CN117173190A - Insulator infrared damage inspection system based on image processing - Google Patents
Insulator infrared damage inspection system based on image processing Download PDFInfo
- Publication number
- CN117173190A CN117173190A CN202311454353.4A CN202311454353A CN117173190A CN 117173190 A CN117173190 A CN 117173190A CN 202311454353 A CN202311454353 A CN 202311454353A CN 117173190 A CN117173190 A CN 117173190A
- Authority
- CN
- China
- Prior art keywords
- image
- insulator
- contour
- infrared
- frequency domain
- 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
- 239000012212 insulator Substances 0.000 title claims abstract description 123
- 238000012545 processing Methods 0.000 title claims abstract description 32
- 230000006378 damage Effects 0.000 title claims abstract description 28
- 238000007689 inspection Methods 0.000 title claims abstract description 13
- 238000001914 filtration Methods 0.000 claims abstract description 36
- 230000002457 bidirectional effect Effects 0.000 claims abstract description 20
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 230000004927 fusion Effects 0.000 claims abstract description 11
- 230000009466 transformation Effects 0.000 claims description 50
- 238000003708 edge detection Methods 0.000 claims description 42
- 239000011159 matrix material Substances 0.000 claims description 39
- 238000000034 method Methods 0.000 claims description 35
- 238000009499 grossing Methods 0.000 claims description 29
- 238000004422 calculation algorithm Methods 0.000 claims description 20
- 230000004044 response Effects 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000011426 transformation method Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 4
- 230000001629 suppression Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000011179 visual inspection Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Abstract
The invention discloses an insulator infrared damage inspection system based on image processing, which relates to the technical field of image processing, and comprises the following components: an image acquisition section including a plurality of infrared cameras for acquiring infrared images of insulators, the insulators being divided into a plurality of areas, each insulator acquiring an area infrared image of one area of the insulator; the image fusion part is configured to fuse the regional infrared images to obtain an insulator infrared image; the image bidirectional filtering part is configured to perform bidirectional filtering on the insulator infrared image to obtain a filtered insulator image; and the contour extraction part is configured to extract the contour of the filter insulator image, and the damage detection part is configured to judge that the abnormal region is damaged. The invention improves the precision and reliability of insulator detection, reduces false alarm rate, and realizes real-time monitoring and maintenance decision support, thereby improving the reliability and safety of the power system.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an insulator infrared damage inspection system based on image processing.
Background
With the continued development of power engineering and power equipment, the reliability and safety of power systems has become a key concern. Insulators are an important component in electrical power systems, and their reliability is critical to the operation of the electrical power system. However, in power equipment, the long-term use of the insulator may be affected by various environmental factors, such as severe weather conditions, power load changes, and pollution, which may cause the performance of the insulator to gradually decrease, even to be damaged, thereby affecting the stability and safety of the power system.
To ensure reliable operation of the power system, the insulators need to be inspected and maintained regularly. Conventional insulator inspection methods rely primarily on visual inspection and manual inspection, which suffer from a number of disadvantages. First, visual inspection requires human resources, is time consuming and laborious, and may result in errors or leaks due to human factors. Second, visual inspection can only detect significant damage, often imperceptible to minor defects or concealment problems inside the insulator. Therefore, the traditional insulator detection method has the problems of low efficiency, low accuracy and incapability of real-time monitoring.
In order to overcome the limitations of the conventional methods, some power equipment monitoring techniques and equipment have begun to be widely used in recent years. Among them, infrared (IR) imaging technology has been introduced for detection of insulators. This technique uses infrared radiation from the insulator surface to detect the temperature distribution of the insulator so that potential faults and problems can be identified. However, conventional infrared imaging techniques still have some limitations. It relies primarily on a single infrared camera to capture temperature information of the insulator surface, ignoring the non-uniformity and complexity of the insulator.
Disclosure of Invention
The invention aims to provide an insulator infrared damage detection system based on image processing, which improves the accuracy and reliability of insulator detection, reduces false alarm rate, and realizes real-time monitoring and maintenance decision support, thereby improving the reliability and safety of a power system.
In order to solve the technical problems, the invention provides an insulator infrared damage inspection system based on image processing, which comprises: an image acquisition section including a plurality of infrared cameras for acquiring infrared images of insulators, the insulators being divided into a plurality of areas, each insulator acquiring an area infrared image of one area of the insulator; the image fusion part is configured to fuse the regional infrared images to obtain an insulator infrared image; the image bidirectional filtering part is configured to perform bidirectional filtering on the insulator infrared image to obtain a filtered insulator image; the contour extraction part is configured to extract the contour of the filter insulator image, and specifically comprises: representing the filter insulator image as a two-dimensional matrix, performing matrix smoothing on the two-dimensional matrix, performing edge detection to obtain an edge detection result, extracting a contour, performing contour wavelet transformation on the contour, performing frequency domain filtering on the contour after the contour wavelet transformation to obtain a frequency domain image, performing inverse wavelet transformation on the frequency domain image, and reconstructing the contour; and the damage checking part is configured for calculating gradient difference values among the reconstructed contours, comparing the gradient difference values with respective corresponding preset judging values to judge whether an abnormality occurs, taking the areas of the insulators corresponding to the two contours with the abnormal difference values as abnormal areas, and judging that the abnormal areas are damaged.
Further, the method for fusing the regional infrared images by the image fusion part to obtain the insulator infrared image comprises the following steps: first, a target image is setFinding a regional infrared image using an image feature point matching algorithm>And target image->Corresponding feature point pair ∈>And->The method comprises the steps of carrying out a first treatment on the surface of the Setting a transformation matrix->Using a transformation matrix->Regional infrared image +.>Transformed into the target image coordinate space.
