Insulator detection system and method based on multi-module fusion

Abstract

The invention provides an insulator detection system and method based on multi-module fusion. The system comprises a detection subsystem, an operation subsystem and a processing subsystem. The detection subsystem comprises a defect detection module, a filthy component content detection module and a filthy state detection module; the defect detection module is used for collecting surface image data and internal ultrasonic data of an insulator to be detected, the filth component content detection module is used for collecting hyperspectral image data of the insulator to be detected, and the filth state detection module is used for collecting ultraviolet light spot image data of the insulator to be detected; the operation subsystem is used for controlling at least one of the defect detection module, the filth component content detection module and the filth state detection module to reach a specified detection position according to the detection instruction signal; and the processing subsystem is used for processing the received data and then outputting an insulator detection result. According to the scheme provided by the invention, the multi-dimensional detection of the insulator is realized, the detection efficiency is higher, and the universality is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of insulator detection, and in particular to an insulator detection system and method based on multi-module fusion. Background Art

[0002] Insulators are widely used in power grid systems, mainly used to support conductors and line towers. At present, common insulators can be divided into three categories according to their materials: ceramic, plastic, and glass. Because they are exposed to the natural environment for a long time, dirt easily accumulates on the surface of the insulator. Once they encounter a humid environment, the dirt will dissolve in the water film and adhere to the surface of the insulator, forming an energized surface, which generally reduces the performance of the insulator and even causes flashover of dirt, causing power system operation failures and huge economic losses.

[0003] In the related technologies, insulator detection usually adopts leakage current method, microwave radiation method, hyperspectral and infrared fusion detection method, and infrared and ultraviolet fusion detection method. However, the leakage current method takes a long time to measure on site, the test equipment is expensive, the determination of alarm current requires a lot of on-site records and statistical calculations, and the measurement process is inefficient. The microwave radiation method has high professional requirements, requires professional equipment and technical support, and has poor universality. The hyperspectral and infrared fusion detection method and the infrared and ultraviolet fusion detection method can only reflect limited information about pollution, and the detection results are not accurate and comprehensive enough.

[0004] It is not difficult to find that the current traditional insulator detection solutions have the problems of low detection efficiency, poor universality, and inaccurate and incomplete detection results. Summary of the invention

[0005] The present invention provides an insulator detection system and method based on multi-module fusion, which are used to solve the defects of traditional insulator detection schemes, such as low detection efficiency, poor universality, and inaccurate and incomplete detection results.

[0006] On the one hand, the present invention provides an insulator detection system based on multi-module fusion, comprising: a detection subsystem, an operation subsystem and a processing subsystem; The detection subsystem and the operation subsystem are both connected to the processing subsystem; The detection subsystem includes a defect detection module, a contamination component content detection module and a contamination state detection module; the defect detection module is used to collect surface image data and internal ultrasonic data of the insulator to be inspected, the contamination component content detection module is used to collect high-spectral image data of the insulator to be inspected, and the contamination state detection module is used to collect ultraviolet spot image data of the insulator to be inspected; The operation subsystem is used to control at least one of the defect detection module, the contamination component content detection module and the contamination status detection module to reach a designated detection position according to the detection instruction signal after receiving the detection instruction signal sent by the processing subsystem; The processing subsystem is used to receive at least part of the surface image data, internal ultrasonic data, hyperspectral image data and ultraviolet spot image data, and output insulator detection results after processing the received data.

[0007] According to the insulator detection system based on multi-module fusion provided by the present invention, the defect detection module includes: a surface defect detection component and an internal defect detection component; The surface defect detection component comprises: a high-definition camera and an illumination light source, wherein the high-definition camera is used to capture surface image data of the insulator to be inspected under the illumination of the illumination light source; The internal defect detection component comprises: an ultrasonic transmitting probe and an ultrasonic receiver, wherein the ultrasonic transmitting probe and the ultrasonic receiver are respectively arranged on both sides of the insulator to be inspected, and the ultrasonic transmitting probe is arranged opposite to the ultrasonic receiver; The ultrasonic transmitting probe is used to transmit ultrasonic signals, and the ultrasonic receiver is used to receive ultrasonic signals containing internal ultrasonic data after passing through the insulator to be inspected.

[0008] According to the insulator detection system based on multi-module fusion provided by the present invention, the internal defect detection component further comprises: a signal analyzer and a wavelength meter; The signal analyzer and the wavelength meter are both arranged inside the ultrasonic receiver, and the signal analyzer and the wavelength meter are respectively connected to the ultrasonic receiver and the processing subsystem; The signal analyzer is used to measure the signal strength of the ultrasonic signal received by the ultrasonic receiver, and the wavelength meter is used to measure the signal wavelength of the ultrasonic signal received by the ultrasonic receiver; The processing subsystem is also used to calculate the position distance of the ultrasonic receiver according to the signal strength and the signal wavelength after receiving the signal strength and the signal wavelength, so as to determine the designated detection position of the ultrasonic receiver.

