Fluorescent fingerprint intelligent analysis system for catenary system

Abstract

The application relates to the technical field of detection of electrified railway contact network systems, and discloses a contact network system fluorescent fingerprint intelligent analysis system, which comprises a visual unit, an emitting unit, a deflection unit, a receiving unit, a detection unit and a main control unit. The visual unit calculates relative motion parameters of insulators, drives the deflection unit to perform reverse compensation motion, so that a laser landing point remains stationary relative to a target surface within an integration time. The detection unit respectively collects parallel and vertical polarization components of Rayleigh scattering light and Stokes luminescence after spectral separation. The main control unit calculates a surface roughness factor by using the Rayleigh scattering component, carries out numerical correction on a polarization anisotropy parameter of the Stokes luminescence, eliminates the influence of geometric depolarization, and extracts fluorescent lifetime and rotation-related time characteristics. The application solves the fluorescent detection distortion problem caused by rough surfaces in a dynamic environment, and realizes multi-dimensional diagnosis of the contamination type and the substrate aging degree of the insulator surface.

Description

Technical Field

[0001] This invention relates to the field of detection technology for electrified railway catenary systems, specifically to an intelligent analysis system for fluorescent fingerprints of catenary systems. Background Technology

[0002] The operational status of insulators in the overhead contact system of electrified railways directly affects the safety and stability of the power supply system. Long-term exposure to the outdoor environment causes dirt to accumulate on the surface of insulators, and the substrate is also affected by ultraviolet radiation, temperature changes, and electric field stress, leading to aging. When insulation performance deteriorates, flashover or breakage accidents are highly likely to occur. Therefore, non-contact detection of the contaminant composition on the insulator surface and the aging status of the internal substrate is an important requirement for railway operation and maintenance.

[0003] Currently, non-contact inspection of insulators in overhead contact systems mainly relies on visible light image recognition and ultraviolet discharge imaging technology. Visible light imaging technology primarily identifies surface damage or large-area contamination through appearance, color, and texture features, but it struggles to distinguish the specific chemical composition of the contaminant and cannot detect microscopic physical changes such as hardening or embrittlement within the insulator substrate. While ultraviolet imaging technology can detect early corona discharge phenomena, it only reflects existing electrical insulation defects and lacks early warning capabilities for latent faults where material properties have already degraded before discharge occurs.

[0004] Laser-induced fluorescence (LAF) technology can analyze the spectral characteristics of stimulated emission to elucidate the chemical composition and molecular dynamics of materials, showing great potential for application in materials characterization. Among these, fluorescence anisotropy parameters are key indicators characterizing molecular rotational diffusion capabilities, directly related to the microviscosity and crosslinking density of polymer materials, and theoretically can be used to assess the aging degree of insulator substrates.

[0005] However, there are technical obstacles when applying this technology to the detection of vehicle-mounted contact network systems. The surface of insulators usually has high roughness and is covered with an uneven layer of dirt, which causes multiple scattering of incident and outgoing light at the microscopic interface. The multiple scattering effect causes the polarization state of light to become randomized, resulting in geometric depolarization. This depolarization caused by surface morphology is mixed with rotational depolarization caused by molecular thermal motion, making it impossible for traditional fluorescence detection equipment to accurately separate the true molecular dynamics parameters, leading to misjudgment of the aging state of the substrate.

[0006] Furthermore, vehicle-mounted detection is performed under high-speed moving conditions. The weak fluorescence signal requires a certain integration time to obtain a sufficient signal-to-noise ratio, and the relative motion between the system and the target causes measurement point drift, making it difficult to achieve spatial alignment during multi-pulse accumulation, further reducing the reliability of the detection data. Existing technologies have not yet effectively solved the problem of accurate demodulation of fluorescent fingerprints on rough surfaces under dynamic environments. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides an intelligent fluorescent fingerprint analysis system for overhead contact lines, which solves the problem that in dynamic detection environments, the surface roughness of insulators causes distortion in fluorescence anisotropy measurements, thus affecting the accuracy of insulation condition diagnosis.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solution: a fluorescent fingerprint intelligent analysis system for contact network systems, mainly composed of a vision unit, a transmitting unit, a deflection unit, a receiving unit, a detection unit, and a main control unit.

[0009] The vision unit is used to continuously acquire images of the overhead contact system, identify the location of insulator targets, and calculate the relative motion parameters of the insulator targets with respect to the system in real time.

[0010] The emitting unit is used to generate a linearly polarized near-infrared light sequence.

[0011] The deflection unit is disposed in the emission optical path and is used to drive the galvanometer to perform reverse compensation motion according to the relative motion parameters, so that the landing point of the ultraviolet pulse laser sequence remains stationary relative to the insulator target surface within the integration time window, thereby eliminating the measurement deviation caused by the system motion.

[0012] The receiving unit is used to collect the echo signal reflected by the insulator target and separate the echo signal into Rayleigh scattering light and Stokes emission in the spectral domain.