Further, the transformation matrixCalculated using the following formula:
；
wherein,is the number of feature point pairs +.>Is the sequence number of the feature point pair.
Further, the image bidirectional filtering part performs bidirectional filtering on the insulator infrared image, and the method for obtaining the filtered insulator image comprises the following steps: inputting the insulator infrared image into a preset bidirectional filter to obtain a filter insulator image; the bi-directional filter is expressed using the following formula:
wherein,is the pixel coordinates; />Is the pixel value in the infrared image of the insulator; />Is the pixel value of the filtered insulator image; />Is a fixed filter window comprising +.>Is a coordinate pair of (2); />Is a normalization coefficient for ensuring that the output pixel value is between 0 and 255; />Is a spatial domain weight;is a value range weight.
Further toThe spatial domain weightCalculated using the following formula:
；
wherein,is the spatial domain variance.
Further, the value range weight is calculated by using the following formula:
；
wherein,representing the difference between pixel values; />Is the custom value range standard deviation.
Further, the contour extraction section includes: an image conversion unit configured to represent the filtered insulator image as a two-dimensional matrix; the smoothing unit is configured to perform matrix smoothing on the two-dimensional matrix to obtain a smoothed matrix; an edge detection unit configured to perform edge detection on the smooth matrix, the edge detection result; a contour wavelet transformation unit configured to extract a contour based on an edge detection result, and perform contour wavelet transformation on the contour; and the frequency domain filtering unit is configured to perform frequency domain filtering on the contour after the contour wavelet transformation by using a frequency domain filter to obtain a frequency domain image, wherein the frequency domain filter is expressed by using the following formula:
；
wherein the method comprises the steps ofIs the response of the filter, +.>Is a distance function in the frequency domain, +.>Is a cut-off frequency, +.>Is the order of the filter; />And->Is a frequency domain coordinate representing different frequency components; and a contour reconstruction unit configured to reconstruct a contour by performing inverse wavelet transform on the frequency domain image line.
Further, the contour wavelet transformation unit extracts a contour based on an edge detection result, and the contour wavelet transformation method comprises the following steps:
the contour is extracted based on the edge detection result using the following formula:
；
wherein the method comprises the steps ofRepresenting outline->And->Respectively represent the edge detection result->At->And->Gradient in direction;
the contours are subjected to contour wavelet transformation using the following formula:
；
wherein,is the contour of the contour wavelet transform, +.>Representing wavelet scale, +.>Is a wavelet function, +.>The upper limit of the coordinate value is the pixel abscissa or ordinate.
Further, after obtaining the contour of the contour wavelet transform, before filtering in the frequency domain, the contour of the contour wavelet transform is further thresholded, which specifically includes:
；
wherein,is the result of the thresholding; />Is a threshold parameter.
Further, the edge detection unit performs edge detection on the smooth matrix by using a Canny edge detection algorithm.
The insulator infrared damage inspection system based on image processing has the following beneficial effects: according to the invention, the infrared image information of the insulator is acquired through the plurality of infrared cameras, so that each infrared camera only acquires the infrared image information of a part of insulator areas, and finally, gradient difference values among different infrared image information areas are found after the infrared image information is fused, the gradient difference values reflect the temperature difference of the different insulator areas, if the difference values deviate from preset values, the occurrence of abnormality can be judged, the infrared temperature analysis of the whole image is not needed, the analysis efficiency is improved, and meanwhile, the part of the insulator with a problem can be more accurately determined, so that the efficiency is further improved. In addition, conventional insulator inspection methods generally capture only significant defects on the surface, but cannot penetrate into subtle problems such as small cracks or blind faults. The contour wavelet transformation technology has excellent detail extraction capability, and can capture tiny contours and features in an image, so that the problem below the surface can be detected, and the detection comprehensiveness is improved. The contour wavelet transform not only extracts details, but also increases the sensitivity of detection. It can recognize weak signals and changes in the image, and even small temperature differences or contour features can be accurately captured. This allows the system to discover potential problems earlier, helping to prevent serious damage to equipment and improving the reliability of the power system. The bidirectional filtering technology can effectively inhibit noise in the image by integrating the spatial domain and the value domain information. This is particularly important for insulator infrared images, as insulators may be subject to contamination or natural factors in harsh environments, resulting in noise in the image. Through noise suppression, the system can more accurately capture the temperature distribution of the insulator, reduce the risk of misjudgment, and further improve the detection reliability.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic system structure diagram of an insulator infrared damage inspection system based on image processing according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1: referring to fig. 1, an image processing-based insulator infrared damage verification system, the system comprising: an image acquisition section including a plurality of infrared cameras for acquiring infrared images of insulators, the insulators being divided into a plurality of areas, each insulator acquiring an area infrared image of one area of the insulator; the image fusion part is configured to fuse the regional infrared images to obtain an insulator infrared image; the image bidirectional filtering part is configured to perform bidirectional filtering on the insulator infrared image to obtain a filtered insulator image; the contour extraction part is configured to extract the contour of the filter insulator image, and specifically comprises: representing the filter insulator image as a two-dimensional matrix, performing matrix smoothing on the two-dimensional matrix, performing edge detection to obtain an edge detection result, extracting a contour, performing contour wavelet transformation on the contour, performing frequency domain filtering on the contour after the contour wavelet transformation to obtain a frequency domain image, performing inverse wavelet transformation on the frequency domain image, and reconstructing the contour; and the damage checking part is configured for calculating gradient difference values among the reconstructed contours, comparing the gradient difference values with respective corresponding preset judging values to judge whether an abnormality occurs, taking the areas of the insulators corresponding to the two contours with the abnormal difference values as abnormal areas, and judging that the abnormal areas are damaged.
