CN115346127A - Dam safety detection method and system - Google Patents
Dam safety detection method and system Download PDFInfo
- Publication number
- CN115346127A CN115346127A CN202211283181.4A CN202211283181A CN115346127A CN 115346127 A CN115346127 A CN 115346127A CN 202211283181 A CN202211283181 A CN 202211283181A CN 115346127 A CN115346127 A CN 115346127A
- Authority
- CN
- China
- Prior art keywords
- image
- dam
- gray
- blocks
- gray level
- 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
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- 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
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a dam safety detection method and a dam safety detection system, belongs to the technical field of image processing, and aims to avoid the situation that radar equipment is used for transmitting electromagnetic waves to realize dam detection through image acquisition, avoid direct processing of dam original images by extracting image characteristics of dam images and processing the image characteristics through a crack recognition model, and improve recognition accuracy.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a dam safety detection method and a dam safety detection system.
Background
The dam is exposed in the air, bears the temperature change, the water pressure, the water scouring, the erosion and the like all the year round, is easy to crack or break, has the problems of dam body deformation, landslide and the like, influences the operation of the dam and also causes potential safety hazards, and therefore the dam needs to be frequently detected, and the stability of the dam structure is guaranteed.
The dam is patrolled by adopting a manual patrolling mode, a certain monitoring effect can be achieved, but for people, a large amount of labor is consumed, negligence easily exists, and the dam cannot be noticed all the time.
The dam is detected by the radar equipment at present, electromagnetic waves are emitted, dam multi-section gradient data are obtained by feeding back electromagnetic wave signals, and dam multi-section gradient data are processed by the processor, so that the dam is detected.
In the prior art, deep learning algorithms such as a neural network are adopted to process dam images, but the difference of image imaging can be caused due to the fact that different equipment are different in shooting weather environments, and therefore the identification precision is low.
Disclosure of Invention
Aiming at the defects in the prior art, the dam safety detection method and the dam safety detection system provided by the invention have the advantages that the dam detection is realized by avoiding the mode of transmitting electromagnetic waves by using radar equipment through image acquisition, and meanwhile, the problem of low identification precision caused by image imaging difference due to different shooting weather environments and different equipment is solved.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that: a dam safety detection method comprises the following steps:
s1, collecting dam images;
s2, extracting image characteristics of the dam image;
and S3, processing the image characteristics by adopting a crack identification model to obtain a crack area of the dam.
The method has the beneficial effects that: according to the dam image detection method, the dam detection is realized in a mode of avoiding using radar equipment to emit electromagnetic waves through image acquisition, image features are extracted from dam images, and then the image features are processed through a crack recognition model, so that the dam original images are prevented from being directly processed, and the recognition accuracy is improved.
Further, the step S2 includes the following sub-steps:
s21, clipping the dam image to obtain a dam building architecture image;
s22, carrying out gray level processing on the dam building architecture image to obtain a dam building architecture gray level image;
s23, aligning the dam building framework gray level image with a pre-stored dam gray level image according to the structural characteristics of the dam;
and S34, screening out image characteristics according to the aligned dam building framework gray level image and the pre-stored dam gray level image.
The beneficial effects of the above further scheme are: the method comprises the steps of cutting a dam image, extracting a dam building framework image, carrying out gray level processing on the dam building framework image to reduce data volume, aligning the dam building framework gray level image with a prestored dam gray level image, and facilitating subsequent same-region blocking after aligning.
Further, the step S34 includes the following sub-steps:
s341, dividing the dam building framework gray level image and the pre-stored dam gray level image into a plurality of image blocks respectively;
s342, forming image block pairs by image blocks of the dam building framework gray level image at the same position and image blocks of a pre-stored dam gray level image;
s343, taking a pixel point of each image block in the image block pair, and calculating a gray coefficient based on the gray value of the pixel point;
s345, correcting all image blocks of the dam building framework gray level image according to the normal gray level coefficient to obtain corrected image blocks;
and S346, screening the corrected image blocks with abnormal gray coefficients as image features.
