer, a second convolution layer, a third convolution layer, a fourth convolution layer, a fifth convolution layer, a weight output module, a first upsampling layer, a first Concat layer, a multiplier M1, a dropout layer and a softmax layer;
the input end of the first convolution layer is used as the input end of the disease area identification unit, and the output end of the first convolution layer is respectively connected with the input end of the second convolution layer and the first input end of the weight output module; the output end of the second convolution layer is respectively connected with the input end of the third convolution layer and the second input end of the weight output module; the input end of the fourth convolution layer is connected with the output end of the third convolution layer, and the output end of the fourth convolution layer is respectively connected with the first input end of the first Concat layer and the input end of the fifth convolution layer; the output end of the fifth convolution layer is connected with the input end of the first up-sampling layer; the second input end of the first Concat layer is connected with the output end of the first up-sampling layer, and the output end of the first Concat layer is connected with the first input end of the multiplier M1; the second input end of the multiplier M1 is connected with the output end of the weight output module, and the output end of the multiplier M1 is connected with the input end of the dropout layer; the input end of the softmax layer is connected with the output end of the dropout layer, and the output end of the softmax layer is used as the output end of the disease area identification unit.
The beneficial effects of the above further scheme are: in the invention, the characteristics of abnormal data are extracted layer by utilizing multilayer convolution, the shallow characteristics are input into the weight output module, the self-adaptive weight is obtained according to the shallow characteristics, and the deep characteristics are multiplied by the weight, so that the self-adaptive attention is applied, and the attention of the characteristics is enhanced.
Further, the weight output module includes: a second upsampling layer, a second Concat layer, a sixth convolution layer, a seventh convolution layer, and a weight layer;
the input end of the second upsampling layer is used as a second input end of the weight output module; the first input end of the second Concat layer is used as the first input end of the weight output module, the second input end of the second Concat layer is connected with the output end of the second upsampling layer, and the output end of the second Concat layer is connected with the input end of the sixth convolution layer; the input end of the seventh convolution layer is connected with the output end of the sixth convolution layer, and the output end of the seventh convolution layer is connected with the input end of the weight layer; the output end of the weight layer is used as the output end of the weight output module.
Further, the expression of the weight layer is:
,
wherein ,for the input of weight layersGo out (I)>As an arctangent function, +.>A seventh convolution layer +.>Output characteristics,/->The number of features is output for the seventh convolutional layer.
The beneficial effects of the above further scheme are: according to the invention, the weight which should be applied by the multiplier M1 is calculated by utilizing the condition of the output characteristics of the seventh convolution layer, so that the aim of realizing different attention according to different inputs is fulfilled, and the accuracy of monitoring the disease area is improved.
Further, the loss function of the disease area identifying unit during training is as follows:
,
,
wherein ,for loss function->For the total number of counted neighbor exercises, +.>Is->Predictive value output by disease area identification unit during secondary training, < >>Is->Label value corresponding to training time, +.>Is a natural constant which is used for the production of the high-temperature-resistant ceramic material,is a proportional coefficient->For the number of the current training times>Is the number of adjacent training times.
The beneficial effects of the above further scheme are: the proportionality coefficient of the invention changes along with the difference value of the predicted value and the label value, and when the difference value is larger, the proportionality coefficient is smaller, and the exponential function is thatThe value is more, and the person is added with the herb>The value is less, the rapid iteration of parameters in the disease area identification unit is realized, the training speed is accelerated, and when the difference value is smaller, the proportionality coefficient is larger,/-in>The value is more, and the person is added with the herb>The invention has the advantages of low value, slow conversion of parameters in the disease area identification unit, gradual finding of a better value, and self-adaptive adjustment of training degrees in different training periods compared with the existing sectional loss function.
The beneficial effects of the invention are as follows: according to the invention, the noise data in the bridge image is filtered by the filtering unit, the filtering image is subjected to ageing correction according to the ageing coefficient of the bridge, so that the influence of service years and the noise data is removed, and the identification precision is improved.
Drawings
FIG. 1 is a system block diagram of a bridge defect monitoring system;
fig. 2 is a schematic structural view of a disease area recognition unit.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, a bridge defect monitoring system includes: the device comprises a bridge image acquisition unit, a filtering unit, an ageing correction unit, an abnormal data screening unit and a disease area identification unit;
the bridge image acquisition unit is used for acquiring the bridge image to obtain a bridge image; the filtering unit is used for filtering the bridge image to obtain a filtered image; the aging correction unit is used for performing aging correction on the filtered image according to the aging coefficient to obtain a corrected image; the abnormal data screening unit is used for screening abnormal data from the corrected image; the disease area identification unit is used for carrying out target identification on the abnormal data to obtain the disease type.