In particular, an insulator is a component commonly found in electrical or electrical equipment and is primarily used to isolate current in the circuit and prevent current from escaping or leaking to places where it should not flow. Insulators are typically used to separate an electrical conductor (e.g., a wire or lead) from a support structure (e.g., a utility pole or cable rack) to ensure that current flows along a predetermined path without causing a short circuit or danger. Damage or failure of the insulator may lead to a power outage or failure of the power system, which is a significant problem for the reliability of the power supply. The health state of the insulator is monitored by using the insulator infrared damage inspection system regularly, so that the risks of power failure and power interruption can be reduced, and the stability of the power system is improved.
The main function of the image acquisition part is to acquire an infrared image of the insulator. These images show the temperature distribution of the insulator surface. By means of the infrared image, it is possible to identify whether there is a temperature anomaly or an abnormal hot spot of the insulator, which may be a sign of damage or malfunction of the insulator.
Wavelet transformation is a signal processing technique for decomposing a signal or image into components of different frequencies. It is similar to a fourier transform, but the wavelet transform has some locality and can better capture short-time variations in the signal or image. In the contour wavelet transform, contour information is first extracted from an input image. Contours are typically represented as binary images in which edges or object boundaries are displayed in a distinct manner. Then, wavelet transformation is applied to the contour image. The wavelet transform decomposes the image into sub-images of different scales, which contain different frequency information of the contours. These sub-images typically comprise a low frequency part and a high frequency part, wherein the high frequency part contains information of details and contours. In the frequency domain, the information of the contour can be better analyzed and processed. Wavelet transforms can highlight edges and features of contours making them easier to identify and quantify. By wavelet transformation, useful features in the contour may be enhanced, such as sharpness of edges and saliency of shape features. The main role of the contour wavelet transform is to enhance the contour information in the image. It can make the outline clearer, more easily identified, and facilitate the subsequent image analysis task. By wavelet transformation, useful features in the contours can be better extracted. These features may be used for applications such as image classification, object detection and recognition. The wavelet transform may also be used to reduce noise, remove interference and noise in the image, to obtain cleaner profile information. Wavelet transforms allow images to be analyzed on different scales, which is particularly useful for processing images of objects or structures having different sizes. The multi-scale representation of the profile information may provide a more comprehensive analysis.
Example 2: on the basis of the above embodiment, the method for fusing the regional infrared image to obtain the insulator infrared image by the image fusion part includes: first, a target image is setFinding a regional infrared image using an image feature point matching algorithm>And target image->Corresponding feature point pair ∈>And->The method comprises the steps of carrying out a first treatment on the surface of the Setting a transformation matrix->Using a transformation matrixRegional infrared image +.>Transformed into the target image coordinate space.
Specifically, first, a target image is determinedThis is typically a reference image, which serves as the target for image fusion. This target image should contain the same insulation as the infrared image of the region to be fusedSub-but possibly photographed under different conditions. Using image feature point matching algorithms, such as SIFT (scale invariant feature transform) or SURF (speed stable feature transform), to find regional infrared images +.>And target image->Corresponding feature point pair ∈>And->. These pairs of feature points will help determine the association between the two images. According to the characteristic point pairs found->And->The transformation matrix is calculated using an image registration technique, such as the RANSAC (random sample consensus) algorithm>. This transformation matrix describes how the regional infrared image is +.>Transformed into the target image coordinate space so that the two can be aligned. Using the calculated transformation matrix +.>Regional infrared image +.>Transform it to map it to the target image +.>Is defined in the coordinate space of (a). Thus, the regional infrared image +.>Will be->Alignment. The multiple region infrared images may then be fused together using image fusion techniques (e.g., weighted averaging or superposition) to generate a composite insulator infrared image.
The main function of the fusion part is to fuse a plurality of regional infrared images into a complete insulator infrared image. This facilitates the operator to view the entire insulator in one image without having to view each area separately. By fusing together the infrared images of the different areas, the system can provide more information, including the temperature distribution of the different parts and the overall state of the insulator. This helps to more fully assess the health of the insulator. The fused single image is easier to analyze, and an operator can more easily identify abnormalities or problems, thereby reducing the operational burden. By aligning and fusing the images, consistency among different areas can be ensured, and thus, the accuracy and reliability of insulator state detection are improved.
Example 3: on the basis of the above embodiment, the transformation matrixCalculated using the following formula:
；
wherein,is the number of feature point pairs +.>Is the sequence number of the feature point pair.
Specifically, firstly, a regional infrared image is found through a characteristic point matching algorithmAnd target image->Corresponding feature point pair +.>And->. These feature points represent similar or corresponding locations in the two images. The goal is to find a transformation matrix +.>So that the characteristic points in the regional infrared image are +.>Feature points mapped into the target image +.>When the sum of the distance errors between all the pairs of feature points is minimal. To achieve this objective, an error function is introduced:
；
the objective of this error function is to minimize the sum of squares of the distance errors between all pairs of feature points. Wherein,representing a transformation matrix->Characteristic points in the infrared image of the region are +.>Feature points mapped into the target image +.>The square of the distance behind. The final goal is to find the most suitable transformation matrix/>To minimize the error function described above and thereby minimize the distance error between pairs of feature points.