[0013] The detection unit is used to collect the parallel polarization components and vertical polarization components of Rayleigh scattered light and Stokes emission relative to the polarization direction of the emitted laser, respectively.

[0014] The main control unit is connected to the detection unit and is used to calculate the surface roughness factor and correct the fluorescence signal.

[0015] Specifically, the main control unit calculates the surface roughness factor based on the parallel polarization component and the perpendicular polarization component of the Rayleigh scattered light, and uses this surface roughness factor to correct the polarization anisotropy parameter of Stokes emission in order to remove the depolarization effect caused by the surface geometry. The main control unit further extracts rotational characteristics that characterize the physical structure and fluorescence lifetime characteristics that characterize the chemical composition, and determines the state of the insulator target based on these characteristics.

[0016] As a preferred embodiment, the galvanometer in the deflection unit performs a nonlinear deflection motion.

[0017] The main control unit calculates the deflection angle sequence based on the independent time variables within the integration time window, the horizontal and vertical distances of the insulator target, and the relative velocity along the track direction, and controls the line of sight to follow the target.

[0018] As a preferred embodiment, the emitting unit includes a solid-state laser and a Glan prism for outputting vertically linearly polarized excitation light.

[0019] The receiving unit includes a Cassegrain telescope and a dichroic mirror, which transmits Stokes light in the visible light band and reflects Rayleigh scattered light in the ultraviolet band.

[0020] As a preferred embodiment, the detection unit incorporates polarization beam splitters in both the Rayleigh detection channel and the fluorescence detection channel. The Rayleigh detection channel, in conjunction with a narrow-band bandpass filter, extracts the orthogonal components of the Rayleigh scattered light. A fluorescence detection channel, in conjunction with a long-pass filter, extracts the orthogonal components of Stokes emission. These components are then converted into photoelectric components by a gated photodetector.

[0021] As a preferred option, the main control unit performs time-domain jitter correction. In the multi-pulse accumulation processing, the main control unit uses the pulse peak time of the parallel polarization component of Rayleigh scattered light as a reference to calculate the time deviation of each acquisition, and aligns and accumulates the Stokes emission data sequence according to the time deviation.

[0022] As a preferred option, the main control unit calculates the ratio of the intensity of the vertically polarized component to the intensity of the parallel polarized component of the Rayleigh scattered light, and uses this ratio as the surface roughness factor.

[0023] The main control unit uses the surface roughness factor to construct a compensation model and calculates the molecular anisotropy attenuation curve after eliminating the influence of surface depolarization.

[0024] As a preferred option, the main control unit uses a grating factor to weight the vertical polarization component of Stokes emission, combines it with the parallel polarization component to construct a full-intensity fluorescence decay curve, and uses a double exponential decay model to fit and obtain the first fluorescence lifetime and the second fluorescence lifetime.

[0025] The main control unit performs single-exponential fitting on the molecular anisotropic decay curve and extracts rotation-related time parameters as physical structure fingerprints.

[0026] As a preferred option, the main control unit constructs a feature vector that includes fluorescence lifetime characteristics, rotational correlation time parameters, and surface roughness factors, calculates the distance between the feature vector and the cluster centers in the fault feature database, and determines the pollution type, hydrophobicity state, or aging degree of the insulator.

[0027] A second aspect of the present invention provides a method for intelligent analysis of fluorescent fingerprints in a contact network system, the method comprising the following steps: Images are acquired using a vision unit, and the relative motion parameters of the insulator target are calculated. The deflection unit is controlled to perform reverse compensation motion, so that the probe beam line of sight tracks the insulator target within the integration time window; The laser emits a linearly polarized ultraviolet pulsed laser sequence and separates and collects Rayleigh scattered light and Stokes emission from the echo signal; The parallel polarization components and vertical polarization components of Rayleigh scattered light and Stokes emission were detected respectively. The ratio of the vertical polarization component to the parallel polarization component of Rayleigh scattered light is calculated as a surface roughness factor, and this surface roughness factor is used to correct the polarization anisotropy parameter of Stokes emission. Fluorescence lifetime characteristics and rotation-related time parameters were extracted, and the insulator state was determined by combining them with the surface roughness factor.

[0028] This invention provides an intelligent analysis system for fluorescent fingerprints in overhead contact lines. It offers the following advantages: 1. This invention achieves staring tracking of high-speed moving insulator targets through the cooperation of deflection unit and vision unit. The system uses reverse compensation motion to offset the relative displacement of the detection vehicle, and constructs quasi-static measurement conditions in a dynamic environment. This ensures that the multi-pulse laser sequence accurately falls on the same position on the target surface, effectively solving the problem of spatial inconsistency in signal accumulation in high-speed detection. The signal-to-noise ratio of weak fluorescence signals is significantly improved through time-domain accumulation, ensuring the effectiveness of detection data in the vehicle environment.