[0009] According to the insulator detection system based on multi-module fusion provided by the present invention, the contamination component content detection module includes: a fill light, a photosensitive sensor and a hyperspectrometer; The fill light is used to illuminate the insulator to be inspected according to the compensated brightness according to the ambient brightness data detected by the photosensitive sensor; The hyperspectrometer is used to capture hyperspectral image data of the insulator to be inspected under the illumination of the fill light.

[0010] According to the insulator detection system based on multi-module fusion provided by the present invention, the contamination state detection module includes: at least one ultraviolet imager; The at least one ultraviolet imager is used to capture ultraviolet spot image data of the insulator to be inspected.

[0011] According to the insulator detection system based on multi-module fusion provided by the present invention, the operation subsystem includes: a controller and a plurality of mechanical arms; The controller is connected to the plurality of mechanical arms respectively, and the defect detection module, the contamination component content detection module and the contamination state detection module are installed on at least one of the plurality of mechanical arms; The controller is used to control at least one of the multiple robot arms to operate so that at least one of the defect detection module, the contamination component content detection module and the contamination state detection module reaches a designated detection position.

[0012] According to the insulator detection system based on multi-module fusion provided by the present invention, the mechanical arm includes a base, a support column, a connecting piece, a first arm body and a second arm body; The support column is arranged on the base, one end of the connecting member is fixedly connected to the support column, the other end of the connecting member is rotatably connected to the first arm body, and the second arm body is rotatably connected to the first arm body.

[0013] According to the insulator detection system based on multi-module fusion provided by the present invention, the processing subsystem includes: a signal receiver and a data processing device; One end of the signal receiver is connected to the defect detection module, the contamination component content detection module and the contamination state detection module respectively, and the other end of the signal receiver is connected to the data processing device.

[0014] On the other hand, the present invention further provides an insulator detection method based on multi-module fusion, the method is based on any one of the above-mentioned insulator detection systems based on multi-module fusion, and the method comprises: Sending a detection instruction signal through the processing subsystem, wherein the detection instruction signal includes at least one of an insulator defect detection instruction, an insulator contamination component content detection instruction, and an insulator contamination state detection instruction; The operation subsystem receives the detection instruction signal, and controls at least one of the defect detection module, the contamination component content detection module, and the contamination status detection module to reach a designated detection position according to the detection instruction signal; The defect detection module is used to collect surface image data and internal ultrasonic data of the insulator to be inspected, and / or the contamination component content detection module is used to collect hyperspectral image data of the insulator to be inspected, and / or the contamination state detection module is used to collect ultraviolet spot image data of the insulator to be inspected; After the processing subsystem processes at least part of the surface image data, the internal ultrasonic data, the hyperspectral image data and the ultraviolet spot image data, the insulator detection result is output.

[0015] According to the insulator detection method based on multi-module fusion provided by the present invention, the surface image data is processed, including: The surface image data are subjected to linear filtering processing, surface defect edge detection, image threshold segmentation, feature extraction, image registration and image fusion to obtain surface defect information of the insulator to be inspected; Processing the internal ultrasonic data includes: Filtering the internal ultrasonic data through a plurality of cascaded digital notch filters to obtain a filtered signal; Processing the filtered signal by wavelet analysis algorithm to obtain internal defect information; Processing the hyperspectral image data includes: Performing black-and-white correction and multivariate scattering correction on the hyperspectral image data respectively to obtain a corrected image; Inputting the corrected image into a previously established and trained pollution component content prediction model to obtain the pollution component content of the insulator to be tested; Processing the ultraviolet spot image data includes: The ultraviolet spot image data are respectively subjected to color space conversion, image segmentation processing, denoising and hole filling processing, and feature extraction to obtain contamination state information.

[0016] The insulator detection system and method based on multi-module fusion provided by the present invention can collect surface image data, internal ultrasonic data, hyperspectral image data and ultraviolet spot image data of the insulator to be inspected by setting a detection subsystem, an operation subsystem and a processing subsystem, and the detection subsystem includes a defect detection module, a contamination component content detection module and a contamination state detection module. After the processing subsystem processes the received data, the obtained insulator detection result comprehensively considers various information such as surface contamination, internal contamination, contamination component content and contamination state of the insulator to be inspected, and can reflect the contamination adhesion of the insulator to be inspected from multiple dimensions, thereby improving the accuracy and comprehensiveness of the insulator detection result, and realizing multi-environment, multi-angle and multi-dimensional detection of the insulator at the same time, with higher detection efficiency and more universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the structural schematic diagrams of the insulator detection system based on multi-module fusion provided by an embodiment of the present invention; Figure 2 This is the second structural schematic diagram of the insulator detection system based on multi-module fusion provided by an embodiment of the present invention; Figure 3 is a schematic diagram of the structure of the first mechanical arm and the devices mounted thereon; Figure 4 is a schematic diagram of the structure of the second mechanical arm and the devices mounted thereon; Figure 5 It is a flow chart of an insulator detection method based on multi-module fusion provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Combine the following Figures 1 to 5 The detailed scheme of the insulator detection system and method based on multi-module fusion provided by the embodiment of the present invention is described.