The main function of this formula is to calculate the transformation matrixRealize the infrared image of the area +.>With the target imageAlignment tasks. By minimizing the distance error between the feature point pairs, it is ensured that the region infrared image and the target image remain as consistent as possible when aligned. By calculating the optimal transformation matrix->It is possible to ensure that the distance error between the pairs of feature points is minimized, thereby maintaining consistency between images. This helps ensure that the fused image appears natural and seamless without significant discontinuities. By aligning the region infrared image and the target image, the fusion portion can more accurately superimpose them together, thereby improving the quality and usability of the fused insulator infrared image.
Example 4: on the basis of the above embodiment, the image bidirectional filtering section performs bidirectional filtering on the insulator infrared image, and the method for obtaining the filtered insulator image includes: inputting the insulator infrared image into a preset bidirectional filter to obtain a filter insulator image; the bi-directional filter is expressed using the following formula:
wherein,is the pixel coordinates; />Is the pixel value in the infrared image of the insulator; />Is the pixel value of the filtered insulator image; />Is a fixed filter window comprising +.>Is a coordinate pair of (2); />Is a normalization coefficient for ensuring that the output pixel value is between 0 and 255; />Is a spatial domain weight;is a value range weight.
Specifically, in image processing, a bi-directional filter is a filter for smoothing and denoising an image. It combines spatial domain weights and value domain weights to ensure that the smoothing process considers both spatial distances between pixels and similarities between pixel values. The input image is an insulator infrared image in whichRepresenting pixel coordinates, +.>Representing pixel values in the image. The bi-directional filter uses a fixed filter window +.>Comprising ∈>Space of (2)Coordinate offset->. This window defines the set of pixels to be considered in the filtering process. In order to ensure that the output pixel value is between 0 and 255, a normalization coefficient is introduced>. This coefficient is used to scale the output pixel values to stay within a reasonable range. Spatial domain weight +.>Representing a pixel +.>Spatially and central pixel->Is typically represented by a gaussian function or other filter kernel to determine the degree of smoothness between pixels. Value range weightsRepresents the center pixel +.>And adjacent pixels->Similarity between pixel values. The difference between pixel values is typically used to calculate to determine the degree of smoothness between pixels.
The main function of the bidirectional filter is to smooth and denoise the insulator infrared image. It takes into account the spatial relationship between pixels and the similarity between pixel values to ensure that the smoothing process preserves details in the image and reduces noise. Through bidirectional filtering, the definition and detail of the image can be kept as much as possible while denoising is performed, and the obtained filter insulator image is ensured to have high quality. The filtered insulator image is more suitable for subsequent analysis tasks such as contour extraction and anomaly detection. Denoising and smoothing help reduce interference and errors in subsequent processing.
Example 5: on the basis of the above embodiment, the spatial domain weightsCalculated using the following formula:
；
wherein,is the spatial domain variance.
Specifically, in a bi-directional filter, spatial domain weights are used to determine pixelsSpatially and centrally pixelDistance relation of (c) is determined. The calculation of this weight is based on a gaussian function, the purpose of which is to apply a smaller weight to the pixels farther from the center pixel and a larger weight to the pixels nearer to the center pixel during the filtering process.
Spatial domain weightsIs helpful in determining the importance of each pixel in the filter window relative to the center pixel. Pixels further from the center pixel will be less weighted, thus achieving smoothness over the spatial domain, which helps remove small scale noise in the image. By adjusting the spatial domain variance->The width of the spatial domain weight distribution can be controlled. Greater->The value can lead to wider smoothing, and is suitable for removing larger-scale noiseAcoustic and smooth images, while smaller +.>The values will result in sharper filtering, preserving more detail. Properly adjusting the spatial domain standard deviationThe quality of the image can be improved, so that the noise is reduced and important characteristics and details are maintained when the image is denoised.
This formula describes a two-dimensional gaussian distribution in whichThe coordinate offset of (2) determines the size of the weights. Distance from the center pixel +.>More recent pixels +.>Higher weights are obtained and pixels farther away are given lower weights. The shape of the Gaussian distribution is->Control, less->Will result in a steeper weight drop, greater +.>Resulting in a more gradual weight drop. Spatial domain weight +.>The main function of (a) is to control the degree of smoothness applied by the bi-directional filter at different locations in the image. The weight determines the contribution of the neighborhood pixels considered in the filtering process, the pixels further from the center pixel being less affected by smoothing. By using a gaussian distribution, the weights help to preserve important details of the image while smoothing the image. Pixels closer to the center pixel are in smoothingPlay a greater role, while pixels farther away contribute less in smoothing, thereby maintaining the sharpness of the image. By adjusting the spatial domain variance->The intensity of the smoothing process can be controlled. Less->Will result in a stronger smoothing, but a larger +.>Resulting in weaker smoothing. This allows the algorithm to be adapted to the requirements of different images and applications.
Example 6: on the basis of the above embodiment, the value range weight is calculated using the following formula:
；
wherein,representing the difference between pixel values; />Is the custom value range standard deviation.
Specifically, the value range weightsThe main function of (a) is to control the degree of smoothing of the bi-directional filter between different pixel values in the image. If the difference between the pixel values is small, the weight is large, so that the smoothing effect is enhanced; if the difference between pixel values is large, the weights will be small, leaving more detail. The value range weights help to preserve details in the image while smoothing the image and reduce the effects of noise by taking into account the similarity between pixel values. A smaller difference will result in a higher weight, thereby enhancing the smoothing effect; while larger differences result in lower weights to preserve more detail. Custom parameters +.>: by adjusting the standard deviation of the value range->The sensitivity of the smoothing process can be controlled. Less->Will result in a stronger smoothing, but a larger +.>Resulting in weaker smoothing. This allows the algorithm to be adapted to the requirements of different images and applications. The principle and effect of the formula is to calculate the range weights by means of gaussian distribution functions to smooth the image in bi-directional filtering and to control the degree of smoothing. The method is helpful for removing noise in the image, retaining details of the image and improving the image quality. This is very useful for image processing and denoising applications.