[0029] 2. By quantifying the Rayleigh scattering depolarization ratio, which is sensitive to surface morphology, the system can remove the geometric multiple scattering component from the total depolarization signal and restore the true fluorescence anisotropic decay curve. This solves the technical problem that traditional fluorescence detection is difficult to distinguish between macroscopic surface scattering and microscopic molecular rotation, enabling the system to accurately obtain physical fingerprints reflecting the viscosity and crosslinking density of the insulator substrate under non-contact conditions.

[0030] 3. This invention constructs a multidimensional feature fingerprint that includes fluorescence lifetime, rotational correlation time, and surface roughness factor, enabling comprehensive diagnosis of insulator condition. The system uses fluorescence lifetime to distinguish the differences in chemical composition of biological, organic, and inorganic contaminants, and uses rotational correlation time to characterize the degree of micro-hardening and embrittlement inside the substrate. This multi-physical quantity fusion analysis method avoids the one-sidedness of single feature diagnosis and can simultaneously identify the type of external contamination on the insulator surface and the degradation of the physicochemical properties of the internal substrate. Attached Figure Description

[0031] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a schematic diagram of the visual guidance and dynamic trajectory synchronous gaze control process of the present invention; Figure 4 This is a schematic diagram of the signal acquisition timing control and time-domain preprocessing logic of the present invention; Figure 5 This is a schematic diagram of the multidimensional fingerprint demodulation and fault diagnosis process for roughness correction according to the present invention.

[0032] Among them, 100 is the main control unit; 200 is the vision unit; 201 is the camera; 202 is the image processor; 300 is the transmitting unit; 301 is the laser; 302 is the polarizer; 400 is the deflection unit; 401 is the galvanometer; 402 is the driver; 500 is the receiving unit; 501 is the telescope; 502 is the beam splitter; 600 is the detection unit; 601 is the Rayleigh detector; and 602 is the fluorescence detector. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Example: Please see the appendix Figure 1 The present invention provides a fluorescent fingerprint intelligent analysis system for overhead contact lines, comprising: a main control unit 100, a vision unit 200, a transmitting unit 300, a deflection unit 400, a receiving unit 500, and a detection unit 600.

[0035] The main control unit 100 establishes electrical connections with the vision unit 200, the transmission unit 300, the deflection unit 400, and the detection unit 600, respectively, and is used to perform system timing synchronization control, motion trajectory calculation, and spectral data analysis.

[0036] The main control unit 100 is configured as a combination architecture of field-programmable gate array and industrial computer, or an embedded processing platform.

[0037] The vision unit 200 is located at the front end of the system, and its field of view covers the preset scanning area of ​​the overhead contact system lines.

[0038] The vision unit 200 includes a camera 201 and an image processor 202. The camera 201 is used to acquire continuous image frames of the overhead contact line system in real time.

[0039] The image processor 202 is used to identify insulator targets in image frames and calculate the relative position coordinates and relative velocity vector of the insulator targets relative to the system.

[0040] The emitting unit 300 is used to generate a fluorescent probe beam for exciting the surface of the insulating medium.

[0041] The emitting unit 300 includes a laser 301 and a polarizer 302. The laser 301 is configured to emit a 355nm nanosecond-level pulsed laser. The polarizer 302 is located in the output path of the laser 301 and is used to modulate the pulsed laser into linearly polarized light.

[0042] The deflection unit 400 is disposed on the optical axis of the common optical path of the transmitting unit 300 and the receiving unit 500. The deflection unit 400 includes a galvanometer 401 and its driver 402. The driver 402 receives control commands from the main control unit 100 and controls the galvanometer 401 to perform deflection scanning in the horizontal and vertical directions to change the projection angle of the detection beam.

[0043] The receiving unit 500 is used to collect reflected light and excited fluorescence from the surface of the insulator target. The receiving unit 500 includes a telescope 501 and a beam splitter 502. The telescope 501 focuses the long-distance echo signal, and the beam splitter 502 is set in the optical path to separate the echo signal into a first optical path and a second optical path in the spectrum. The first optical path transmits Rayleigh scattered light with the same wavelength as the emission wavelength, and the second optical path transmits light with a wavelength greater than the Stokes emission wavelength.

[0044] The detection unit 600 is used to convert optical signals into electrical signals. The detection unit 600 includes a Rayleigh detector 601 and a fluorescence detector 602.

[0045] Both Rayleigh detector 601 and fluorescence detector 602 are equipped with polarization beam splitters, which decompose the incident light into a parallel component parallel to the emission polarization direction and a perpendicular component perpendicular to the emission polarization direction, respectively, and then perform photoelectric conversion through a photomultiplier tube or a single-photon detector.