The specific filtering process of the filtering unit comprises the following steps: setting upTaking the central pixel point under the filter window as the pixel point to be filtered, calculating the distance characteristic value of the pixel point to be filtered and other pixel points in the filter window, and carrying out filter processing on the pixel point to be filtered according to the distance characteristic value and the filter threshold value to obtain the filter pixel point.
The expression of the distance characteristic value is as follows:
,
wherein ,for distance characteristic value, ++>For the pixel value of the pixel to be filtered, < >>Dividing pixel values for a filter windowExternal->A pixel value.
The expression of the filtering process is as follows:
,
wherein ,for filtering the pixel values of the pixel points, < >>Is a filtering threshold.
According to the invention, the filtering threshold value is set, and when the difference value between the distance characteristic value and the filtering threshold value is larger, the difference between the pixel values of the pixel points to be filtered and the pixel values of the peripheral pixel points is larger, so that the pixel points are proved to be abnormal points, the pixel values of the peripheral pixel points are emphasized to be taken for optimizing the pixel values of the pixel points to be filtered, meanwhile, the filtering threshold value is set, and the texture characteristic can be still reserved for the texture pixel points, so that the characteristic of a lesion area can be still reserved and cannot be filtered in the filtering process.
The expression of the aging correction unit is as follows:
,
wherein ,to correct the%>Pixel value of each pixel, +.>For filtering the picture->Pixel value of each pixel, +.>Is the aging coefficient.
The specific acquisition method of the aging coefficient comprises the following steps: bridge images at different use time points are acquired, the change condition of the pixel values at different time points is determined according to the pixel values of the bridge images at different use time points, and the least square method can be adopted for fitting according to the change rule to obtain an aging coefficient. The aging factor is used to reduce the effect of age on pixel values.
The abnormal data screening unit includes: the device comprises a gray level processing module, a space transformation module, an abnormal frequency screening module and a reconstruction module;
the gray processing module is used for carrying out gray processing on the corrected image to obtain a gray image;
the space transformation module is used for converting the real space of the gray level diagram into the frequency space to obtain frequency space data;
the abnormal frequency screening module is used for subtracting the frequency space data from the stored standard frequency space data to obtain abnormal frequency space data;
the reconstruction module is used for reconstructing the abnormal frequency space data to obtain abnormal data.
According to the invention, a gray level diagram is obtained through gray level processing, so that real-frequency conversion is facilitated, real space to frequency space conversion is carried out on the gray level diagram, frequency space data is obtained, the frequency space data represents data components, the frequency space data is subtracted from stored standard frequency space data, abnormal component data is obtained, the abnormal component data is reconstructed, and the frequency space is converted into real space, so that abnormal data is obtained.
In the invention, the stored standard frequency space data is the data after the real-frequency conversion of the disease-free bridge image.
As shown in fig. 2, the disease area identifying unit includes: the weight output module comprises a first convolution layer, a second convolution layer, a third convolution layer, a fourth convolution layer, a fifth convolution layer, a weight output module, a first upsampling layer, a first Concat layer, a multiplier M1, a dropout layer and a softmax layer;
the input end of the first convolution layer is used as the input end of the disease area identification unit, and the output end of the first convolution layer is respectively connected with the input end of the second convolution layer and the first input end of the weight output module; the output end of the second convolution layer is respectively connected with the input end of the third convolution layer and the second input end of the weight output module; the input end of the fourth convolution layer is connected with the output end of the third convolution layer, and the output end of the fourth convolution layer is respectively connected with the first input end of the first Concat layer and the input end of the fifth convolution layer; the output end of the fifth convolution layer is connected with the input end of the first up-sampling layer; the second input end of the first Concat layer is connected with the output end of the first up-sampling layer, and the output end of the first Concat layer is connected with the first input end of the multiplier M1; the second input end of the multiplier M1 is connected with the output end of the weight output module, and the output end of the multiplier M1 is connected with the input end of the dropout layer; the input end of the softmax layer is connected with the output end of the dropout layer, and the output end of the softmax layer is used as the output end of the disease area identification unit.
In the invention, the characteristics of abnormal data are extracted layer by utilizing multilayer convolution, the shallow characteristics are input into the weight output module, the self-adaptive weight is obtained according to the shallow characteristics, and the deep characteristics are multiplied by the weight, so that the self-adaptive attention is applied, and the attention of the characteristics is enhanced.