The beneficial effects of the above further scheme are: the dam building framework gray image is aligned with the pre-stored dam gray image, so that the image block pairs are divided to belong to the same position of the dam, pixel points of each image block in the image block pairs are taken, the gray coefficient between the two paired image blocks is calculated according to the gray value of each image block, and the gray coefficient represents the condition of image imaging difference caused by different equipment in shooting weather environments between the current image block and the image block of the pre-stored dam gray image. Because cracks appear in some dam structures, the gray coefficient of the image block is obviously abnormal to that of most other image blocks. Screening out a small number of corrected image blocks with abnormal gray coefficients as image features, and correcting all image blocks of the dam building framework gray images by adopting normal gray coefficients in the step S345 according to the image features, so that the influence of shooting different equipment in weather environments is eliminated as much as possible from image data input into a crack identification model subsequently.
Further, the formula for calculating the gamma in S343 is:
wherein the content of the first and second substances,
is as follows
For the gamma corresponding to the image block,
for pre-storing dam gray level image
The first of a picture block
The gray value of each pixel point is calculated,
constructing gray scale images for dam buildings
Image block number one
The gray value of each pixel point is calculated,
the pixel points in the image block.
The beneficial effects of the above further scheme are: the different influences of different equipment in the shooting weather environment are on the whole dam image, so that the dam image is changed into a gray-scale image, and different influence degrees of different equipment in the shooting weather environment are represented by the difference between the gray-scale value of the gray-scale image and the gray-scale value of a background gray-scale image.
Further, the normal gamma in S345 is: a gamma between an upper gamma threshold and a lower gamma threshold.
The beneficial effects of the above further scheme are: the proportion of the partial image features with cracks in the whole dam image is small, so that after blocking, most of calculated gray coefficients are in the same level, and the small part of calculated gray coefficients are in an abnormal level, so that an upper limit gray coefficient threshold value and a lower limit gray coefficient threshold value can be set to screen out normal gray coefficients.
Further, the S346 includes the following sub-steps:
s3461, selecting a pair of image blocks, where the pair of image blocks satisfies the following condition: the difference value between the gray value sum of the corrected image blocks corresponding to the image blocks and the gray value sum of the image blocks of the pre-stored dam gray image is smaller than a set threshold value;
s3462, screening out the corrected image blocks with the difference values smaller than the threshold value in the step S3461 as central image blocks;
s3463, calculating the difference value between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block;
s3464, judging whether the absolute value of the difference between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block is larger than a difference threshold value, if so, screening the adjacent corrected image block without participating in subsequent traversal, and if not, jumping to S3465;
s3465, selecting any adjacent corrected image block as a next central image block, and jumping to the step S3463 until all the corrected image blocks are traversed;
s3466, using the corrected image block screened in step S3464 as an image feature.
The beneficial effects of the above further scheme are: in step S3461, the selection condition must be satisfied: the difference value between the gray value sum of the corrected image blocks corresponding to the image blocks and the gray value sum of the image blocks of the pre-stored dam gray image is smaller than a set threshold value, so that the selected image blocks do not contain crack features. After correction, if there is no crack feature in the corrected image blocks, the difference between the total gray scale value of the same location and the total gray scale value of the background image block is very small, so that the corrected image block without crack feature is selected as the first central image block in step S3461, and the central image block is compared with the gray scale coefficients of other adjacent corrected image blocks to screen out the corrected image blocks with a large difference from the gray scale coefficient of the central image block, so that there are crack features in the corrected image blocks. If the adjacent corrected image block has the same gray coefficient as the central image block, the adjacent corrected image block is indicated to have no crack feature, and one adjacent corrected image block without crack feature can be selected as the next traversal object until all corrected image blocks are found.
Further, the crack recognition model in the S3 adopts a neural network.
Further, the step S3 includes the following sub-steps:
s31, fusing all image characteristics belonging to the same dam image to obtain fused characteristics;
and S32, inputting the fusion characteristics into a crack identification model to obtain a crack area of the dam.
The beneficial effects of the above further scheme are: since the plurality of images are divided into the plurality of image blocks, the plurality of image blocks are fused in step S31, and all crack feature images belonging to the same dam image are identified by the crack identification model.
Further, the fusion features in S31 are:
wherein the content of the first and second substances,
in order to fuse the features of the image,
a first clipping matrix of 0 and 1,
is composed of 0 and 1
The matrix is clipped in such a way that,
is composed of 0 and 1
The matrix is clipped in such a way that,
the product of the Hadamard is used as the target,
is a first image characteristic，
Is as follows
The characteristics of the image are shown in the figure,
is as follows
The characteristics of the image are shown in the figure,
is the number of image features.