[0046] See attached document Figure 2 This invention provides a method for intelligent analysis of fluorescent fingerprints in overhead contact systems, comprising the following steps: S100 and vision unit 200 continuously scan the contact network system in front. When an insulator target is detected to enter the detection range, the relative speed difference and distance parameters between the insulator target and the current detection vehicle are calculated in real time, and the motion parameters are transmitted to the main control unit 100. S200 and main control unit 100 calculate the relative motion angular velocity curve based on motion parameters and send a drive signal to deflection unit 400. The galvanometer 401 performs reverse compensation motion based on the drive signal, so that the landing point of the probe beam is relatively stationary and locked on the region of interest on the surface of the insulator target within a preset integration time window. S300. During the integration time window when the galvanometer 401 is locked, the main control unit 100 triggers the transmitting unit 300 to continuously transmit a series of pulses and sends a nanosecond-level time-gated signal to the detection unit 600 to control the detector to start acquisition only during the echo arrival time period. S400 and Rayleigh detector 601 collect the parallel and vertical components of the Rayleigh scattered light intensity corresponding to each pulse, while fluorescence detector 602 simultaneously collects the attenuation curves of the parallel and vertical components of the fluorescence intensity corresponding to each pulse. S500 and main control unit 100 perform time-domain cumulative averaging processing on multiple sets of pulse data collected within a single integration time window to generate high signal-to-noise ratio average Rayleigh scattering intensity data and average fluorescence decay waveform data. S600 and main control unit 100 use average Rayleigh scattering intensity data to calculate the physical roughness factor of the insulator surface, and use the physical roughness factor to correct the polarization anisotropy parameter in the fluorescence attenuation waveform data, eliminating the depolarization effect caused by surface geometry. S700 and main control unit 100 extract physical structural features characterizing molecular rotation properties based on the corrected polarization anisotropy parameters, and extract fluorescence lifetime features characterizing chemical components based on the total fluorescence intensity decay curve. The S800 and main control unit 100 compare the physical structure characteristics and fluorescence lifetime characteristics with the preset fault characteristic database to determine the type of contamination components, hydrophobicity status, and substrate aging degree of the insulator target surface, and output diagnostic results containing the information.

[0047] See attached document Figure 3 In this embodiment, the system establishes a nonlinear mapping relationship between the beam scanning angular velocity and the relative linear motion velocity of the vehicle, thereby achieving temporary spatial locking of the detection beam's line of sight to the moving target. The basic principle is to use the rotational motion of the galvanometer to compensate for the change rate of the viewing angle caused by the translational motion of the vehicle, so that within a preset time window, the displacement of the laser spot relative to the surface of the target being measured is less than the spot diameter, thus constructing quasi-static measurement conditions in a fast-moving dynamic scene.

[0048] The specific implementation process includes the following steps: S101, Target acquisition and region of interest extraction: The camera 201 in the vision unit 200 continuously acquires grayscale or color images of the overhead contact system line at a frame rate of not less than 60fps, and the image processor 202 performs target detection on the acquired image frames.

[0049] The image processor 202 uses a feature matching algorithm based on deep learning or an edge detection algorithm based on gradient to identify the contour features of the insulator in the image. After identifying the insulator contour, the image processor 202 determines the region of interest based on the bounding rectangle of the insulator and calculates the centroid coordinates of the region of interest based on the geometric moments of the pixel coordinate system, and uses it as the target point to be measured.

[0050] S102. Calculation of relative motion parameters. The image processor 202 calculates the relative motion vector of the insulator target relative to the detection system based on the time derivative of the target point pixel coordinates in consecutive image frames, combined with the intrinsic parameter matrix of the camera 201 and the calibrated installation height information.

[0051] The system pre-constructs a three-dimensional spatial coordinate system with the laser emission center as the origin, where the X-axis is parallel to the direction of track extension and the Z-axis is perpendicular to the track plane.

[0052] Image processor 202 outputs the current moment in real time. Horizontal distance of the target in the coordinate system Vertical distance and the relative velocity along the X-axis. The relative velocity It is the sum of the projections of the vehicle's running speed vector and the insulator's swing speed vector onto the X-axis.

[0053] S103, Reverse compensation motion curve generation: The main control unit 100 receives motion parameters from the vision unit 200 and generates the drive control function for the deflection unit 400.

[0054] In order to counteract the vehicle's displacement The resulting line-of-sight deviation requires galvanometer 401 to generate a compensation angle that varies non-linearly with time. Integration time window starting from Inside, the deflection angle of galvanometer 401 in the X-axis direction The following relationship must be satisfied: ; in: It is an independent time variable within the integration time window, and its value range is... ; in The value ranges from 5ms to 20ms, and this value is determined based on the laser's repetition rate and the required cumulative photon count signal-to-noise ratio threshold. The starting time of integration The horizontal straight-line distance between the insulator target and the system; The calculated relative tangential velocity is in The value within the window is approximated as a constant; The vertical distance between the system and the overhead contact line is obtained through laser ranging module or visual structured cursor calibration. The initial installation tilt angle of the system is the angle between the zero position of the galvanometer and the vertical direction, which is a fixed calibration value. for The mechanical deflection angle of the galvanometer relative to the optical axis at any given time.

[0055] S104. The gaze scan is executed and synchronized with the timing sequence. The main control unit 100 calculates the deflection angle sequence. The signal is converted into an analog voltage signal that varies continuously with time by a digital-to-analog converter and applied to the galvanometer 401 by a driver 402.