[0021] Figure 1 It is a structural schematic diagram of an insulator detection system based on multi-module fusion provided by an embodiment of the present invention.

[0022] like Figure 1 As shown, the insulator detection system based on multi-module integration provided by the embodiment of the present invention specifically includes: a detection subsystem 110 , an operation subsystem 120 and a processing subsystem 130 .

[0023] The detection subsystem 110 and the operation subsystem 120 are both connected to the processing subsystem 130 .

[0024] The detection subsystem 110 includes a defect detection module 1101, a contamination component content detection module 1102 and a contamination state detection module 1103; the defect detection module 1101 is used to collect surface image data and internal ultrasonic data of the insulator to be inspected, the contamination component content detection module 1102 is used to collect high-spectral image data of the insulator to be inspected, and the contamination state detection module 1103 is used to collect ultraviolet spot image data of the insulator to be inspected.

[0025] The operation subsystem 120 is used to control at least one of the defect detection module 1101, the contamination component content detection module 1102 and the contamination state detection module 1103 to reach a designated detection position according to the detection instruction signal after receiving the detection instruction signal sent by the processing subsystem 130.

[0026] The processing subsystem 130 is used to receive at least part of the surface image data, the internal ultrasonic data, the hyperspectral image data and the ultraviolet spot image data, and output the insulator detection result after processing the received data.

[0027] It can be understood that since the detection subsystem 110 can collect surface image data, internal ultrasonic data, hyperspectral image data and ultraviolet spot image data of the insulator to be inspected, accordingly, the insulator detection result includes at least part of the surface defect information, internal defect information, contamination component content and contamination status information of the insulator to be inspected, thereby realizing multi-dimensional and efficient detection of contamination adhesion on the insulator, and the inspection result is more accurate and comprehensive.

[0028] In one embodiment, the defect detection module specifically includes: a surface defect detection component and an internal defect detection component.

[0029] like Figure 2 As shown, the surface defect detection component specifically includes: a high-definition camera 15 and an illumination light source 16 . The high-definition camera 15 is used to capture surface image data of the insulator 20 to be inspected under the illumination of the illumination light source 16 .

[0030] The internal defect detection component specifically includes: an ultrasonic transmitting probe 22 and an ultrasonic receiver 17 . The ultrasonic transmitting probe 22 and the ultrasonic receiver 17 are respectively arranged on both sides of the insulator 20 to be inspected, and the ultrasonic transmitting probe 22 and the ultrasonic receiver 17 are arranged opposite to each other.

[0031] The ultrasonic transmitting probe 22 is used to transmit ultrasonic signals, and the ultrasonic receiver 17 is used to receive ultrasonic signals containing internal ultrasonic data after passing through the insulator 20 to be inspected.

[0032] In practical applications, the illumination light source 16 may be an LED lamp, which is mainly used to illuminate the insulator to be inspected during the shooting process of the high-definition camera 15, so as to make the shooting environment brighter and improve the clarity of the surface image data obtained by shooting.

[0033] In one embodiment, the internal defect detection component may further include: a signal analyzer and a wavelength meter.

[0034] The signal analyzer and the wavelength meter are both arranged inside the ultrasonic receiver, and are connected to the ultrasonic receiver and the processing subsystem respectively.

[0035] The signal analyzer is used to measure the signal strength of the ultrasonic signal received by the ultrasonic receiver, and the wavelength meter is used to measure the signal wavelength of the ultrasonic signal received by the ultrasonic receiver.

[0036] The processing subsystem is also used to calculate the position distance of the ultrasonic receiver according to the signal strength and the signal wavelength after receiving the signal strength and the signal wavelength, so as to determine the designated detection position of the ultrasonic receiver.

[0037] It is understandable that the signal analyzer and the wavelength meter can be wirelessly connected to the processing subsystem. The settings of the signal analyzer and the wavelength meter can ensure that the ultrasonic receiver 17 is always located opposite to the ultrasonic transmitting probe 22, so that the ultrasonic receiver 17 can stably receive ultrasonic signals.

[0038] In this embodiment, the position distance required by the ultrasonic receiver can be calculated by the following formula: (1) in, is the signal strength of the ultrasonic signal received by the ultrasonic receiver, is the signal strength of the ultrasonic signal emitted by the ultrasonic transmitting probe. is the path cost exponent, usually in free space , is the wavelength, is the attenuation value caused by other factors, It is the distance between the ultrasonic receiver and the ultrasonic transmitting probe.

[0039] In one embodiment, see Figure 2 The pollution component content detection module specifically includes: a fill light 18 , a photosensor (not shown) and a hyperspectrometer 19 .

[0040] The fill light 18 is used to illuminate the insulator 20 to be inspected according to the compensated brightness based on the ambient brightness data detected by the photosensor.

[0041] The hyperspectrometer 19 is used to capture hyperspectral image data of the insulator 20 to be inspected under the illumination of the fill light 18 .