The beneficial effects of the above further scheme are: by means of a first clipping matrix
To the first
Clipping matrix
Respectively for the first image characteristics
To the first
Image features
And cutting to obtain the fracture fusion characteristics.
A system of a dam safety detection method is characterized by comprising the following steps: the method comprises the steps of collecting a dam image unit, an image feature extraction unit and a crack area identification unit;
the dam image acquisition unit is used for acquiring dam images;
the image feature extraction unit is used for extracting image features of the dam image;
the crack area identification unit is used for processing image characteristics by adopting a crack identification model to obtain a crack area of the dam.
The system has the beneficial effects that: according to the dam image detection method, the dam detection is realized in a mode of avoiding using radar equipment to emit electromagnetic waves through image acquisition, image features are extracted from dam images, and then the image features are processed through a crack recognition model, so that the dam original images are prevented from being directly processed, and the recognition accuracy is improved.
Drawings
FIG. 1 is a flow chart of a dam safety inspection method;
fig. 2 is a system block diagram of a dam safety inspection system.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the 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 it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
Example 1:
as shown in fig. 1, a dam safety detection method includes the following steps:
s1, collecting dam images;
s2, extracting image characteristics of the dam image;
and S3, processing the image characteristics by adopting a crack identification model to obtain a crack area of the dam.
The embodiment 1 of the invention has the following beneficial effects: according to the dam image detection method, the dam detection is realized in a mode of avoiding using radar equipment to emit electromagnetic waves through image acquisition, image features are extracted from dam images, and then the image features are processed through a crack recognition model, so that the dam original images are prevented from being directly processed, and the recognition accuracy is improved.
Example 2:
the method for S2 in the embodiment 1 comprises the following substeps:
s21, clipping the dam image to obtain a dam building architecture image;
s22, carrying out gray level processing on the dam building architecture image to obtain a dam building architecture gray level image;
s23, aligning the dam building framework gray level image with a pre-stored dam gray level image according to the structural characteristics of the dam;
and S34, screening out image characteristics according to the aligned dam building framework gray level image and the pre-stored dam gray level image.
The beneficial effects of the embodiment 2 are as follows: the method comprises the steps of cutting a dam image, extracting a dam building framework image, carrying out gray level processing on the dam building framework image to reduce data volume, aligning the dam building framework gray level image with a prestored dam gray level image, and facilitating subsequent same-region blocking after aligning.
Example 3:
aiming at S34 in the embodiment 2, the method comprises the following substeps:
s341, dividing the dam building framework gray level image and the pre-stored dam gray level image into a plurality of image blocks respectively;
s342, forming image block pairs by image blocks of the dam building framework gray level image at the same position and image blocks of a pre-stored dam gray level image;
s343, taking a pixel point of each image block in the image block pair, and calculating a gray coefficient based on the gray value of the pixel point;
s345, correcting all image blocks of the dam building framework gray level image according to the normal gray level coefficient to obtain corrected image blocks;
and S346, screening the corrected image blocks with abnormal gray coefficients as image features.
The beneficial effects of the embodiment 3 are as follows: the dam building framework gray image is aligned with the pre-stored dam gray image, so that the image block pairs are divided to belong to the same position of the dam, pixel points of each image block in the image block pairs are taken, the gray coefficient between the two paired image blocks is calculated according to the gray value of each image block, and the gray coefficient represents the condition of image imaging difference caused by different equipment in shooting weather environments between the current image block and the image block of the pre-stored dam gray image. And because cracks already appear in some dam structures, the gray coefficient of the image block is obviously abnormal to that of most other image blocks. Screening out a small number of corrected image blocks with abnormal gray coefficients as image features, and correcting all image blocks of the dam building framework gray images by adopting normal gray coefficients in the step S345 according to the image features, so that the influence of shooting different equipment in weather environments is eliminated as much as possible from image data input into a crack identification model subsequently.
Example 4:
the formula for calculating the gamma in S343 in embodiment 3 is:
wherein the content of the first and second substances,
is as follows
For the gamma corresponding to the image block,
for pre-storing dam gray level image
The first of a picture block
The gray value of each pixel point is calculated,
constructing gray scale images for dam buildings
The first of a picture block
The gray value of each pixel point is calculated,
the pixel points in the image block.