[0056] Under the action of the drive signal, the galvanometer 401 performs an angular velocity opposite to that of the vehicle. A rotational motion matching the rate of change of the line of sight is employed, during which the emission direction of the laser beam is continuously adjusted over time, resulting in the projection coordinates of the laser spot center onto the insulator surface. satisfy ,in It is one-tenth the diameter of the laser spot.

[0057] S105, Multi-pulse excitation window locked, the main control unit 100 will control the time period during which the galvanometer 401 performs reverse compensation motion. Defined as a gaze window, within which the main control unit 100 sends a trigger signal to the transmitting unit 300, controlling the laser 301 to operate at a repetition frequency. emission A pulse sequence.

[0058] Among them, the number of pulses From the formula It is determined that the typical value is 50 to 200 pulses. Due to the synchronization compensation effect of the galvanometer 401, this... Multiple laser pulses overlap in space and strike the same area on the surface of the insulator, achieving multiple excitation of the same physical point.

[0059] when When the current target is observed, the system terminates the observation task, the galvanometer 401 resets, and waits for the next insulator target to enter.

[0060] This embodiment employs an orthogonally cascaded optical path architecture combining spectral frequency division and polarization beam splitting. Its basic principle lies in utilizing the wavelength shift differences generated when different substances interact with light to separate elastically scattered light from inelastic fluorescence, and using polarization-selective elements to extract the spin angular momentum characteristics of photons in orthogonal geometric dimensions.

[0061] The specific implementation method is described as follows: S201. High-purity linearly polarized excitation light is generated. The laser 301 in the emitting unit 300 is a diode-pumped solid-state laser, outputting ultraviolet laser with a center wavelength of 355nm and a pulse width between 0.5ns and 1.5ns. Since the polarization degree of the laser itself is usually insufficient to support high-precision anisotropy calculations, a polarizer 302 is placed in the output path of the laser 301. The polarizer 302 is an air-gap Glan-Taylor prism, configured to modulate the polarization state of the emitted laser into a vertically linearly polarized state. The extinction ratio of the modulated excitation light is configured to be better than 10. 5 :1, to ensure that the polarization purity of the initial excitation light meets the detection requirements for trace anisotropic changes.

[0062] S202. Echo signal collection and collimation: The receiving unit 500 employs a large-aperture Cassegrain telescope as the main focusing optical system 501, utilizing its long focal length to compress the receiving field of view. A field stop is provided on the back focal plane of the main focusing optical system 501. The aperture size of this field stop matches the imaging size of the excitation spot at the target, used to perform spatial filtering and physically remove stray light from the environmental background outside the field of view. The beam passing through the field stop is transformed into a parallel beam by a collimating lens group. The divergence angle of this parallel beam is controlled within 3 mrad to meet the incident angle requirements of subsequent interference filters.

[0063] S203, Spectral dual-domain separation. Beam splitter 502 is positioned in the collimated optical path, using a dichroic mirror as the core beam-splitting element. The cutoff wavelength of the dichroic mirror is set to 360 nm, and its spectral characteristics are configured as follows: It has reflectivity for ultraviolet light with a wavelength of 355nm±2nm >99.5%, with transmittance of visible light in the wavelength range of 380nm to 700nm. >95%, after the echo signal is separated into reflected light path and transmitted light path in space by the dichroic mirror. The reflected light path transmits elastic sharp scattered light containing the physical morphology characteristics of the surface, and the transmitted light path transmits Stokes luminescence containing the chemical composition characteristics.

[0064] S204. Polarization analysis of the Rayleigh scattering channel: In the reflected light path, a narrow-band bandpass filter with a center wavelength of 355 nm and a full width at half maximum (FWHM) of 10 nm is set, with an out-of-band suppression depth greater than 4, to filter out solar background light interference. Subsequently, the beam is incident on Rayleigh detector 601. Rayleigh detector 601 contains a polarization beam splitter cube (PBS), which has an extinction ratio better than 1000:1.

[0065] PBS separates the incident light into two orthogonal components: a Rayleigh parallel component whose polarization direction is parallel to the excitation light polarization direction. ) and the Rayleigh vertical component whose polarization direction is perpendicular to the excitation light polarization direction ( .

[0066] S205. Polarization analysis of the fluorescence channel: In the transmission optical path, a long-pass filter with a cutoff wavelength of 400nm is set. The filter's blocking capability for the 355nm excitation wavelength needs to reach the OD6 level to prevent the high-energy excitation light residue from masking the weak fluorescence signal. The processed broadband fluorescence signal is incident on the fluorescence detector 602.

[0067] The fluorescence detector 602 also uses a polarizing beam splitter to separate the incident fluorescence into two orthogonal components: a parallel component whose polarization direction is parallel to the excitation light polarization direction. ) and the vertical component of fluorescence with a polarization direction perpendicular to the excitation light polarization direction ( .

[0068] S206, nanosecond-level time-resolved detection, the detection unit 600 contains four sets of independent high-sensitivity photodetectors, corresponding to the four optical path components mentioned above. The photodetectors are selected from photomultiplier tubes with gated function, and their response band covers 300nm to 650nm, with an anode pulse rise time of less than 1.5ns.