[0042] In one embodiment, see Figure 2 The contamination state detection module specifically includes: at least one ultraviolet imager 21.

[0043] At least one ultraviolet imager 21 is used to capture ultraviolet spot image data of the insulator 20 to be inspected.

[0044] Figure 2 In the system example shown, one ultraviolet imager 21 is provided. In actual applications, multiple ultraviolet imagers 21 may also be provided. The specific number of the provided ultraviolet imagers 21 may be reasonably set according to actual scene requirements and is not specifically limited here.

[0045] In one embodiment, the operation subsystem specifically includes: a controller and a plurality of robotic arms.

[0046] The controller is connected to the plurality of mechanical arms respectively, and the defect detection module, the contamination component content detection module and the contamination state detection module are installed on at least one of the plurality of mechanical arms.

[0047] The controller is used to control at least one of the multiple mechanical arms to operate so that at least one of the defect detection module, the contamination component content detection module and the contamination state detection module reaches a designated detection position.

[0048] like Figure 2 As shown, three robotic arms are provided in the present embodiment, namely a first robotic arm 7, a second robotic arm 8 and a third robotic arm 9, and all three robotic arms are connected to a controller 4. Specifically, the controller 4 is connected to the first robotic arm 7 via a first transmission optical fiber 6, the controller 4 is connected to the second robotic arm 8 via a second transmission optical fiber 60, and the controller 4 is connected to the third robotic arm 9 via a third transmission optical fiber 61.

[0049] like Figure 3 As shown, the hyperspectrometer 19, the fill light 18 and the photosensor are all installed on the first mechanical arm 7. Figure 4 As shown, the high-definition camera 15 , the lighting source 16 and the ultrasonic receiver 17 are all installed on the second robotic arm 8 , and the ultraviolet imager 21 and the ultrasonic transmitting probe 22 are all installed on the third robotic arm 9 .

[0050] In practical applications, a mounting platform may be provided on each robotic arm to facilitate the arrangement and installation of multiple devices thereon.

[0051] In a specific implementation, Figure 4 As shown, taking the second robotic arm as an example, the robotic arm specifically includes a base 10 , a support column 11 , a connecting piece 12 , a first arm body 13 and a second arm body 14 .

[0052] The support column 11 is disposed on the base 10 , one end of the connecting member 12 is fixedly connected to the support column 11 , the other end of the connecting member 12 is rotatably connected to the first arm body 13 , and the second arm body 14 is rotatably connected to the first arm body 13 .

[0053] In some embodiments, the connecting member 12 may be a connecting ring, and the supporting column 11 is connected to the connecting ring via a slot to meet the rotation requirement of the first arm body 13 .

[0054] In this embodiment, a plurality of universal wheels can be provided at the bottom of the base 10 so that the entire robot arm can move freely, or the entire robot arm can be installed on a movable pan-tilt platform through the base 10 so that the entire robot arm can move freely. By providing the connecting member 12, the first arm body 13 can rotate freely around the top of the support column 11. In addition, the second arm body 14 is rotatably connected to the first arm body 13, and the angles of the two can also be adjusted.

[0055] In some embodiments, a mounting platform may be fixedly provided on one end of the second arm body 14 away from the first arm body 13 , so as to facilitate fixing a plurality of devices on the entire robotic arm through the mounting platform.

[0056] In one embodiment, see Figure 2 The processing subsystem specifically includes: a signal receiver 3 and a data processing device 1.

[0057] One end of the signal receiver 3 is connected to the defect detection module, the contamination component content detection module and the contamination state detection module respectively, and the other end of the signal receiver 3 is connected to the data processing device 1 .

[0058] like Figure 2 As shown, the data processing device 1 is connected to the signal receiver 3 through a first signal cable 2, the data processing device 1 is connected to the controller 4 through a second signal cable 200, the signal receiver 3 is connected to the ultraviolet imager 21 through a first dedicated cable 5, the signal receiver 3 is connected to the spectrometer 19 through a second dedicated cable 50, the signal receiver 3 is connected to the high-definition camera 15 through a third dedicated cable 51, and the signal receiver 3 is connected to the ultrasonic receiver 17 through a fourth dedicated cable 52.

[0059] In practical applications, the data processing device 1 may be a computer, a server or other device having data transmission, reception and data processing functions.

[0060] Based on the same general inventive concept, the present invention also protects an insulator detection method based on multi-module fusion. The insulator detection method based on multi-module fusion provided by the present invention is described below. The insulator detection method based on multi-module fusion described below and the insulator detection system based on multi-module fusion described above can refer to each other.

[0061] Figure 5 It is a flow chart of an insulator detection method based on multi-module fusion provided in an embodiment of the present invention.

[0062] like Figure 5 As shown, the insulator detection method based on multi-module fusion provided in an embodiment of the present invention may be executed by the insulator detection system based on multi-module fusion provided in the above embodiments. The method mainly includes the following steps: Step 210: issuing a detection instruction signal through the processing subsystem, wherein the detection instruction signal includes at least one of an insulator defect detection instruction, an insulator contamination component content detection instruction, and an insulator contamination state detection instruction.