The beneficial effects of the embodiment 4 are as follows: the different influences of different equipment in the shooting weather environment are on the whole dam image, so that the dam image is changed into a gray-scale image, and different influence degrees of different equipment in the shooting weather environment are represented by the difference between the gray-scale value of the gray-scale image and the gray-scale value of a background gray-scale image.
Example 5:
the normal gamma for S345 in embodiment 3 is: a gamma between the upper gamma threshold and the lower gamma threshold.
The beneficial effects of the embodiment 5 are as follows: the proportion of partial image features with cracks in the whole dam image is small, so that after blocking, most of calculated gray coefficients are in the same level, and the small part of calculated gray coefficients are in an abnormal level, so that an upper limit gray coefficient threshold value and a lower limit gray coefficient threshold value can be set to screen out normal gray coefficients.
Example 6:
the step for S346 in example 3 includes the following substeps:
s3461, selecting a pair of image blocks, where the pair of image blocks satisfies the following condition: the difference value between the gray value sum of the corrected image blocks corresponding to the image blocks and the gray value sum of the image blocks of the pre-stored dam gray image is smaller than a set threshold value;
s3462, screening out the corrected image blocks with the difference values smaller than the threshold value in the step S3461 as central image blocks;
s3463, calculating the difference between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block;
s3464, judging whether the absolute value of the difference between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block is larger than a difference threshold value, if so, screening the adjacent corrected image block without participating in subsequent traversal, and if not, jumping to S3465;
s3465, selecting one adjacent corrected image block as a next central image block, and jumping to the step S3463 until all the corrected image blocks are traversed;
s3466, using the corrected image block screened in step S3464 as an image feature.
The beneficial effects of the embodiment 6 are as follows: in step S3461, the selection condition must be satisfied: the difference value between the gray value sum of the corrected image blocks corresponding to the image blocks and the gray value sum of the image blocks of the pre-stored dam gray image is smaller than a set threshold value, and therefore the selected image blocks do not contain crack features. After correction, if there is no crack feature in the corrected image blocks, the difference between the total gray scale value at the same position and the total gray scale value of the background image block is very small, so that the corrected image block without crack feature is selected as the first central image block through step S3461, and the corrected image blocks with a larger difference from the gray scale coefficient of the central image block are selected through comparing the gray scale coefficients of the central image block and other adjacent corrected image blocks, so that there is a crack feature in the corrected image blocks. If the adjacent corrected image block has the same gray coefficient as the central image block, the adjacent corrected image block is indicated to have no crack feature, and one adjacent corrected image block without crack feature can be selected as the next traversal object until all corrected image blocks are found.
For the crack recognition models in the above embodiments 1 to 6, a neural network is used in S3.
Example 7:
the method for S3 in the embodiment 1 comprises the following substeps:
s31, fusing all image characteristics belonging to the same dam image to obtain fused characteristics;
and S32, inputting the fusion characteristics into a crack identification model to obtain a crack area of the dam.
The beneficial effects of the embodiment 7 are as follows: since the plurality of images are divided into the plurality of image blocks, the plurality of image blocks are fused in step S31, and all crack feature images belonging to the same dam image are recognized by the crack recognition model.
Example 8:
the fusion characteristics in S31 for example 7 are:
wherein, the first and the second end of the pipe are connected with each other,
in order to fuse the features of the image,
a first clipping matrix of 0 and 1,
is composed of 0 and 1
The matrix is clipped by a clipping matrix that,
is 0 and 1 composed of
The matrix is clipped by a clipping matrix that,
the product of the Hadamard is used as the target,
in order to be a first image feature,
is as follows
The characteristics of the image are shown in the figure,
is as follows
The characteristics of the image are such that,
is the number of image features.
The beneficial effects of the embodiment 8 are as follows: by means of a first clipping matrix
To the first
Clipping matrix
Respectively to the first image characteristics
To the first
Image features
And cutting to obtain the fracture fusion characteristics.
Example 9:
for the above embodiments 1 to 8, the loss function of the crack identification model is as follows:
wherein, the first and the second end of the pipe are connected with each other,
in order to be a function of the loss,
for the output of the crack identification model,
for a desired output corresponding to the output of the crack identification model,
the two-dimensional image data are combined to form an intersection,
in order to be a union set,
is a first weight coefficient of the first weight coefficient,
is a second weight coefficient, and is,
the abscissa of the geometric center point of the output for the crack identification model,
as the abscissa of the geometric center point of the desired output,
for the ordinate of the geometric center point of the output of the crack identification model,
as the ordinate of the geometric center point of the desired output,
the distance between the farthest point on the output of the model and the farthest point corresponding to the desired output is identified for the crack.