[0069] The detector's gating terminal is connected to the main control unit 100 and configured to... (The sentence is incomplete and requires more context to translate accurately.) Always open, among which The system is dynamically adjusted based on real-time ranging results. The analog signal output by the detector is transmitted to a high-speed acquisition card, which is configured with a sampling rate of no less than 1 GSps (1 billion samples per second) to fully record the transient characteristics of the fluorescence decay waveform.

[0070] See attached document Figure 4 This embodiment employs a multi-frame temporal-domain aligned cumulative averaging method. Its basic principle is based on the statistical characteristics of random noise: The signal components are coherent as pulses are repeatedly superimposed, while random noise is uncorrelated and grows at the root mean square. This is achieved by analyzing the same physical target... The system signal-to-noise ratio will be increased by accumulating repeated measurements. This increases the rate by several times, making the fluorescence decay tailing feature, which was previously submerged in noise, resolvable. The specific implementation process includes the following steps: S301. Dynamic distance gating timing generation: The timing generator inside the main control unit 100 uses the synchronization signal of the laser emission pulse as the reference time zero point. The main control unit 100 generates the timing based on the real-time vertical distance calculated by the visual unit 200 within the current gaze cycle. and horizontal distance Using the formula Calculate the time delay of the laser pulse to and from the target, where The speed is the speed of light. To shield near-field atmospheric scattering signals and suppress background noise, the system uses an electronically gated window with a specific width. The start time of the gating signal Set as: ; in, The reserved background sampling preamble time is set to a range of 10ns to 20ns. This time period is used to collect the dark current reference before the signal arrives. The time length is set to cover 3 to 5 times the fluorescence lifetime. For silicone rubber insulating materials, this value is configured to be 100ns to 200ns. The main control unit 100 sends the generated nanosecond-level TTL gated signal to the gain control terminal of the detection unit 600.

[0071] S302, Multi-channel parallel waveform digitization: During the gating window opening period, the four analog voltage signals output by the detection unit 600 enter the high-speed acquisition card. Under the FPGA logic control of the main control unit 100, the acquisition card discretizes the analog waveform with a sampling rate of not less than 1GSps and a vertical resolution of not less than 10-bit.

[0072] High vertical resolution is used to ensure that the amplitude of the weak fluorescence tail signal is greater than the minimum quantization step, for the first... laser pulses The system records lengths of The sequence of numbers, where , The sampling frequency.

[0073] S303, Time-Domain Jitter Correction Based on Rayleigh Peak. Due to nanosecond-level jitter at the laser's emission time, ... There is a slight misalignment in the data set on the timeline.

[0074] This embodiment utilizes Rayleigh scattering signals with no lifetime delay as an absolute time reference. The main control unit 100... Rayleigh parallel component sequence acquired in the second acquisition Perform a peak search to determine the index of the sampling point where the pulse peak is located. .

[0075] Then, the index is compared with the preset reference index. Deviation between The main control unit 100 determines the amount of deviation based on this deviation. A linear index shift operation is performed on the two synchronously acquired fluorescence channel data sequences, and zero-filling is applied to the resulting gaps. Precise alignment of the secondary pulse data on the time axis.

[0076] S304, pixel-level time-domain accumulation and averaging: After time-domain alignment is completed, the accumulator inside the FPGA performs pixel-level time-domain accumulation and averaging within the view window. N The digital sequence of the next pulse is used for point-to-point summation and averaging.

[0077] Average signal of fluorescence channel The calculation logic follows the formula below: in: This indicates the 1st digit after averaging. Signal strength at each sampling point, subscript Represents parallel or perpendicular polarization channels; This represents the total number of valid pulses within the gaze window, typically ranging from 50 to 200. For the first The original sampled data sequence corresponding to the next pulse; This refers to the alignment offset calculated in step S303; For the first The background baseline value collected in this study is calculated using a leading time window. Arithmetic mean of internal data: .

[0078] S305. Saturation detection and anomaly removal: During the accumulation process, the main control unit 100 monitors the signal amplitude of the Rayleigh channel in real time, and the system presets a saturation threshold. It is 95% of the full-scale voltage of the data acquisition card.

[0079] If the Rayleigh scattering intensity of a certain pulse exceeds If the data saturation occurs, the system will remove all channel data for that pulse from the accumulation queue and not include it in the total count. This is done to prevent truncated and distorted waveforms from affecting the accuracy of subsequent fitting parameters.

[0080] See attached document Figure 5 This embodiment employs a multidimensional feature decoupling algorithm with Rayleigh-fluorescence joint correction. Its basic principle is that fluorescence anisotropy attenuation reflects the rotational diffusion capability of molecules, but the photon multiple scattering effect caused by surface roughness produces additional geometric depolarization, thus falsely reducing the observed anisotropy value.

[0081] This embodiment utilizes Rayleigh scattering signals, which are sensitive to surface morphology, to quantify this geometric interference and restore the true molecular dynamics parameters from the original signal.