[0063] Step 220: receiving the detection instruction signal through the operation subsystem, and controlling at least one of the defect detection module, the contamination component content detection module and the contamination status detection module to reach a designated detection position according to the detection instruction signal.

[0064] Step 230: collecting surface image data and internal ultrasonic data of the insulator to be inspected through the defect detection module, and / or collecting hyperspectral image data of the insulator to be inspected through the contamination component content detection module, and / or collecting ultraviolet spot image data of the insulator to be inspected through the contamination state detection module.

[0065] Step 240: After processing at least part of the surface image data, the internal ultrasonic data, the hyperspectral image data, and the ultraviolet spot image data through the processing subsystem, the insulator detection result is output.

[0066] In one embodiment, processing the surface image data specifically includes: The surface image data are processed by linear filtering, surface defect edge detection, image threshold segmentation, feature extraction, image registration and image fusion to obtain the surface defect information of the insulator to be inspected.

[0067] Specifically, the defect types corresponding to the surface defect information of the insulator in this embodiment include: contamination, cracks, pores, air gaps, and the like.

[0068] In this embodiment, combined with Figure 2 The data processing device 1 issues an insulator surface defect detection instruction. After receiving the insulator surface defect detection instruction, the controller 4 controls the second mechanical arm 8 to reach the specified position, turns on the illumination light source 16 to provide illumination for the high-definition camera 15, and the high-definition camera 15 performs real-time detection of the surface of the insulator 20. The signal receiver 3 transmits the received surface image data to the data processing device 1 for processing.

[0069] Among them, the linear filtering process for the surface image data can be used to smooth and enhance the collected surface image data by linear filtering. The linear filtering formula is as follows: (2) in, is the filtered image data, is the surface image data before filtering, is the number of pixels in the filter template.

[0070] After that, the surface defect edge detection is performed on the filtered image data. In the surface defect edge detection link, the Sobel operator can be used to perform defect recognition and defect edge positioning on the filtered image data. The Sobel operator is specifically expressed as follows: (3) in, is the gradient of the filtered image data in the horizontal direction, is the gradient of the filtered image data in the vertical direction, yes The transposed matrix of .

[0071] Subsequently, the image data after surface defect detection is subjected to image threshold segmentation. In the image threshold segmentation stage, the image data after surface defect detection can be segmented by image threshold to separate the image foreground and image background, and highlight the features of the defects in the image. The image threshold segmentation formula is as follows: (4) in, is the image after image threshold segmentation, is the pixel value in the image before image threshold segmentation, is the image threshold.

[0072] Next, feature extraction is performed on the image after image threshold segmentation. Specifically, the image after image threshold segmentation can be segmented using the minimum covering circle method to find the smallest circle that contains the defect feature, thereby determining the boundary of the defect and extracting the shape of the defect in the image.

[0073] Finally, the image after feature extraction is subjected to image registration and image fusion. Specifically, the collected surface image data and the extracted image features can be subjected to image registration and image fusion to ensure the spatial consistency of the image information, improve the image resolution, and enhance the image defect features. If after image registration, it is found that there are obvious differences between the collected surface image data and the extracted image features, the image features need to be extracted again.

[0074] In a specific implementation, the internal ultrasonic data is processed, specifically including: Firstly, the internal ultrasonic data is filtered through multiple cascaded digital notch filters to obtain a filtered signal.

[0075] Then, the filtered signal is processed by wavelet analysis algorithm to obtain the internal defect information.

[0076] In this embodiment, combined with the attached Figure 2 The data processing device 1 issues an insulator internal defect detection instruction. After receiving the insulator internal defect detection instruction, the controller 4 controls the third mechanical arm 9 to reach the specified position, and controls the second mechanical arm 8 to move in coordination according to the calculated position information, ensuring that the ultrasonic receiver 17 and the ultrasonic transmitting probe 22 are in opposite positions. The ultrasonic transmitting probe 22 performs real-time detection on the inside of the insulator 20, and the signal receiver 3 transmits the received ultrasonic signal to the data processing device 1 for processing.

[0077] This embodiment mainly uses a digital notch filter to filter out narrowband signals in internal ultrasonic data in the ultrasonic signal. The function used by the digital notch filter is: (5) in, It is the center frequency of the digital notch filter. At this frequency, the amplitude gain of the signal is attenuated to almost zero, so the digital notch filter has a good filtering effect on narrowband noise.

[0078] In practical applications, ultrasonic signals contain interference of different frequencies. Therefore, this embodiment adopts a cascade method of multiple digital notch filters. Specifically, this embodiment cascades four digital notch filters to suppress interference of multiple frequencies. The frequency of the four-cascade filter obtained by cascading the four digital notch filters is: (6) in, is the frequency of the four-cascade filter, , , , Figure 2 is the transfer function of four cascaded digital notch filters.