Example 7: on the basis of the above embodiment, the contour extraction section includes: an image conversion unit configured to represent the filtered insulator image as a two-dimensional matrix; the smoothing unit is configured to perform matrix smoothing on the two-dimensional matrix to obtain a smoothed matrix; an edge detection unit configured to perform edge detection on the smooth matrix, the edge detection result; a contour wavelet transformation unit configured to extract a contour based on an edge detection result, and perform contour wavelet transformation on the contour; and the frequency domain filtering unit is configured to perform frequency domain filtering on the contour after the contour wavelet transformation by using a frequency domain filter to obtain a frequency domain image, wherein the frequency domain filter is expressed by using the following formula:
；
wherein the method comprises the steps ofIs the response of the filter, +.>Is a distance function in the frequency domain, +.>Is a cut-off frequency, +.>Is the order of the filter; />And->Is a frequency domain coordinate representing different frequency components; and a contour reconstruction unit configured to reconstruct a contour by performing inverse wavelet transform on the frequency domain image line.
Specifically, in frequency domain image processing, one image may be decomposed into components of different frequency components. Each point in the frequency domainA component of a particular frequency is represented whose magnitude corresponds to the intensity of the frequency component. Distance functionRepresenting the frequency domain coordinates +.>The distance to the origin of the frequency domain is typically represented using euclidean distance or other suitable distance metric. Cut-off frequency->Is a parameter that controls the frequency response of the filter. It determines which frequency components will be emphasized or suppressed in the frequency domain. Less->Will result in a strong filtering effect, greater +.>Will result in weakerFiltering effect. Order->The sharpness of the filter is controlled. Greater->The value will lead to a steeper frequency response, whereas a smaller +.>The values will result in a smoother frequency response.
The function of the formula of the frequency domain filter is based on the frequency domain coordinatesDistance function of>To adjust the weights of the different frequency components in the frequency domain. The filter passes the distance function in the formula +.>Different frequency components in the frequency domain are weighted. The distance function determines which frequency components are to be enhanced and which are to be suppressed. This helps the filter to emphasize or suppress signals in a particular frequency range. Cut-off frequency->The value of (2) determines the frequency response range of the frequency domain filter. Smaller and smallerWill select lower frequency components, and larger +.>Higher frequency components will be selected. This allows the filter to adjust the signal in different frequency ranges. Order of filter->The degree of steepness of the frequency response of the filter is determined. Greater->The values will result in steeper filters and more accurate selection of signals in a particular frequency range. Less->The values will result in a smoother frequency response of the filter.
Example 8: on the basis of the above embodiment, the contour wavelet transformation unit extracts a contour based on an edge detection result, and the method for performing contour wavelet transformation on the contour includes:
the contour is extracted based on the edge detection result using the following formula:
；
wherein the method comprises the steps ofRepresenting outline->And->Respectively represent the edge detection result->At->And->Gradient in direction; first, the input of the formula is an image processed by the edge detection algorithm, expressed as +.>. This image contains information about the position of the edges in the image, typically pixels with high gradient values at the edges. The core principle of the formula is based on the edge detection result +.>Contour information is calculated. It uses a gradient calculation method in image processing, wherein +.>And->Respectively represent image +.>Gradient in horizontal and vertical directions. +.>Representing profile information, calculated by: />: this item indicates that the image is +.>And->Gradient product in the direction, which helps to capture the crossover variation of the edges. When both gradients change at the same time, this term will take a large positive value. />And->These two items represent the image at +.>And->Square of gradient in direction. They help to capture the amplitude of the edge variation. When the gradient change is large, both will take a large positive value. By the above calculation, ++>The combination of cross-over variation and variation amplitude information can be regarded as the intensity of the profile. Edge +.>The value is larger, indicating that the edge is obvious, and +.>The value is small.
Calculation of profile information by formulaIt represents the intensity of the contours in the image. The value at the contour is larger and the value at the background is smaller, so that the contour in the image can be separated from the background. The formula considers gradient information of the edge in different directions, and the magnitude of the gradient. This helps emphasize the true edge features in the image and reduces the effects of non-edge noise. Since crossover variations and variations in amplitude are considered in the formula, it can be used to extract the contours of a particular shape or feature. Different profile features will be +.>Is shown in different ways. The extracted contour information may be used for further image analysis tasks such as anomaly detection, object segmentation, etc. The enhanced profile information helps to more accurately identify and segment the object.
The contours are subjected to contour wavelet transformation using the following formula:
；
wherein,is the contour of the contour wavelet transform, +.>Representing wavelet scale, +.>Is a wavelet function, +.>The upper limit of the coordinate value is the pixel abscissa or ordinate.
First, the contour extraction formula has calculated contour information in the imageThis is an image representing a contour in which the pixel values of the edge regions are higher and the pixel values of the background regions are lower. Wavelet transformation is a multi-scale analysis method that decomposes a signal or image into components of different scales and frequencies. The purpose of the contour wavelet transformation here is to add contour information +.>And converted to the wavelet domain for further analysis. +.>Is a wavelet function that is used to decompose the profile information at different locations and scales. The wavelet function determines how the signal is decomposed and reconstructed. Parameter->Representing the scale of the wavelet transform. Less->The values correspond to the high frequency details, but are larger +.>The values correspond to low frequency components and can therefore be used to control the degree of resolution. Double summation in the formula is +/for each pixel of contour information>A series of wavelet transform operations are performed, including position and scale changes. The wheels are summed by doubleThe profile information is decomposed and reorganized at different locations and scales.