The beneficial effects of the embodiment 9 are as follows: output by crack recognition model
And expected output
Of intersection of
And output of
And expected output
Union of (1)
To measure the output
And expected output
Meanwhile, the distance between the images is measured by the geometric central point of the output of the crack identification model and the expected geometric central point, and the overall loss is measured by the loss of the two aspects.
Example 10:
for embodiments 1 to 9, as shown in fig. 2, a system of a dam safety detection method is provided, including: the method comprises the steps of collecting a dam image unit, an image feature extraction unit and a crack area identification unit;
the dam image acquisition unit is used for acquiring a dam image;
the image feature extraction unit is used for extracting image features of the dam image;
the crack area identification unit is used for processing image characteristics by adopting a crack identification model to obtain a crack area of the dam.
The beneficial effects of the embodiment 10 are as follows: according to the dam image detection method, the dam detection is realized in a mode of avoiding using radar equipment to emit electromagnetic waves through image acquisition, image features are extracted from dam images, and then the image features are processed through a crack recognition model, so that the dam original images are prevented from being directly processed, and the recognition accuracy is improved.
Claims (10)
1. A dam safety detection method is characterized by comprising the following steps:
s1, collecting dam images;
s2, extracting image characteristics of the dam image;
and S3, processing the image characteristics by adopting a crack identification model to obtain a crack area of the dam.
2. The dam safety detection method according to claim 1, wherein the step S2 comprises the following substeps:
s21, clipping the dam image to obtain a dam building architecture image;
s22, carrying out gray level processing on the dam building framework image to obtain a dam building framework gray level image;
s23, aligning the gray level image of the dam building framework with a pre-stored gray level image of the dam according to the structural characteristics of the dam;
and S34, screening out image characteristics according to the aligned dam building framework gray level image and the pre-stored dam gray level image.
3. The dam safety detection method according to claim 2, wherein the step S34 comprises the following substeps:
s341, dividing the dam building framework gray level image and the pre-stored dam gray level image into a plurality of image blocks respectively;
s342, forming an image block pair by the image blocks of the dam building framework gray level image and the image blocks of the pre-stored dam gray level image at the same position;
s343, taking a pixel point of each image block in the image block pair, and calculating a gray coefficient based on the gray value of the pixel point;
s345, correcting all image blocks of the dam building framework gray level image according to the normal gray level coefficient to obtain corrected image blocks;
and S346, screening the corrected image blocks with abnormal gray coefficients as image features.
4. The dam safety detection method according to claim 3, wherein the formula for calculating the gamma in S343 is:
wherein the content of the first and second substances,
is a first
For the gamma corresponding to the image block,
for pre-storing dam gray level image
The first of a picture block
The gray value of each pixel point is calculated,
constructing gray scale images for dam buildings
The first of a picture block
The gray value of each pixel point is calculated,
are pixels in the image block.
5. The dam safety detection method according to claim 3, wherein the normal gamma in S345 is: a gamma between the upper gamma threshold and the lower gamma threshold.
6. The dam safety detection method according to claim 3, wherein the S346 comprises the following substeps:
s3461, selecting a pair of image blocks, where the image blocks satisfy the following conditions: the difference value between the gray value sum of the corrected image blocks corresponding to the image blocks and the gray value sum of the image blocks of the pre-stored dam gray image is smaller than a set threshold value;
s3462, screening the corrected image blocks with the difference values smaller than the threshold value in the step S3461 to be used as central image blocks;
s3463, calculating the difference value between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block;
s3464, judging whether the absolute value of the difference between the gray coefficient of the central image block and the gray coefficient of the adjacent corrected image block is larger than a difference threshold value, if so, screening the adjacent corrected image block without participating in subsequent traversal, and if not, jumping to S3465;
s3465, selecting one adjacent corrected image block as a next central image block, and jumping to the step S3463 until all the corrected image blocks are traversed;
s3466, using the corrected image block screened in step S3464 as an image feature.
7. The dam safety detection method according to claim 1, wherein the crack recognition model in S3 is a neural network.
8. The dam safety detection method according to claim 1, wherein the step S3 comprises the following substeps:
s31, fusing all image characteristics belonging to the same dam image to obtain fused characteristics;
and S32, inputting the fusion characteristics into a crack identification model to obtain a crack area of the dam.