[0079] Afterwards, this embodiment uses wavelet analysis to process the filtered ultrasonic signal, that is, the filtered signal. For high-frequency signals, the time window of wavelet analysis is automatically reduced, and for low-frequency signals, the time window is automatically widened. The required continuous wavelet transform function is: (7) in, are the coefficients of the wavelet transform, is the scale parameter, is the scaling factor, is the translation factor, is the original signal, that is, the filtered signal, is the complex conjugate wavelet function.

[0080] In a specific implementation, the hyperspectral image data is processed, specifically including: Firstly, the hyperspectral image data is subjected to black-and-white correction and multivariate scattering correction respectively to obtain the corrected image.

[0081] Then, the corrected image is input into a previously established and trained pollution component content prediction model to obtain the pollution component content of the insulator to be tested.

[0082] Specifically, insulators in this embodiment can be divided into ceramic insulators, glass insulators and plastic insulators according to their materials.

[0083] In this embodiment, the data processing device 1 issues an insulator contamination component content detection instruction. After receiving the insulator contamination component content detection instruction, the controller 4 controls the first mechanical arm 7 to reach the specified position and turns on the fill light 18. The fill light 18 provides lighting in real time according to the ambient light. The hyperspectrometer 19 performs real-time detection of surface contamination of the insulator 20. The signal receiver 3 transmits the collected hyperspectral image data to the data processing device 1 for component content analysis processing.

[0084] For the black-and-white correction link, this embodiment uses black-and-white correction to remove image noise and dark current in the band with weak light intensity distribution, so as to realize reflectivity correction of hyperspectral image data. The required correction algorithm is specifically as follows: (8) in, is the original hyperspectral image data, is the all-black calibration image data, is the all-white calibration image data, This is the image data after black and white correction.

[0085] In the multivariate scattering correction link, this embodiment uses multivariate scattering correction to correct the hyperspectral image data, eliminate the influence of scattering and increase the signal-to-noise ratio. The required algorithm implementation process is specifically as follows: The first step is to calculate the average spectrum. The calculation formula is as follows: (9) In the formula, represents the average spectrum, Indicates i Sample No. j The spectral value of each spectral band, n represents the number of samples selected.

[0086] In the second step, the average spectrum is used for univariate linear regression. The specific formula is as follows: (10) In the formula, Represents the average spectrum vector obtained by averaging the original spectra of all samples at each band. is the sample number ( samples), yes dimensional matrix, representing a single sample spectral vector, Represents the average spectrum vector obtained by averaging the original spectra of all samples at each band. is the sample number (i.e. samples), and Represents the hyperspectral spectrum of each sample With the average spectral vector The relative offset coefficient and translation obtained after univariate linear regression are: express dimensional spectral data matrix, is the number of samples selected, The number of bands collected for the spectrum.

[0087] The third step is to use the relative offset coefficient and translation obtained after univariate linear regression to perform multivariate scattering correction. The specific formula is as follows: (11) in, represents the spectral data matrix after multivariate scattering correction, yes dimensional matrix, representing a single sample spectral vector, and Represents the hyperspectral spectrum of each sample With the average spectral vector The relative offset coefficient and translation obtained after performing univariate linear regression.

[0088] After the correction of the above two links, the corrected image can be obtained. Subsequently, insulator models of different materials and pollution component content prediction models are established, and the above models are transfer-corrected to improve the applicability to different insulator models. The pollution component content prediction model is repeatedly trained using known pollution component content samples, and the pollution component content prediction model is verified by residual samples or cross-validation methods to evaluate the prediction accuracy and generalization ability of the pollution component content prediction model.

[0089] After obtaining the trained and verified pollution component content prediction model, the corrected image can be input into the pre-established and trained pollution component content prediction model to obtain the pollution component content of the insulator to be tested output by the pollution component content prediction model.

[0090] In some embodiments, after obtaining the content of the pollutant components, the level of the pollutant component content can be further determined. In this embodiment, the pollutant component content can be divided into four levels: A, B, C, and D according to the range of the pollutant component content. The specific level division scheme can be reasonably set according to actual application requirements and is not specifically limited here.

[0091] In a specific implementation, the ultraviolet spot image data is processed, specifically including: The ultraviolet spot image data are subjected to color space conversion, image segmentation, denoising, hole filling and feature extraction to obtain the contamination status information.

[0092] In this embodiment, the data processing device 1 issues an insulator contamination status detection instruction. After receiving the insulator contamination status detection instruction, the controller 4 controls the third robotic arm 9 to reach a specified position, and the ultraviolet imager 21 performs real-time detection of surface contamination of the insulator 20. The signal receiver 3 transmits the collected ultraviolet spot image data to the data processing device 1 for analysis and processing.

[0093] In the color space conversion link, this embodiment converts the original RGB color space corresponding to the ultraviolet spot image data into the HSI color space, which can better digitally process the image color, reduce the color information contained in the image, and thus shorten the calculation time. In practical applications, the conversion formula from RGB space to HSI space is specifically as follows: (12) (13) (14) in, Indicates red, Indicates green, Indicates blue, Indicates color tone, S Indicates saturation, I Indicates brightness.