The weight output module comprises: a second upsampling layer, a second Concat layer, a sixth convolution layer, a seventh convolution layer, and a weight layer;
the input end of the second upsampling layer is used as a second input end of the weight output module; the first input end of the second Concat layer is used as the first input end of the weight output module, the second input end of the second Concat layer is connected with the output end of the second upsampling layer, and the output end of the second Concat layer is connected with the input end of the sixth convolution layer; the input end of the seventh convolution layer is connected with the output end of the sixth convolution layer, and the output end of the seventh convolution layer is connected with the input end of the weight layer; the output end of the weight layer is used as the output end of the weight output module.
The expression of the weight layer is as follows:
,
wherein ,for the output of the weight layer, +.>As an arctangent function, +.>A seventh convolution layer +.>Output characteristics,/->The number of features is output for the seventh convolutional layer.
According to the invention, the weight which should be applied by the multiplier M1 is calculated by utilizing the condition of the output characteristics of the seventh convolution layer, so that the aim of realizing different attention according to different inputs is fulfilled, and the accuracy of monitoring the disease area is improved.
The loss function of the disease area identification unit during training is as follows:
,
,
wherein ,for loss function->For the total number of counted neighbor exercises, +.>Is->Predictive value output by disease area identification unit during secondary training, < >>Is->Label value corresponding to training time, +.>Is a natural constant which is used for the production of the high-temperature-resistant ceramic material,is a proportional coefficient->For the number of the current training times>Is the number of adjacent training times.
The proportionality coefficient of the invention changes along with the difference value of the predicted value and the label value, and when the difference value is larger, the proportionality coefficient is smaller, and the exponential function is thatThe value is more, and the person is added with the herb>The value is less, the rapid iteration of parameters in the disease area identification unit is realized, the training speed is accelerated, and when the difference value is smaller, the proportionality coefficient is larger,/-in>The value is more, and the person is added with the herb>The invention has the advantages of low value, slow conversion of parameters in the disease area identification unit, gradual finding of a better value, and self-adaptive adjustment of training degrees in different training periods compared with the existing sectional loss function.
According to the invention, the noise data in the bridge image is filtered by the filtering unit, the filtering image is subjected to ageing correction according to the ageing coefficient of the bridge, so that the influence of service years and the noise data is removed, and the identification precision is improved.
Claims (10)
1. A bridge defect monitoring system, comprising: the device comprises a bridge image acquisition unit, a filtering unit, an ageing correction unit, an abnormal data screening unit and a disease area identification unit;
the bridge image acquisition unit is used for acquiring the bridge image to obtain a bridge image; the filtering unit is used for filtering the bridge image to obtain a filtered image; the aging correction unit is used for performing aging correction on the filtered image according to the aging coefficient to obtain a corrected image; the abnormal data screening unit is used for screening abnormal data from the corrected image; the disease area identification unit is used for carrying out target identification on the abnormal data to obtain the disease type.
2. The bridge defect monitoring system according to claim 1, wherein the specific filtering process of the filtering unit comprises: setting upTaking the central pixel point under the filter window as the pixel point to be filtered, calculating the distance characteristic value of the pixel point to be filtered and other pixel points in the filter window, and carrying out filter processing on the pixel point to be filtered according to the distance characteristic value and the filter threshold value to obtain the filter pixel point.
3. The bridge defect monitoring system of claim 2, wherein the distance eigenvalue is expressed as:
, wherein ,/>For distance characteristic value, ++>For the pixel value of the pixel to be filtered, < >>Divide the pixel value for the filter window +.>External->A pixel value.
4. A bridge defect monitoring system according to claim 3, wherein the expression of the filtering process is:
, wherein ,/>For filtering the pixel values of the pixel points, < >>Is a filtering threshold.
5. The bridge defect monitoring system according to claim 1, wherein the expression of the aging correction unit is:
, wherein ,/>To correct the%>Pixel value of each pixel, +.>For filtering the picture->Pixel value of each pixel, +.>Is the aging coefficient.
6. The bridge defect monitoring system according to claim 1, wherein the abnormal data screening unit comprises: the device comprises a gray level processing module, a space transformation module, an abnormal frequency screening module and a reconstruction module;
the gray processing module is used for carrying out gray processing on the corrected image to obtain a gray image;
the space transformation module is used for converting the real space of the gray level diagram into the frequency space to obtain frequency space data;
the abnormal frequency screening module is used for subtracting the frequency space data from the stored standard frequency space data to obtain abnormal frequency space data;
the reconstruction module is used for reconstructing the abnormal frequency space data to obtain abnormal data.