9. The dam safety detection method according to claim 8, wherein the fusion features in S31 are:
wherein the content of the first and second substances,
in order to fuse the features of the image,
a first clipping matrix of 0 and 1,
is 0 and 1 composed of
The matrix is clipped in such a way that,
is composed of 0 and 1
The matrix is clipped in such a way that,
the product of the Hadamard is used as the target,
is a feature of the first image that is characteristic of the first image,
is a first
The characteristics of the image are such that,
is as follows
The characteristics of the image are shown in the figure,
is the number of image features.
10. A system of the dam safety detection method according to any one of claims 1 to 9, comprising: the method comprises the steps of collecting a dam image unit, an image feature extraction unit and a crack area identification unit;
the dam image acquisition unit is used for acquiring a dam image;
the image feature extraction unit is used for extracting image features of the dam image;
the crack area identification unit is used for processing image characteristics by adopting a crack identification model to obtain a crack area of the dam.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211283181.4A CN115346127B (en) | 2022-10-20 | 2022-10-20 | Dam safety detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211283181.4A CN115346127B (en) | 2022-10-20 | 2022-10-20 | Dam safety detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115346127A true CN115346127A (en) | 2022-11-15 |
| CN115346127B CN115346127B (en) | 2023-01-24 |
Family
ID=83957391
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211283181.4A Active CN115346127B (en) | 2022-10-20 | 2022-10-20 | Dam safety detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115346127B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105954292A (en) * | 2016-04-29 | 2016-09-21 | 河海大学常州校区 | Underwater structure surface crack detection device and method based on compound-eye bionic vision |
| CN107169953A (en) * | 2017-04-07 | 2017-09-15 | 西安电子科技大学 | Bridge concrete surface crack detection method based on HOG features |
| CN108182681A (en) * | 2018-01-12 | 2018-06-19 | 石家庄铁道大学 | A kind of distress in concrete detection method for the learning network that transfinited based on multilayer |
| CN109087305A (en) * | 2018-06-26 | 2018-12-25 | 汕头大学 | A kind of crack image partition method based on depth convolutional neural networks |
| CN109241985A (en) * | 2017-07-11 | 2019-01-18 | 普天信息技术有限公司 | A kind of image-recognizing method and device |
| CN112465817A (en) * | 2020-12-17 | 2021-03-09 | 大连海事大学 | Pavement crack detection method based on directional filter |
| CN113237886A (en) * | 2021-04-26 | 2021-08-10 | 安徽猎寻科技有限公司 | New energy automobile wheel tread defect detection method based on photoelectric measurement |
| CN113744268A (en) * | 2021-11-04 | 2021-12-03 | 深圳市城市交通规划设计研究中心股份有限公司 | Crack detection method, electronic device and readable storage medium |
| CN113781418A (en) * | 2021-08-30 | 2021-12-10 | 大连地铁集团有限公司 | Subway image anomaly detection method and system based on comparison and storage medium |
| CN114359542A (en) * | 2021-11-26 | 2022-04-15 | 广州大学 | Concrete structure crack identification method based on computer vision and deep learning |
| CN114387233A (en) * | 2021-12-29 | 2022-04-22 | 山东大学 | Sand mold defect detection method and system based on machine vision |
| CN115063423A (en) * | 2022-08-18 | 2022-09-16 | 启东市嘉信精密机械有限公司 | Mechanical casting cold and hot crack self-adaptive identification method based on computer vision |
| CN115082467A (en) * | 2022-08-22 | 2022-09-20 | 山东亿昌装配式建筑科技有限公司 | Building material welding surface defect detection method based on computer vision |
-
2022
- 2022-10-20 CN CN202211283181.4A patent/CN115346127B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105954292A (en) * | 2016-04-29 | 2016-09-21 | 河海大学常州校区 | Underwater structure surface crack detection device and method based on compound-eye bionic vision |
| CN107169953A (en) * | 2017-04-07 | 2017-09-15 | 西安电子科技大学 | Bridge concrete surface crack detection method based on HOG features |
| CN109241985A (en) * | 2017-07-11 | 2019-01-18 | 普天信息技术有限公司 | A kind of image-recognizing method and device |
| CN108182681A (en) * | 2018-01-12 | 2018-06-19 | 石家庄铁道大学 | A kind of distress in concrete detection method for the learning network that transfinited based on multilayer |