[0094] In the image segmentation process, the maximum divergence difference threshold segmentation method can be used to segment the ultraviolet spot image data. When the inter-class variance reaches the maximum and the intra-class variance reaches the minimum, the divergence difference is the largest, and the maximum divergence difference threshold segmentation method is to set the gray value when the divergence difference reaches the maximum as the image segmentation threshold. When the variance of these two classes is the largest, it means that the separation effect is the best.

[0095] In the denoising and hole filling processing stage, this embodiment uses corrosion and expansion methods to process the ultraviolet spot image data. Corrosion can filter out redundant noise in the image, and expansion can fill small holes in the image to make the spot more complete.

[0096] In the feature extraction stage, this embodiment will extract the spot area in the ultraviolet spot image data, and pre-establish the corresponding relationship between the spot area and the contamination state. Different spot areas correspond to different contamination states, and the spot area range corresponding to different contamination states is set. In practical applications, after determining the spot area, the corresponding contamination state can be determined according to the range of the spot area.

[0097] In some embodiments, the spot area can be divided into three levels: small, medium and large. The contamination state of the insulator (20) includes four levels: level one, level two, level three and level four. The specific division details can be reasonably set according to actual needs.

[0098] Through the above process, one of the areas to be inspected on the insulator to be inspected can be inspected. After the inspection is completed, other areas to be inspected are continuously inspected until all the areas to be inspected on the insulator to be inspected are inspected.

[0099] In some embodiments, before performing insulator inspection, an insulator inspection database can be established in the data processing device 1. Each cleaning area of ​​the insulator corresponds to a detection model and is stored in the insulator inspection database. Each detection model contains key information such as the defect category of the cleaning area, the insulator contamination material, the contamination component content level and the contamination state. If an area to be inspected of an insulator is inspected and the defect category is contamination and bubbles, the insulator contamination material is ceramic insulator, the contamination component content level is Class B, the spot area is small, and the contamination state is Class III, then the data processing device 1 will match the target model in the insulator inspection database as follows: bubbles, contamination-Class B ceramic insulator-Small Class III, and each area to be inspected of the insulator will be prepared for subsequent insulator repair and laser cleaning.

[0100] When the insulator detection task is completed, the data processing device 1 will issue a stop detection instruction. After receiving the stop detection instruction, the controller 4 controls the first robotic arm 7, the second robotic arm 8 and the third robotic arm 9 to return to the starting position and stop working. The data processing device 1 will record the detection information of each area to be inspected, and then match the corresponding detection model in the database and save it, and then the data processing device 1 stops working.

[0101] In summary, the insulator detection system and method based on multi-module fusion provided by the embodiments of the present invention have at least the following advantages: First, the present invention adopts a multi-module fusion technology of an insulator defect detection module, an insulator contamination component content detection module and an insulator contamination state detection module to realize insulator detection, which can realize the detection of various information such as insulator surface defects, internal defects, contamination component content, contamination state, etc., and realizes multi-environment, multi-angle and multi-dimensional detection of insulators.

[0102] Second, before insulator inspection, the present invention can establish a huge insulator inspection database based on defect categories, insulator contamination materials, contamination component content levels and contamination status levels, and fuse a large amount of random data detected during training into a detection model and put it into the insulator inspection database. An insulator is divided into many cleaning areas, each cleaning area corresponds to a detection model, and a detection model corresponds to a set of insulator repair methods and laser cleaning parameters. The database information is continuously improved and expanded through deep learning, thereby improving the applicability and accuracy of insulator inspection.

[0103] Third, the present invention realizes detection based on three robotic arms. The robotic arms can be installed on a pan-tilt platform or the mobile wheels can be directly installed on the base. The coordinated movement of the robotic arms can realize synchronous real-time detection of the three detection modules, which effectively reduces the detection time, improves the insulator detection efficiency, and realizes efficient detection of insulators.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An insulator detection system based on multi-module fusion, characterized in that: include: Detection subsystem, operation subsystem and processing subsystem; The detection subsystem and the operation subsystem are both connected to the processing subsystem; The detection subsystem includes a defect detection module, a contamination component content detection module and a contamination state detection module; the defect detection module is used to collect surface image data and internal ultrasonic data of the insulator to be inspected, the contamination component content detection module is used to collect high-spectral image data of the insulator to be inspected, and the contamination state detection module is used to collect ultraviolet spot image data of the insulator to be inspected; The operation subsystem is used to control at least one of the defect detection module, the contamination component content detection module and the contamination status detection module to reach a designated detection position according to the detection instruction signal after receiving the detection instruction signal sent by the processing subsystem; The processing subsystem is used to receive at least part of the surface image data, internal ultrasonic data, hyperspectral image data and ultraviolet spot image data, and output insulator detection results after processing the received data.