[0082] The specific implementation process includes the following steps: S401, Surface roughness factor calibration calculation, the main control unit 100 reads the average Rayleigh scattering parallel component intensity obtained in the previous steps. and the average Rayleigh scattering vertical component intensity In physics, backscattering from an ideal smooth surface mainly maintains the polarization state of the incident light, while a rough surface causes the incident photons to undergo multiple reflections and refractions at the microscopic interface, resulting in the randomization of the polarization state.

[0083] System defines surface roughness factor The depolarization ratio of the Rayleigh scattering signal is calculated using the following formula: in, This is the dark current noise floor value of the detector under no-light conditions, which is obtained by sampling through the lead time window in step S301.

[0084] This is a dimensionless parameter, with a value ranging from 0 to 1. The larger the value, the rougher the micro-geometry of the insulator surface.

[0085] S402, Construction of full-intensity fluorescence decay curves and chemical fingerprint extraction, with the main control unit 100 utilizing time-domain aligned parallel fluorescence components. and fluorescence vertical component Construct full-intensity decay curves to characterize the total radiant energy of fluorescent molecules. .

[0086] To eliminate the difference in transmission efficiency of detectors and optical systems for light waves with different polarization directions, a grating factor was introduced into the calculation. grating factor The system constant is pre-calibrated by measuring a standard unpolarized light source or using the tail-matching method with long-lifetime fluorescent samples. Its typical value is usually between 0.8 and 1.2. The formula for calculating the full intensity is: Subsequently, the main control unit 100 uses the Levenberg-Marquardt nonlinear least squares algorithm to perform... A double exponential decay model was fitted. The fitted model is as follows: ; In the formula, This is the residual DC background component. and These represent the amplitude weights of the two components. Through fitting calculations, the system extracts the first fluorescence lifetime. Second fluorescence lifetime .

[0087] These two parameters directly correspond to the chemical energy level structure of the luminescent group and are used to distinguish between biological, organic, and inorganic pollution.

[0088] S403. Anisotropic modulation based on roughness correction: In order to extract molecular rotation information reflecting the microscopic physical state of the insulator substrate, time-resolved fluorescence anisotropy needs to be calculated. Since multiple scattering caused by rough surfaces can lead to spurious attenuation of anisotropic values, this embodiment utilizes the roughness factor obtained in step S401. Construct a modified model to calculate the true molecular anisotropy. : In the formula, The geometric coupling coefficient is obtained through regression analysis on a set of standard samples with identical chemical composition but known surface roughness. This correction step mathematically compensates for the gain of the original measurements using a roughness factor, thus eliminating the interference of macroscopic surface morphology on the measurement of microscopic molecular rotation.

[0089] S404, Physical structure fingerprint extraction. Main control unit 100 performs correction on the anisotropic attenuation curve. Perform single-exponential fitting to extract rotational correlation times. The fitting formula is as follows: ; in, The initial anisotropy at time zero has a theoretical maximum value of 0.4; The remaining anisotropy under restricted rotation.

[0090] Physically, this corresponds to the relaxation time of fluorescent molecules rotating within an insulating medium. For polymer substrates, The value is related to the material's microviscosity and crosslinking density according to the Stokes-Einstein-Debye equation.

[0091] When insulators age, harden, become brittle, or lose their hydrophobicity, the movement of molecular chains is restricted, leading to... The value increases significantly. The system will extract... The value is defined as a physical structure fingerprint.

[0092] S405, Multidimensional State Mapping and Fault Diagnosis: The main control unit 100 constructs a multidimensional feature vector containing chemical and physical fingerprints. The system memory pre-stores a fault feature database trained on a large number of samples, which contains feature vector clustering centers corresponding to different health states.

[0093] The main control unit 100 calculates the current feature vector. Calculate the Mahalanobis distance to each cluster center and determine the target state as the closest fault type.

[0094] The specific diagnostic logic is as follows: like Within the range of 2ns to 5ns and It was determined to be covered by organic oil stains; like and It was determined to be surface powdering and aging of the substrate; like A value greater than 1.5 times the baseline health value is considered to indicate deep aging or hardening of the insulator's internal substrate. If detected The long-lived components were identified as containing biological contaminants (such as algae).

[0095] S406. Diagnostic result output: The main control unit 100 packages the determined fault type, the calculated aging level score, and the corresponding insulator geographical coordinate information into a diagnostic report, and sends it to the upper-level operation and maintenance management system through the vehicle communication interface to provide a basis for decision-making for the cleaning or replacement of the contact network system.