The wavelet transform allows for multi-scale analysis of the profile information. By adjusting scale parametersFeatures of different scales in the image may be selectively analyzed. This facilitates detection and analysis of targets or structures of different sizes. The wavelet transform may extract features in the profile information. Different wavelet scales and frequency components can capture different characteristics of the contour, such as thickness of the edge, details of the texture, etc.
Example 9: on the basis of the above embodiment, after obtaining the contour of the contour wavelet transform, before filtering in the frequency domain, the thresholding is further performed on the contour of the contour wavelet transform, which specifically includes:
；
wherein,is the result of the thresholding; />Is a threshold parameter.
Specifically, first, in the above embodiment, the contour wavelet transform has been performed to obtain the wavelet domain representation of the contour. This representation contains information of the contour at different scales and frequencies. Thresholding is a common image processing technique for selectively retaining or discarding pixels based on their intensity values. Here, the result of the contour wavelet transformation +.>Will be thresholded to reduce noise or preserve profile information of interest. The key parameter for thresholding is the threshold parameter +.>. It decides which pixel values are to be preserved and which are to be discarded. Greater->The value will retain more information and smaller +.>The value will filter out weaker signals.
By going below a threshold valueThe thresholding helps to remove noise and insignificant signals, thereby improving image quality. Thresholding can enhance the contour features of interest, preserve the pixel values of the contour region, and suppress background and noise. By appropriate selection of threshold parameters->Specific features or targets in the image can be selectively extracted, and other irrelevant information can be filtered out. By zeroing out the smaller wavelet coefficients, thresholding can achieve information compression, reducing storage and transmission costs.
Example 10: based on the above embodiment, the edge detection unit performs edge detection on the smoothed matrix using a Canny edge detection algorithm.
Specifically, the Canny edge detection algorithm is a classical image processing algorithm for detecting edges in images. It is based on a number of steps including gaussian filtering, computing gradients, non-maxima suppression and edge tracking. The Canny edge detection algorithm requires a gray scale image as input, where the smoothing matrix takes the role of the input image. The Canny edge detection algorithm detects edges in the smoothed matrix and produces a binary image in which white pixels represent detected edges and black pixels represent non-edge regions.
The Canny algorithm is widely recognized as a method capable of accurately detecting edges. It is able to identify the true edge features in the image by multi-step processing, including gradient computation and non-maxima suppression. The Canny algorithm performs gaussian filtering prior to edge detection to help combat noise in the image, thereby reducing false detections. The Canny algorithm refines the detected edges by non-maxima suppression and edge tracking, making them more accurate and fine. The binary image output by the Canny algorithm is very suitable for subsequent processing, such as contour extraction and feature analysis.
The present invention has been described in detail above. The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to facilitate an understanding of the method of the present invention and its core ideas. It should be noted that it will be apparent to those skilled in the art that various modifications and adaptations of the invention can be made without departing from the principles of the invention and these modifications and adaptations are intended to be within the scope of the invention as defined in the following claims.
Claims (10)
1. Insulator infrared damage verification system based on image processing, characterized in that, the system includes: an image acquisition section including a plurality of infrared cameras for acquiring infrared images of insulators, the insulators being divided into a plurality of areas, each insulator acquiring an area infrared image of one area of the insulator; the image fusion part is configured to fuse the regional infrared images to obtain an insulator infrared image; the image bidirectional filtering part is configured to perform bidirectional filtering on the insulator infrared image to obtain a filtered insulator image; the contour extraction part is configured to extract the contour of the filter insulator image, and specifically comprises: representing the filter insulator image as a two-dimensional matrix, performing matrix smoothing on the two-dimensional matrix, performing edge detection to obtain an edge detection result, extracting a contour, performing contour wavelet transformation on the contour, performing frequency domain filtering on the contour after the contour wavelet transformation to obtain a frequency domain image, performing inverse wavelet transformation on the frequency domain image, and reconstructing the contour; and the damage checking part is configured for calculating gradient difference values among the reconstructed contours, comparing the gradient difference values with respective corresponding preset judging values to judge whether an abnormality occurs, taking the areas of the insulators corresponding to the two contours with the abnormal difference values as abnormal areas, and judging that the abnormal areas are damaged.
2. The image processing-based insulator infrared damage inspection system as claimed in claim 1, wherein the image fusion part, the method for fusing the regional infrared image to obtain the insulator infrared image comprises the following steps: first, a target image is setFinding a regional infrared image using an image feature point matching algorithm>And target image->Corresponding feature point pair ∈>Andthe method comprises the steps of carrying out a first treatment on the surface of the Setting a transformation matrix->Using a transformation matrix->Regional infrared image +.>Transformed into the target image coordinate space.
3. The image processing-based insulator infrared damage verification system of claim 2, wherein the transformation matrixCalculated using the following formula:
；
wherein,is the number of feature point pairs +.>Is the sequence number of the feature point pair.
4. The insulator infrared damage detection system based on image processing as claimed in claim 3, wherein the image bidirectional filtering section performs bidirectional filtering on the insulator infrared image, and the method for obtaining the filtered insulator image includes: inputting the insulator infrared image into a preset bidirectional filter to obtain a filter insulator image; the bi-directional filter is expressed using the following formula:
wherein,is the pixel coordinates; />Is the pixel value in the infrared image of the insulator; />Is the pixel value of the filtered insulator image; />Is a fixed filter window comprising +.>Is a coordinate pair of (2); />Is a normalization coefficient for ensuring that the output pixel value is between 0 and 255; />Is a spatial domain weight;is a value range weight.
5. The image processing based insulator infrared damage verification system of claim 4, wherein the spatial domain weightsCalculated using the following formula:
；
wherein,is the spatial domain variance.
6. The image processing-based insulator infrared damage verification system of claim 5, wherein the value range weights are calculated using the following formula:
；
wherein,representing the difference between pixel values; />Is the custom value range standard deviation.
7. The image processing-based insulator infrared damage verification system of claim 6, wherein the contour extraction section includes: an image conversion unit configured to represent the filtered insulator image as a two-dimensional matrix; the smoothing unit is configured to perform matrix smoothing on the two-dimensional matrix to obtain a smoothed matrix; an edge detection unit configured to perform edge detection on the smooth matrix, the edge detection result; a contour wavelet transformation unit configured to extract a contour based on an edge detection result, and perform contour wavelet transformation on the contour; and the frequency domain filtering unit is configured to perform frequency domain filtering on the contour after the contour wavelet transformation by using a frequency domain filter to obtain a frequency domain image, wherein the frequency domain filter is expressed by using the following formula:
；
wherein the method comprises the steps ofIs the response of the filter, +.>Is a distance function in the frequency domain, +.>Is a cut-off frequency, +.>Is the order of the filter; />And->Is a frequency domain coordinate representing different frequency components; contour reconstructionAnd a unit configured to inverse wavelet transform the frequency domain image line to reconstruct a contour.
8. The image processing-based insulator infrared damage inspection system of claim 7, wherein the contour wavelet transformation unit extracts a contour based on an edge detection result, and the contour wavelet transformation method comprises:
the contour is extracted based on the edge detection result using the following formula:
；
wherein the method comprises the steps ofRepresenting outline->And->Respectively represent the edge detection result->At->And->Gradient in direction;
the contours are subjected to contour wavelet transformation using the following formula:
；
wherein,wheel for contour wavelet transformationProfile (I)>Representing wavelet scale, +.>Is a wavelet function, +.>The upper limit of the coordinate value is the pixel abscissa or ordinate.
9. The image processing-based insulator infrared damage detection system of claim 8, wherein after obtaining the contour of the contour wavelet transform, the contour of the contour wavelet transform is further thresholded before filtering in the frequency domain, and specifically comprises:
；
wherein,is the result of the thresholding; />Is a threshold parameter.
10. The image processing-based insulator infrared damage detection system of claim 9, wherein the edge detection unit performs edge detection on the smoothed matrix using a Canny edge detection algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311454353.4A CN117173190B (en) | 2023-11-03 | 2023-11-03 | Insulator infrared damage inspection system based on image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311454353.4A CN117173190B (en) | 2023-11-03 | 2023-11-03 | Insulator infrared damage inspection system based on image processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117173190A true CN117173190A (en) | 2023-12-05 |
| CN117173190B CN117173190B (en) | 2024-02-02 |
Family
ID=88941735
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311454353.4A Active CN117173190B (en) | 2023-11-03 | 2023-11-03 | Insulator infrared damage inspection system based on image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117173190B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050110988A1 (en) * | 2003-11-05 | 2005-05-26 | Hidetoshi Nishiyama | Method and apparatus for inspecting defects of patterns |
| CN103700072A (en) * | 2013-12-17 | 2014-04-02 | 北京工业大学 | Image denoising method based on self-adaptive wavelet threshold and two-sided filter |
| CN107578418A (en) * | 2017-09-08 | 2018-01-12 | 华中科技大学 | A kind of indoor scene profile testing method of confluent colours and depth information |
| CN107729907A (en) * | 2016-08-12 | 2018-02-23 | 南京理工大学 | A kind of fault recognition method based on infra-red thermal imaging system |
| CN108830819A (en) * | 2018-05-23 | 2018-11-16 | 青柠优视科技（北京）有限公司 | A kind of image interfusion method and device of depth image and infrared image |
| CN111260616A (en) * | 2020-01-13 | 2020-06-09 | 三峡大学 | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization |
| CN112001260A (en) * | 2020-07-28 | 2020-11-27 | 国网湖南省电力有限公司 | Cable trench fault detection method based on infrared and visible light image fusion |
| CN113469989A (en) * | 2021-07-14 | 2021-10-01 | 广东电网有限责任公司 | Method, system, equipment and medium for extracting power transmission conductor in remote sensing image |
| CN113592729A (en) * | 2021-06-30 | 2021-11-02 | 国网吉林省电力有限公司延边供电公司 | Infrared image enhancement method for electrical equipment based on NSCT domain |
| CN113837974A (en) * | 2021-09-28 | 2021-12-24 | 国网上海市电力公司 | NSST (non-subsampled contourlet transform) domain power equipment infrared image enhancement method based on improved BEEPS (Bayesian particle swarm optimization) filtering algorithm |
| CN113870135A (en) * | 2021-09-28 | 2021-12-31 | 国网上海市电力公司 | NSST domain infrared image enhancement method based on longicorn stigma optimization algorithm |
| CN114155161A (en) * | 2021-11-01 | 2022-03-08 | 富瀚微电子(成都)有限公司 | Image denoising method and device, electronic equipment and storage medium |
| CN114353880A (en) * | 2022-01-21 | 2022-04-15 | 国网河南省电力公司电力科学研究院 | Strain insulator string wind-induced vibration online monitoring system and method |
| CN115661044A (en) * | 2022-09-30 | 2023-01-31 | 国网山西省电力公司大同供电公司 | Multi-source fusion-based substation power equipment fault detection method |
| CN116167999A (en) * | 2023-02-22 | 2023-05-26 | 南昌大学 | Distribution line zero-value insulator infrared thermal image detection method based on image matching |
-
2023
- 2023-11-03 CN CN202311454353.4A patent/CN117173190B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050110988A1 (en) * | 2003-11-05 | 2005-05-26 | Hidetoshi Nishiyama | Method and apparatus for inspecting defects of patterns |
| CN103700072A (en) * | 2013-12-17 | 2014-04-02 | 北京工业大学 | Image denoising method based on self-adaptive wavelet threshold and two-sided filter |
| CN107729907A (en) * | 2016-08-12 | 2018-02-23 | 南京理工大学 | A kind of fault recognition method based on infra-red thermal imaging system |
| CN107578418A (en) * | 2017-09-08 | 2018-01-12 | 华中科技大学 | A kind of indoor scene profile testing method of confluent colours and depth information |
| CN108830819A (en) * | 2018-05-23 | 2018-11-16 | 青柠优视科技（北京）有限公司 | A kind of image interfusion method and device of depth image and infrared image |
| CN111260616A (en) * | 2020-01-13 | 2020-06-09 | 三峡大学 | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization |
| CN112001260A (en) * | 2020-07-28 | 2020-11-27 | 国网湖南省电力有限公司 | Cable trench fault detection method based on infrared and visible light image fusion |
| CN113592729A (en) * | 2021-06-30 | 2021-11-02 | 国网吉林省电力有限公司延边供电公司 | Infrared image enhancement method for electrical equipment based on NSCT domain |
| CN113469989A (en) * | 2021-07-14 | 2021-10-01 | 广东电网有限责任公司 | Method, system, equipment and medium for extracting power transmission conductor in remote sensing image |
| CN113837974A (en) * | 2021-09-28 | 2021-12-24 | 国网上海市电力公司 | NSST (non-subsampled contourlet transform) domain power equipment infrared image enhancement method based on improved BEEPS (Bayesian particle swarm optimization) filtering algorithm |
| CN113870135A (en) * | 2021-09-28 | 2021-12-31 | 国网上海市电力公司 | NSST domain infrared image enhancement method based on longicorn stigma optimization algorithm |
| CN114155161A (en) * | 2021-11-01 | 2022-03-08 | 富瀚微电子(成都)有限公司 | Image denoising method and device, electronic equipment and storage medium |
| CN114353880A (en) * | 2022-01-21 | 2022-04-15 | 国网河南省电力公司电力科学研究院 | Strain insulator string wind-induced vibration online monitoring system and method |
| CN115661044A (en) * | 2022-09-30 | 2023-01-31 | 国网山西省电力公司大同供电公司 | Multi-source fusion-based substation power equipment fault detection method |
| CN116167999A (en) * | 2023-02-22 | 2023-05-26 | 南昌大学 | Distribution line zero-value insulator infrared thermal image detection method based on image matching |
Non-Patent Citations (4)
| Title |
|---|
| ZHONGHE REN 等: "State of the Art in Defect Detection Based on Machine Vision", 《INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING-GREEN TECHNOLOGY》, pages 1 - 31 * |
| 李顺远: "电力巡检图像中绝缘子故障识别方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》, no. 02, pages 042 - 918 * |
| 杨蔺: "航拍绝缘子图像识别及故障检测方法设计与研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》, no. 02, pages 042 - 273 * |
| 石杰: "基于梯度图像融合的接触网绝缘子故障检测", 《红外技术》, vol. 45, no. 10, pages 1106 - 1117 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117173190B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921176B (en) | Pointer instrument positioning and identifying method based on machine vision | |
| Ziou et al. | Edge detection techniques-an overview | |
| Acharjya et al. | Study and comparison of different edge detectors for image segmentation | |
| Juneja et al. | Performance evaluation of edge detection techniques for images in spatial domain | |
| Yang et al. | Defect detection in magnetic tile images based on stationary wavelet transform | |
| GB2569751A (en) | Static infrared thermal image processing-based underground pipe leakage detection method | |
| Krishnan et al. | A survey on different edge detection techniques for image segmentation | |
| Yang et al. | Crack detection in magnetic tile images using nonsubsampled shearlet transform and envelope gray level gradient | |
| Rani et al. | Detection of edges using mathematical morphological operators | |
| Kumar et al. | A conventional study of edge detection technique in digital image processing | |
| Reddy et al. | Comparative analysis of common edge detection algorithms using pre-processing technique | |
| Bhadauria et al. | Comparison between various edge detection methods on satellite image | |
| CN110490872B (en) | Foreign matter detection method and system for processing equipment | |
| CN109948629B (en) | GIS equipment X-ray image fault detection method based on SIFT features | |
| CN104966283A (en) | Imaging layered registering method | |
| CN111221996A (en) | Instrument screen visual detection method and system | |
| CN117197700A (en) | Intelligent unmanned inspection contact net defect identification system | |
| CN117173190B (en) | Insulator infrared damage inspection system based on image processing | |
| Rana et al. | Review on traditional methods of edge detection to morphological based techniques | |
| Zheng et al. | Research on edge detection algorithm in digital image processing | |
| Islam et al. | A study on edge detection techniques for natural image segmentation | |
| Wang et al. | Improved morphological band-pass filtering algorithm and its application in circle detection | |
| CN114926446A (en) | Belt tearing edge detection method | |
| Janoriya et al. | Critical review on edge detection techniques in spatial domain on low illumination images | |
| Ye et al. | Research on detection method of tower corrosion based on hough transform |
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 |