| CN109087305A (en) * | 2018-06-26 | 2018-12-25 | 汕头大学 | A kind of crack image partition method based on depth convolutional neural networks |
| CN112465817A (en) * | 2020-12-17 | 2021-03-09 | 大连海事大学 | Pavement crack detection method based on directional filter |
| CN113237886A (en) * | 2021-04-26 | 2021-08-10 | 安徽猎寻科技有限公司 | New energy automobile wheel tread defect detection method based on photoelectric measurement |
| CN113781418A (en) * | 2021-08-30 | 2021-12-10 | 大连地铁集团有限公司 | Subway image anomaly detection method and system based on comparison and storage medium |
| CN113744268A (en) * | 2021-11-04 | 2021-12-03 | 深圳市城市交通规划设计研究中心股份有限公司 | Crack detection method, electronic device and readable storage medium |
| CN114359542A (en) * | 2021-11-26 | 2022-04-15 | 广州大学 | Concrete structure crack identification method based on computer vision and deep learning |
| CN114387233A (en) * | 2021-12-29 | 2022-04-22 | 山东大学 | Sand mold defect detection method and system based on machine vision |
| CN115063423A (en) * | 2022-08-18 | 2022-09-16 | 启东市嘉信精密机械有限公司 | Mechanical casting cold and hot crack self-adaptive identification method based on computer vision |
| CN115082467A (en) * | 2022-08-22 | 2022-09-20 | 山东亿昌装配式建筑科技有限公司 | Building material welding surface defect detection method based on computer vision |
Non-Patent Citations (3)
| Title |
|---|
| CAO VU DUNG等: "Autonomous concrete crack detection using deep fully convolutional neural network", 《AUTOMATION IN CONSTRUCTION》 * |
| 王泽矫等: "基于深度学习的大坝裂缝检测方法研究", 《水利规划与设计》 * |
| 范新南等: "基于劳伦茨信息值的水下大坝裂缝提取算法", 《计算机与现代化》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115346127B (en) | 2023-01-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109902633B (en) | Abnormal event detection method and device based on fixed-position camera monitoring video | |
| CN108537215A (en) | A kind of flame detecting method based on image object detection | |
| JP4675949B2 (en) | Method and apparatus for measuring crack width of structures and products using image processing technique | |
| KR100773393B1 (en) | Real-time Monitoring System and Method for DAM | |
| CN105279772B (en) | A kind of trackability method of discrimination of infrared sequence image | |
| US11657491B2 (en) | Learning data collection apparatus, learning data collection method, and program | |
| CN109559324B (en) | Target contour detection method in linear array image | |
| CN110008947B (en) | Granary grain quantity monitoring method and device based on convolutional neural network | |
| CN110766058B (en) | Battlefield target detection method based on optimized RPN (resilient packet network) | |
| CN109360190B (en) | Building damage detection method and device based on image superpixel fusion | |
| CN115639248A (en) | System and method for detecting quality of building outer wall | |
| CN103927553A (en) | Coal and rock recognition method based on multi-scale micro-lamination and contrast ratio joint distribution | |
| CN114462646B (en) | Pole number plate identification method and system based on contact network safety inspection | |
| CN115512247A (en) | Regional building damage grade assessment method based on image multi-parameter extraction | |
| CN114219773A (en) | Pre-screening and calibration method for bridge crack detection data set | |
| CN112580542A (en) | Steel bar counting method based on target detection | |
| CN115862259A (en) | Fire alarm early warning system based on temperature monitoring | |
| CN117635604B (en) | Method for detecting welding quality of constructional engineering steel structure | |
| CN115346127B (en) | Dam safety detection method and system | |
| CN110084795A (en) | A kind of infrared image blind pixel detection method and system based on background | |
| CN107369163B (en) | Rapid SAR image target detection method based on optimal entropy dual-threshold segmentation | |
| CN110837775A (en) | Underground locomotive pedestrian and distance detection method based on binarization network | |
| CN110794397A (en) | Target detection method and system based on camera and radar | |
| CN112784703B (en) | Multispectral-based personnel action track determination method | |
| CN112461345B (en) | Truck scale rolling line out-of-bounds detection method based on LSD (least squares distortion) linear detection 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 |