7. The bridge defect monitoring system according to claim 1, wherein the defect area identifying unit includes: the weight output module comprises a first convolution layer, a second convolution layer, a third convolution layer, a fourth convolution layer, a fifth convolution layer, a weight output module, a first upsampling layer, a first Concat layer, a multiplier M1, a dropout layer and a softmax layer;
the input end of the first convolution layer is used as the input end of the disease area identification unit, and the output end of the first convolution layer is respectively connected with the input end of the second convolution layer and the first input end of the weight output module; the output end of the second convolution layer is respectively connected with the input end of the third convolution layer and the second input end of the weight output module; the input end of the fourth convolution layer is connected with the output end of the third convolution layer, and the output end of the fourth convolution layer is respectively connected with the first input end of the first Concat layer and the input end of the fifth convolution layer; the output end of the fifth convolution layer is connected with the input end of the first up-sampling layer; the second input end of the first Concat layer is connected with the output end of the first up-sampling layer, and the output end of the first Concat layer is connected with the first input end of the multiplier M1; the second input end of the multiplier M1 is connected with the output end of the weight output module, and the output end of the multiplier M1 is connected with the input end of the dropout layer; the input end of the softmax layer is connected with the output end of the dropout layer, and the output end of the softmax layer is used as the output end of the disease area identification unit.
8. The bridge defect monitoring system of claim 7, wherein the weight output module comprises: a second upsampling layer, a second Concat layer, a sixth convolution layer, a seventh convolution layer, and a weight layer;
the input end of the second upsampling layer is used as a second input end of the weight output module; the first input end of the second Concat layer is used as the first input end of the weight output module, the second input end of the second Concat layer is connected with the output end of the second upsampling layer, and the output end of the second Concat layer is connected with the input end of the sixth convolution layer; the input end of the seventh convolution layer is connected with the output end of the sixth convolution layer, and the output end of the seventh convolution layer is connected with the input end of the weight layer; the output end of the weight layer is used as the output end of the weight output module.
9. The bridge defect monitoring system of claim 8, wherein the weight layer has the expression:
, wherein ,/>For the output of the weight layer, +.>As an arctangent function, +.>A seventh convolution layer +.>Output characteristics,/->The number of features is output for the seventh convolutional layer.
10. The bridge defect monitoring system of claim 1, wherein the defect area recognition unit is trained to have a loss function of:
,
, wherein ,/>For loss function->For the total number of counted neighbor exercises, +.>Is->Predictive value output by disease area identification unit during secondary training, < >>Is->Label value corresponding to training time, +.>Is natural constant (18)>Is a proportional coefficient->For the number of the current training times>Is the number of adjacent training times.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310977695.8A CN116682072B (en) | 2023-08-04 | 2023-08-04 | Bridge disease monitoring system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310977695.8A CN116682072B (en) | 2023-08-04 | 2023-08-04 | Bridge disease monitoring system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116682072A true CN116682072A (en) | 2023-09-01 |
CN116682072B CN116682072B (en) | 2023-10-20 |
Family
ID=87784098
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310977695.8A Active CN116682072B (en) | 2023-08-04 | 2023-08-04 | Bridge disease monitoring system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116682072B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006287623A (en) * | 2005-03-31 | 2006-10-19 | Sharp Corp | Input image correction device, method, and program, image processor, image formation apparatus, and recording medium |
KR20150109083A (en) * | 2014-03-19 | 2015-10-01 | 국방과학연구소 | Method of Distance Adaptive Detection Threshold Value for Infrared Target Detection System |
WO2017120796A1 (en) * | 2016-01-13 | 2017-07-20 | 富士通株式会社 | Pavement distress detection method and apparatus, and electronic device |
CN110070008A (en) * | 2019-04-04 | 2019-07-30 | 中设设计集团股份有限公司 | Bridge disease identification method adopting unmanned aerial vehicle image |
CN111401358A (en) * | 2020-02-25 | 2020-07-10 | 华南理工大学 | Instrument dial plate correction method based on neural network |
CN111985499A (en) * | 2020-07-23 | 2020-11-24 | 东南大学 | High-precision bridge apparent disease identification method based on computer vision |