[0096] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A fluorescent fingerprint intelligent analysis system for overhead contact lines, characterized in that, include: The vision unit is used to continuously acquire images of the overhead contact system, identify insulator targets, and calculate the relative motion parameters of the insulator targets with respect to the system in real time. The emitting unit is used to generate a high-purity near-infrared light sequence; A deflection unit, located in the emission optical path, is used to drive the galvanometer to perform reverse compensation motion according to the relative motion parameters, so that the landing point of the near-infrared light sequence remains stationary relative to the insulator target surface within a preset integration time window. The receiving unit is used to collect the echo signal reflected by the insulator target and perform spectral separation on the echo signal to obtain Rayleigh scattered light of elastic scattering and Stokes luminescence of inelastic scattering. The detection unit is used to collect the parallel polarization component and the perpendicular polarization component of the Rayleigh scattered light and the Stokes emission light relative to the polarization direction of the emitted laser, respectively. The main control unit is used to calculate the surface roughness factor based on the parallel polarization component and the perpendicular polarization component of Rayleigh scattered light, use the surface roughness factor to correct the polarization anisotropy parameter of Stokes emission to eliminate the influence of geometric depolarization, extract rotational characteristics characterizing the physical structure and fluorescence lifetime characteristics characterizing the chemical composition, and determine the state of the insulator target based on the characteristics.

2. The contact network system fluorescent fingerprint intelligent analysis system according to claim 1, characterized in that, The galvanometer in the deflection unit is configured to generate a deflection angle that varies non-linearly with time. The deflection angle is calculated by the main control unit based on independent time variables within the integration time window, the horizontal and vertical distances of the insulator target, and the relative velocity along the track direction, in order to counteract the line-of-sight deviation caused by the system's translation.

3. The contact network system fluorescent fingerprint intelligent analysis system according to claim 1, characterized in that, The emitting unit includes a solid-state laser and a Glan prism arranged in series for outputting excitation light in a vertically linearly polarized state; The receiving unit includes a telescope and a dichroic mirror, which is configured to transmit Rayleigh scattering light in the near-infrared band and Stokes emission light in the visible band, thereby achieving spectral domain separation.

4. The contact network system fluorescent fingerprint intelligent analysis system according to claim 3, characterized in that, The detection unit includes a Rayleigh detection channel and a fluorescence detection channel; The Rayleigh detection channel is equipped with a narrow-band bandpass filter and a first polarization beam splitter to separate the parallel and perpendicular components of the Rayleigh scattered light. The fluorescence detection channel is equipped with a long-pass filter and a second polarization beam splitter to separate the parallel and vertical components of Stokes emission. Each component is coupled to a photomultiplier tube with nanosecond-level gating function for acquisition.

5. The contact network system fluorescent fingerprint intelligent analysis system according to claim 1, characterized in that, The main control unit is configured to perform time-domain jitter correction: During the multi-pulse accumulation process, the peak moment of the parallel polarization component of Rayleigh scattered light is used as the absolute time reference. The time deviation of each acquisition relative to the preset reference is calculated, and the synchronously acquired Stokes emission data sequence is indexed, shifted and aligned according to the time deviation. Then, time-domain accumulation and averaging are performed.

6. The intelligent analysis system for fluorescent fingerprints of overhead contact lines according to claim 1, characterized in that, The main control unit calculates the surface roughness factor in the following way: The intensity of the vertical polarization component and the intensity of the parallel polarization component of the Rayleigh scattered light after time-domain accumulation are obtained. After deducting the background noise, the ratio between the two is calculated. This ratio is defined as the surface roughness factor characterizing the micro-geomorphology of the insulator surface.

7. The contact network system fluorescent fingerprint intelligent analysis system according to claim 6, characterized in that, The main control unit corrects the polarization anisotropy parameters of Stokes emission in the following way: The original anisotropy value is calculated based on the measured Stokes emission parallel polarization component and vertical polarization component. A gain compensation model is constructed using the surface roughness factor to calculate the true molecular anisotropy attenuation curve after eliminating surface multiple scattering interference.

8. The contact network system fluorescent fingerprint intelligent analysis system according to claim 1, characterized in that, The main control unit extracts fluorescence lifetime characteristics in the following way: The vertical polarization component of Stokes emission is weighted using a pre-calibrated grating factor and superimposed with the parallel polarization component to construct a full-intensity fluorescence decay curve. The full-intensity fluorescence decay curve is fitted using a double exponential decay model, and the first fluorescence lifetime and the second fluorescence lifetime are analyzed as chemical fingerprints.

9. The contact network system fluorescent fingerprint intelligent analysis system according to claim 7, characterized in that, The main control unit extracts rotational characteristic features in the following way: The modified real molecular anisotropic decay curve was fitted by a single exponential method to extract the rotational correlation time parameter, which was then used as a physical structural fingerprint to characterize the microviscosity and crosslinking density of the insulator substrate.

10. The contact network system fluorescent fingerprint intelligent analysis system according to claim 9, characterized in that, The main control unit determines the target state of the insulator in the following way: A multidimensional feature vector containing the fluorescence lifetime characteristics, the rotational correlation time parameters, and the surface roughness factor is constructed. The Mahalanobis distance between the multidimensional feature vector and the cluster centers in the preset fault feature database is calculated. Based on the principle of closest distance, the type of contamination, hydrophobicity, or substrate aging degree of the insulator surface is determined.

Images