2. The insulator detection system based on multi-module fusion according to claim 1 is characterized in that: The defect detection module includes: a surface defect detection component and an internal defect detection component; The surface defect detection component comprises: a high-definition camera and an illumination light source, wherein the high-definition camera is used to capture surface image data of the insulator to be inspected under the illumination of the illumination light source; The internal defect detection component comprises: an ultrasonic transmitting probe and an ultrasonic receiver, wherein the ultrasonic transmitting probe and the ultrasonic receiver are respectively arranged on both sides of the insulator to be inspected, and the ultrasonic transmitting probe is arranged opposite to the ultrasonic receiver; The ultrasonic transmitting probe is used to transmit ultrasonic signals, and the ultrasonic receiver is used to receive ultrasonic signals containing internal ultrasonic data after passing through the insulator to be inspected.

3. The insulator detection system based on multi-module fusion according to claim 2 is characterized in that: The internal defect detection assembly also includes: a signal analyzer and a wavelength meter; The signal analyzer and the wavelength meter are both arranged inside the ultrasonic receiver, and the signal analyzer and the wavelength meter are respectively connected to the ultrasonic receiver and the processing subsystem; The signal analyzer is used to measure the signal strength of the ultrasonic signal received by the ultrasonic receiver, and the wavelength meter is used to measure the signal wavelength of the ultrasonic signal received by the ultrasonic receiver; The processing subsystem is also used to calculate the position distance of the ultrasonic receiver according to the signal strength and the signal wavelength after receiving the signal strength and the signal wavelength, so as to determine the designated detection position of the ultrasonic receiver.

4. The insulator detection system based on multi-module fusion according to claim 1 is characterized in that: The pollution component content detection module includes: a fill light, a photosensor and a hyperspectrometer; The fill light is used to illuminate the insulator to be inspected according to the compensated brightness according to the ambient brightness data detected by the photosensitive sensor; The hyperspectrometer is used to capture hyperspectral image data of the insulator to be inspected under the illumination of the fill light.

5. The insulator detection system based on multi-module fusion according to claim 1 is characterized in that: The contamination state detection module includes: at least one ultraviolet imager; The at least one ultraviolet imager is used to capture ultraviolet spot image data of the insulator to be inspected.

6. The insulator detection system based on multi-module fusion according to claim 1 is characterized in that: The operation subsystem includes: a controller and a plurality of mechanical arms; The controller is connected to the plurality of mechanical arms respectively, and the defect detection module, the contamination component content detection module and the contamination state detection module are installed on at least one of the plurality of mechanical arms; The controller is used to control at least one of the multiple robot arms to operate so that at least one of the defect detection module, the contamination component content detection module and the contamination state detection module reaches a designated detection position.

7. The insulator detection system based on multi-module fusion according to claim 6 is characterized in that: The mechanical arm comprises a base, a support column, a connecting piece, a first arm body and a second arm body; The support column is arranged on the base, one end of the connecting member is fixedly connected to the support column, the other end of the connecting member is rotatably connected to the first arm body, and the second arm body is rotatably connected to the first arm body.

8. The insulator detection system based on multi-module fusion according to claim 1 is characterized in that: The processing subsystem includes: a signal receiver and a data processing device; One end of the signal receiver is connected to the defect detection module, the contamination component content detection module and the contamination state detection module respectively, and the other end of the signal receiver is connected to the data processing device.

9. An insulator detection method based on multi-module fusion, characterized in that: The method is based on the insulator detection system based on multi-module fusion according to any one of claims 1 to 8, and the method comprises: Sending a detection instruction signal through the processing subsystem, wherein the detection instruction signal includes at least one of an insulator defect detection instruction, an insulator contamination component content detection instruction, and an insulator contamination state detection instruction; The operation subsystem receives the detection instruction signal, and controls at least one of the defect detection module, the contamination component content detection module, and the contamination status detection module to reach a designated detection position according to the detection instruction signal; The defect detection module is used to collect surface image data and internal ultrasonic data of the insulator to be inspected, and / or the contamination component content detection module is used to collect hyperspectral image data of the insulator to be inspected, and / or the contamination state detection module is used to collect ultraviolet spot image data of the insulator to be inspected; After the processing subsystem processes at least part of the surface image data, the internal ultrasonic data, the hyperspectral image data and the ultraviolet spot image data, the insulator detection result is output.

10. The insulator detection method based on multi-module fusion according to claim 9, characterized in that: Processing the surface image data includes: The surface image data are subjected to linear filtering processing, surface defect edge detection, image threshold segmentation, feature extraction, image registration and image fusion to obtain surface defect information of the insulator to be inspected; Processing the internal ultrasonic data includes: Filtering the internal ultrasonic data through a plurality of cascaded digital notch filters to obtain a filtered signal; Processing the filtered signal by wavelet analysis algorithm to obtain internal defect information; Processing the hyperspectral image data includes: Performing black-and-white correction and multivariate scattering correction on the hyperspectral image data respectively to obtain a corrected image; Inputting the corrected image into a previously established and trained pollution component content prediction model to obtain the pollution component content of the insulator to be tested; Processing the ultraviolet spot image data includes: The ultraviolet spot image data are respectively subjected to color space conversion, image segmentation processing, denoising and hole filling processing, and feature extraction to obtain contamination state information.

Images