CN113689396A (en) * | 2021-08-20 | 2021-11-23 | 深圳先进技术研究院 | Cell fluorescence image thresholding method, system, terminal and storage medium |
CN115713488A (en) * | 2022-10-27 | 2023-02-24 | 浙江大学 | Bridge apparent disease pixel level identification method and system based on instance segmentation |
CN115995056A (en) * | 2023-03-22 | 2023-04-21 | 南京航空航天大学 | Automatic bridge disease identification method based on deep learning |
-
2023
- 2023-08-04 CN CN202310977695.8A patent/CN116682072B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006287623A (en) * | 2005-03-31 | 2006-10-19 | Sharp Corp | Input image correction device, method, and program, image processor, image formation apparatus, and recording medium |
KR20150109083A (en) * | 2014-03-19 | 2015-10-01 | 국방과학연구소 | Method of Distance Adaptive Detection Threshold Value for Infrared Target Detection System |
WO2017120796A1 (en) * | 2016-01-13 | 2017-07-20 | 富士通株式会社 | Pavement distress detection method and apparatus, and electronic device |
CN110070008A (en) * | 2019-04-04 | 2019-07-30 | 中设设计集团股份有限公司 | Bridge disease identification method adopting unmanned aerial vehicle image |
CN111401358A (en) * | 2020-02-25 | 2020-07-10 | 华南理工大学 | Instrument dial plate correction method based on neural network |
CN111985499A (en) * | 2020-07-23 | 2020-11-24 | 东南大学 | High-precision bridge apparent disease identification method based on computer vision |
CN113689396A (en) * | 2021-08-20 | 2021-11-23 | 深圳先进技术研究院 | Cell fluorescence image thresholding method, system, terminal and storage medium |
CN115713488A (en) * | 2022-10-27 | 2023-02-24 | 浙江大学 | Bridge apparent disease pixel level identification method and system based on instance segmentation |
CN115995056A (en) * | 2023-03-22 | 2023-04-21 | 南京航空航天大学 | Automatic bridge disease identification method based on deep learning |
Non-Patent Citations (4)
Title |
---|
ZHANG LINGXIN 等: "A review of the research and application of deep learning-based computer vision in structural damage detection", 《EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION》, pages 1 - 21 * |
张宁 等: "基于机器视觉的送纱管同轴度检测研究", 《机械设计与制造》, pages 1 - 6 * |
张玥 等: "基于树莓派的智能监控系统设计与实现", 《计算机测量与控制》, vol. 31, no. 8, pages 122 - 128 * |
徐发东: "基于图像处理的桥梁表面裂缝病害检测的研究", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅱ辑)》, pages 034 - 908 * |
Also Published As
Publication number | Publication date |
---|---|
CN116682072B (en) | 2023-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2019080367A1 (en) | Method for evaluating health status of mechanical device | |
CN111429403B (en) | Automobile gear finished product defect detection method based on machine vision | |
CN111007566B (en) | Curvature-driven diffusion full-convolution network seismic data bad channel reconstruction and denoising method | |
CN111044570A (en) | Defect identification and early warning method and device for power equipment and computer equipment | |
CN110119677B (en) | Image classification network-based carbon fiber composite core cable damage detection method | |
CN112305379A (en) | Mode identification method and system for GIS insulation defect | |
CN116842364A (en) | Rolling bearing fault method for asynchronous motor | |
CN116682072B (en) | Bridge disease monitoring system | |
CN114820438A (en) | Method and system for detecting surface material defects of silencer | |
CN114724045A (en) | Method for inverting submarine topography of shallow sea on land | |
CN115290326A (en) | Rolling bearing fault intelligent diagnosis method | |
CN116612103B (en) | Intelligent detection method and system for building structure cracks based on machine vision | |
CN117272753A (en) | Power transmission line icing galloping early warning method based on process parameters | |
CN116012687A (en) | Image interaction fusion method for identifying tread defects of wheel set | |
CN117079005A (en) | Optical cable fault monitoring method, system, device and readable storage medium | |
CN115099291B (en) | Building energy-saving monitoring method | |
CN115410003A (en) | Power distribution room temperature early warning method, system and medium based on deep learning and infrared detection | |
CN112435219B (en) | Pavement crack identification method based on transposition neural network interlayer feature fusion | |
CN115144211A (en) | Fault detection method and device based on discrete wavelet transform and gated cyclic unit | |
CN115660140A (en) | Multi-mode set climate prediction method based on artificial intelligence correction | |
CN113192018B (en) | Water-cooled wall surface defect video identification method based on fast segmentation convolutional neural network | |
CN114563671A (en) | High-voltage cable partial discharge diagnosis method based on CNN-LSTM-Attention neural network | |
CN113723210A (en) | Method and system for detecting leakage edge equipment of water supply pipe in intelligent park | |
CN114240948B (en) | Intelligent segmentation method and system for structural surface damage image | |
CN111754426B (en) | Automatic restoration method for mural shedding disease based on genetic algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |