Eddy current data compression transmission optimization method

Abstract

The invention discloses an eddy current data compression transmission optimization method, relates to the technical field of data compression transmission, and solves the technical problems that a compression algorithm lacks scenario adaptation and data transmission lacks priority and link adaptation. Scene demand-algorithm characteristic accurate matching is achieved, compressed data overall reconstruction errors and defect area reconstruction errors are reduced, delay and compression ratio requirements of different scenes are met, data priorities are judged in a multi-dimensional mode, it is ensured that defect data acquisition resources are inclined, key information is prevented from being lost, and the method is suitable for large-scale popularization and application. A transmission protocol is optimized for a short / medium / long distance scene, the protocol adaptability and the transmission efficiency are improved, a transmission scene-link state-data priority three-dimensional linkage transmission strategy is constructed, the bandwidth occupation proportion is dynamically adjusted, core data are preferentially transmitted when links are congested, and resources are fully utilized when the links are idle.

Description

Technical Field

[0001] This invention relates to the field of data compression and transmission technology, specifically to an optimized method for eddy current data compression and transmission. Background Technology

[0002] Eddy current testing technology, with its advantages of being non-contact, having a fast response time, and being highly resistant to interference, is widely used in scenarios such as industrial equipment fault monitoring, aerospace component defect detection, and quality control of automotive parts production lines. Its core process involves collecting eddy current signals through sensors, preprocessing and compressing them, and then uploading them to a host computer or cloud platform for analysis, ultimately achieving defect identification and status warning.

[0003] Existing eddy current data processing and transmission technologies have formed preliminary solutions: the preprocessing stage mostly adopts single filtering, fixed alignment and normalization; the compression stage mostly uses general compression algorithms without being adapted to different scenarios; the transmission stage mostly uses fixed protocols without dynamically adjusting based on data value priority and real-time link status. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an optimized method for eddy current data compression and transmission, which solves the problems of compression algorithms lacking scenario-specific adaptability and data transmission lacking priority and link adaptation.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an eddy current data compression and transmission optimization method, which specifically includes the following steps:

[0006] Step 1: Collect raw eddy current data, and perform noise reduction filtering, signal alignment and data normalization on the raw data in sequence to obtain preprocessed data;

[0007] Step 2: Based on the application scenario of eddy current detection, quantify the core indicators of the scenario and select the corresponding standard compression algorithm. Perform compression processing on the preprocessed data to obtain compressed data.

[0008] Step 3: Perform structured processing, priority classification, and transmission protocol adaptation on the compressed data;

[0009] Step 4: Based on the current transmission scenario and the identified transmission status, perform differentiated transmission processing on high, medium and low priority compressed data according to the preset bandwidth occupancy ratio, queue scheduling strategy and transmission parameters to generate data transmission information.

[0010] As a further embodiment of the present invention, the denoising filtering includes a combination of adaptive median filtering and wavelet soft thresholding. The wavelet soft thresholding uses the Sym25 wavelet basis, and the threshold is set to σ=median(|x|) / 0.6745, where x is the signal after adaptive median filtering.

[0011] Signal alignment: Based on the cross-correlation algorithm, using the defect-free standard eddy current signal as a reference, the phase and time sequence of multiple sets of acquired data are aligned, and the matching threshold of the cross-correlation algorithm is ≥0.8;

[0012] Data normalization: For unsigned eddy current data, linear normalization is used, calculated using the formula x. norm =(xx min ) / (x max -x min） The data is mapped to the [0, 1] interval; for eddy current data containing positive and negative defect features, Z-score normalization is used.

[0013] As a further aspect of the present invention, the method for obtaining compressed data is as follows:

[0014] Low sampling + real-time compression scenario: The core quantitative indicators include sampling rate ≤ 60% of the target Nyquist rate, total latency of transmission + compression + reconstruction ≤ 10ms, overall reconstruction error ≤ 5% and defect detection rate ≥ 95%, and the compression algorithm is selected and compression is performed based on the indicators;

[0015] High precision + low distortion scenario: The core quantitative indicators include overall reconstruction error ≤3%, defect area reconstruction error ≤2%, signal-to-noise ratio ≥25dB, and defect feature consistency. Compression algorithms are selected and compression is performed in combination with data characteristics.

[0016] High compression ratio + complex scenarios: The core quantitative indicators include real-time scene compression ratio ≤15:1, offline scene compression ratio ≥20:1, overall reconstruction error ≤6%, and defect area reconstruction error ≤4%. Compression algorithms are selected and compression is performed in combination with scene complexity.

[0017] As a further aspect of the present invention, the specific method of priority classification is as follows:

[0018] High-priority data: defect feature matching degree ≥80%, data variance ≥3 times the mean of stationary data, reconstruction error ≤3%, and latest processing time for real-time early warning scenarios ≤10ms;

[0019] Medium priority data: defect feature matching degree ≤30%, data variance ≤2 times the mean of stationary data, reconstruction error ≤5%, latest processing time ≤50ms;

[0020] Low-priority data: defect feature matching degree = 0%, data type is auxiliary data, and there is no limit to the latest processing time.

[0021] As a further aspect of the present invention, the specific method of transmission protocol adaptation is as follows:

[0022] For short-distance transmission scenarios, a simplified version of the TCP / IP protocol is used, removing redundant TCP / IP options and reducing the packet header from 20 bytes to 8 bytes. Connections are established through a three-way handshake, and a sliding window mechanism is used for flow control.

[0023] For medium-distance transmission scenarios, a simplified UDP protocol and a custom application layer protocol are adopted. The simplified UDP protocol only retains the source port, destination port, length and checksum fields, and removes optional fields. The custom application layer protocol adds a dual verification mechanism of overall frame header verification and data segment verification.

[0024] For long-distance transmission scenarios, a simplified version of the MQTT protocol is used, with the keep-alive mechanism enabled and redundant functions such as retaining messages and will messages disabled.

[0025] As a further aspect of the present invention, the defect feature matching degree is calculated as follows:

[0026] Feature extraction is performed on the preprocessed data corresponding to the compressed data to obtain the feature vector to be matched; the same feature extraction process is performed on the sample data in the standard defect sample library to obtain the standard feature vector; feature extraction includes frequency domain feature, time-frequency domain feature or impedance feature extraction, and the cosine similarity between the feature vector to be matched and the standard feature vector is calculated, and the cosine similarity is mapped to a defect feature matching degree of 0-100%.

[0027] As a further aspect of the present invention, the method for collecting transmission link indicators and performing link status identification is as follows:

[0028] Short-distance transmission scenarios: Excellent state: U<30%, L<0.1%, D<3ms; Good state: 30%≤U<60%, 0.1%≤L<1%, 3ms≤D<8ms; Medium state: 60%≤U<80%, 1%≤L<5%, 8ms≤D<20ms; Poor state: U≥80%, L≥5%, D≥20ms;

[0029] Mid-range transmission scenario: Excellent state: U<35%, L<0.2%, D<5ms; Good state: 35%≤U<65%, 0.2%≤L<1.5%, 5ms≤D<15ms; Medium state: 65%≤U<85%, 1.5%≤L<6%, 15ms≤D<30ms; Poor state: U≥85%, L≥6%, D≥30ms;

[0030] Long-distance transmission scenarios: Excellent state: U<40%, L<0.5%, D<10ms; Good state: 40%≤U<70%, 0.5%≤L<2%, 10ms≤D<30ms; Medium state: 70%≤U<90%, 2%≤L<8%, 30ms≤D<60ms; Poor state: U≥90%, L≥8%, D≥60ms;

[0031] Where U represents bandwidth utilization, L represents packet loss rate, and D represents transmission delay.

[0032] As a further aspect of the present invention, the frame structure of fixed header + variable data segment + check bit is specifically as follows:

[0033] The fixed header includes a 2-byte device ID, a 1-byte priority identifier, a 1-byte data type identifier, a 2-byte compression parameter index, a 2-byte variable data segment length, and a 2-byte timestamp;

[0034] The variable data segment length is 64-1024 bytes, storing the compressed eddy current data body;

[0035] High-priority data uses 4-byte CRC-32 checksum, while medium-priority and low-priority data use 2-byte CRC-16 checksum.

[0036] As a further aspect of the present invention, the specific selection of the standard compression algorithm is as follows:

[0037] Low sampling + real-time compression scenario: The standard compression algorithm is compressed sensing, the sparse basis of compressed sensing is the Sym8 wavelet basis, the observation matrix is ​​a binary random matrix, the sampling rate coefficient α = 0.1-0.3, and the reconstruction algorithm is orthogonal matching pursuit.

[0038] High-precision + low-distortion scenarios: The standard compression algorithm is an improved wavelet transform, with Sym25 as the wavelet basis and 8-10 decomposition levels. An improved soft thresholding strategy is employed, with the threshold λ = σ × , where σ is the noise standard deviation and N is the data length;

[0039] High compression ratio + complex scenarios: Standard compression algorithms are CS+WT concatenated algorithm, K-SVD+CS algorithm or lightweight autoencoder deep learning algorithm.

[0040] This invention provides an optimized method for eddy current data compression and transmission. Compared with existing technologies, it has the following advantages:

[0041] This invention achieves precise matching between scenario requirements and algorithm characteristics by formulating quantitative indicators for three core scenarios and selecting exclusive standard compression algorithms. It reduces the overall reconstruction error of compressed data and the reconstruction error of defective areas, while meeting the latency and compression ratio requirements of different scenarios. It establishes a quantitative priority classification standard and determines data priority from multiple dimensions to ensure that defective data receives resource allocation and avoids the loss of key information. It optimizes transmission protocols for short / medium / long-distance scenarios to improve protocol adaptability and transmission efficiency. It establishes a standardized link status identification mechanism and constructs a three-dimensional linkage transmission strategy of transmission scenario, link status, and data priority. It dynamically adjusts the bandwidth occupancy ratio, queue scheduling method, and fault tolerance mechanism to achieve priority transmission of core data when the link is congested and full utilization of resources when the link is idle. Attached Figure Description

[0042] Figure 1 This is a flowchart of the steps in the eddy current data compression and transmission optimization method of the present invention;

[0043] Figure 2 This is a flowchart of the closed-loop control process for link coordination and scheduling in this invention. Detailed Implementation

[0044] 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.

[0045] Please see Figure 1 This application provides an optimized method for eddy current data compression and transmission, which specifically includes the following steps:

[0046] Step 1: Acquire eddy current data and preprocess it. Preprocessing includes denoising filtering, signal alignment, and data normalization to obtain preprocessed data. Denoising filtering uses adaptive median filtering combined with wavelet soft thresholding. The specific wavelet basis is Sym25 / Daubechies4, used to balance denoising and feature preservation. The threshold is σ=median. Signal alignment is based on a cross-correlation algorithm, using a defect-free standard signal as a reference, to align the phase / time sequence of multiple sets of acquired data. The cross-correlation threshold must be ≥0.8. Data normalization uses linear normalization and Z-score normalization. The linear normalization calculation formula is x... norm =(xx min ) / (x max -x min） Mapped to the [0, 1] interval, it is only applicable to unsigned data. When there are positive and negative defect signals, Z-score normalization is used.

[0047] Step 2: Acquire preprocessed data and select a compression algorithm based on the corresponding application scenario. For low sampling + real-time compression scenarios, clarify the core requirements and quantification indicators. The core requirements include sampling rate requirements, latency threshold requirements, and accuracy tolerance requirements. Specific sampling rate requirements: whether it needs to be lower than the Nyquist rate, for example, if the highest frequency of the original signal is 500kHz, the target sampling rate is ≤300kHz. Latency threshold: total transmission + compression + reconstruction latency ≤10ms. Accuracy tolerance: overall reconstruction error ≤5%, defect detection rate ≥95%. Based on the core requirements, the compression algorithm is screened to obtain a standard compression algorithm, and data compression processing is performed to obtain compressed data. The specific screening method is as follows:

[0048] Low sampling + real-time compression scenario: The standard compression algorithm is compressed sensing, the sparse basis of compressed sensing is the Sym8 wavelet basis, the observation matrix is ​​a binary random matrix, the sampling rate coefficient α = 0.1-0.3, and the reconstruction algorithm is orthogonal matching pursuit.

[0049] High-precision + low-distortion scenarios: The standard compression algorithm is an improved wavelet transform, with Sym25 as the wavelet basis and 8-10 decomposition levels. An improved soft thresholding strategy is employed, with the threshold λ = σ × , where σ is the noise standard deviation and N is the data length;

[0050] High compression ratio + complex scenarios: standard compression algorithms are CS+WT concatenated algorithm, K-SVD+CS algorithm or lightweight autoencoder deep learning algorithm;

[0051] For high-precision and low-distortion scenarios, the core indicators are first quantified, including reconstruction error: overall ≤3%, defect area ≤2%; signal-to-noise ratio: ≥25dB; defect feature consistency: peak deviation ≤5%, abrupt change time deviation ≤1 sampling point, width deviation ≤10%; data characteristics: whether it is three-dimensional eddy current field data, defect size. Based on the core indicators, the compression algorithm is screened to obtain the standard compression algorithm, and the data is compressed to obtain compressed data.

[0052] For high compression ratio and complex scenarios, key indicators are quantified, including compression ratio targets: 10:1-30:1, ≤15:1 for real-time scenarios, and ≥20:1 for offline scenarios; accuracy tolerance: overall reconstruction error ≤6%, defect area ≤4%; scenario complexity: data volume, defect type, and real-time requirements. Based on these key indicators, compression algorithms are selected to obtain standard compression algorithms, and data compression processing is performed to obtain compressed data.

[0053] Step 3: Acquire compressed data and design a structured frame format. Specifically, design the frame structure as a fixed header + variable data segment + check bit. Then, classify the compressed data into high, medium, and low priority categories. High-priority data represents data containing defect features after compression. The corresponding quantitative judgment condition includes a defect feature matching degree ≥ 80%. The specific calculation method for the defect feature matching degree is as follows: Perform feature extraction on the preprocessed data corresponding to the compressed data to obtain the feature vector to be matched; perform the same feature extraction process on the sample data in the standard defect sample library to obtain the standard feature vector; feature extraction includes frequency domain features, time-frequency domain features, or impedance feature extraction; calculate the cosine similarity between the feature vector to be matched and the standard feature vector; map the cosine similarity to a defect feature matching degree of 0~100%; and ensure that the data variance is ≥ 3 times the mean of stationary data. For data with a reconstruction error ≤3% and a latest processing time ≤10ms in real-time early warning scenarios, medium-priority data represents stable compressed data under normal operating conditions. The corresponding quantitative judgment conditions include a defect feature matching degree ≤30%, data variance ≤2 times the mean of stable data, reconstruction error ≤5%, and a latest processing time ≤50ms. Low-priority data represents redundant verification data and equipment status feedback. The corresponding quantitative judgment conditions include a defect feature matching degree of 0%, data type as auxiliary data, and no limit on the latest processing time. Furthermore, different transmission protocols are selected based on different transmission distance scenarios, and the specific selection methods are as follows:

[0054] For short-distance transmission scenarios, the TCP / IP protocol is adopted, and redundant TCP / IP options are removed. The packet header is reduced from 20 bytes to 8 bytes. A connection is established through a three-way handshake, and a sliding window mechanism is used for flow control.

[0055] For medium-distance transmission scenarios, the UDP protocol is combined with a custom application layer protocol. The UDP packet header is simplified, the source port, destination port, length, and checksum are retained, optional fields are removed, and a custom application layer adds dual verification: overall frame header verification and data segment verification.

[0056] For long-distance transmission scenarios, the MQTT protocol is adopted, the MQTT keep-alive mechanism is enabled, and redundant functions such as retaining messages and will messages are disabled.

[0057] Step 4: Based on the analyzed transmission protocol and compressed data, obtain the quantitative indicators corresponding to the transmission link, specifically including bandwidth utilization U, packet loss rate L, and transmission delay D. Use these quantitative indicators to identify the real-time transmission status of the transmission link. For short-distance transmission scenarios, if bandwidth utilization U < 30%, packet loss rate L < 0.1%, and transmission delay D < 3ms, the corresponding status level is excellent; if 30% ≤ U < 60%, 0.1% ≤ L < 1%, and 3ms ≤ D < 8ms, the corresponding status level is good; if 60% ≤ U < 80%, 1% ≤ L < 5%, and 8ms ≤ D < 20ms, the corresponding status level is medium; if U ≥ 80%, L ≥ 5%, and D ≥ 20ms, the corresponding status level is poor.

[0058] For medium-range transmission scenarios, if U < 35%, L < 0.2%, and D < 5ms, the corresponding status level is excellent; if 35% ≤ U < 65%, 0.2% ≤ L < 1.5%, and 5ms ≤ D < 15ms, the corresponding status level is good; if 65% ≤ U < 85%, 1.5% ≤ L < 6%, and 15ms ≤ D < 30ms, the corresponding status level is average; and if U ≥ 85%, L ≥ 1.5%, and D ≥ 30ms, the corresponding status level is poor.

[0059] For long-distance transmission scenarios, if U < 40%, L < 0.5%, and D < 10ms, the corresponding status level is excellent; if 40% ≤ U < 70%, 0.5% ≤ L < 2%, and 10ms ≤ D < 30ms, the corresponding status level is good; if 70% ≤ U < 90%, 2% ≤ L < 8%, and 30ms ≤ D < 60ms, the corresponding status level is medium; and if U ≥ 90%, L ≥ 8%, and D ≥ 60ms, the corresponding status level is poor.

[0060] Based on the current real-time transmission scenario and combined with the corresponding status judgment information, compressed data of different priorities are processed for transmission, and corresponding data transmission information is generated. Taking the short-distance transmission scenario as an example, if the corresponding transmission status is excellent, a simplified TCP protocol is used for high-priority compressed data, with a bandwidth usage of 30% and a PQ queue priority of level 1. For medium-priority compressed data, a simplified TCP protocol is used, with a bandwidth usage of 60% and a WFQ queue weight of 70%. For low-priority compressed data, an SPI protocol is used, with a bandwidth usage of ≤10% and an LPQ queue.

[0061] If the corresponding transmission status is good, for high-priority compressed data, a simplified TCP protocol is used, with bandwidth usage of 40% and PQ queue priority of level 1; for medium-priority compressed data, a simplified TCP protocol is used, with bandwidth usage of 60% and WFQ queue weight of 60%; for low-priority compressed data, an I2C protocol is used, with bandwidth usage of ≤10% and LPQ queue.

[0062] If the corresponding transmission status is medium, for high-priority compressed data, a simplified TCP protocol with selective retransmission is used, with 60% bandwidth usage and the PQ queue having exclusive access to the channel. For medium-priority compressed data, a simplified TCP protocol is used, with 30% bandwidth usage and the WFQ queue having a weight of 40%. For low-priority compressed data, active transmission is paused.

[0063] If the corresponding transmission status is poor, for high-priority compressed data, dual-link backup is used, bandwidth usage is 80%, and medium / low-priority channels are closed. For medium-priority compressed data, a simplified version of the TCP protocol is used, bandwidth usage is 20%, and the WFQ queue weight is 20%. For low-priority compressed data, active transmission is suspended.

[0064] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0065] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for optimizing eddy current data compression and transmission, characterized in that, The method specifically includes the following steps: Step 1: Collect raw eddy current data, and perform noise reduction filtering, signal alignment and data normalization on the raw data in sequence to obtain preprocessed data; Step 2: Based on the application scenario of eddy current detection, quantify the core indicators of the scenario and select the corresponding standard compression algorithm. For low sampling + real-time compression scenarios, the standard compression algorithm is compressed sensing. For high precision + low distortion scenarios, the standard compression algorithm is improved wavelet transform. For high compression ratio + complex scenarios, the standard compression algorithm is CS+WT cascade algorithm, K-SVD+CS algorithm or lightweight autoencoder deep learning algorithm. Perform compression processing on the preprocessed data to obtain compressed data. Step 3: Perform structured processing, priority classification, and transmission protocol adaptation on the compressed data; Step 4: Based on the current transmission scenario and the identified transmission status, perform differentiated transmission processing on high, medium and low priority compressed data according to the preset bandwidth occupancy ratio, queue scheduling strategy and transmission parameters to generate data transmission information.

2. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The denoising filtering includes a combination of adaptive median filtering and wavelet soft thresholding. The wavelet soft thresholding uses the Sym25 wavelet basis, and the threshold is set to σ=median(|x|) / 0.6745, where x is the signal after adaptive median filtering. Signal alignment: Based on the cross-correlation algorithm, using the defect-free standard eddy current signal as a reference, the phase and time sequence of multiple sets of acquired data are aligned, and the matching threshold of the cross-correlation algorithm is ≥0.8; Data normalization: For unsigned eddy current data, linear normalization is used, calculated using the formula x. norm =(xx min ) / (x max -x min） The data is mapped to the [0, 1] interval; for eddy current data containing positive and negative defect features, Z-score normalization is used.

3. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The compressed data is obtained in the following ways: Low sampling + real-time compression scenario: The core quantitative indicators include sampling rate ≤ 60% of the target Nyquist rate, total latency of transmission + compression + reconstruction ≤ 10ms, overall reconstruction error ≤ 5% and defect detection rate ≥ 95%, and the compression algorithm is selected and compression is performed based on the indicators; High precision + low distortion scenario: The core quantitative indicators include overall reconstruction error ≤3%, defect area reconstruction error ≤2%, signal-to-noise ratio ≥25dB, and defect feature consistency. Compression algorithms are selected and compression is performed in combination with data characteristics. High compression ratio + complex scenarios: The core quantitative indicators include real-time scene compression ratio ≤15:1, offline scene compression ratio ≥20:1, overall reconstruction error ≤6%, and defect area reconstruction error ≤4%. Compression algorithms are selected and compression is performed in combination with scene complexity.

4. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The specific method for priority classification is as follows: High-priority data: defect feature matching degree ≥80%, data variance ≥3 times the mean of stationary data, reconstruction error ≤3%, and latest processing time for real-time early warning scenarios ≤10ms; Medium priority data: defect feature matching degree ≤30%, data variance ≤2 times the mean of stationary data, reconstruction error ≤5%, latest processing time ≤50ms; Low-priority data: defect feature matching degree = 0%, data type is auxiliary data, and there is no limit to the latest processing time.

5. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The specific method for transmission protocol adaptation is as follows: For short-distance transmission scenarios, a simplified version of the TCP / IP protocol is used, removing redundant TCP / IP options and reducing the packet header from 20 bytes to 8 bytes. Connections are established through a three-way handshake, and a sliding window mechanism is used for flow control. For medium-distance transmission scenarios, a simplified UDP protocol and a custom application layer protocol are adopted. The simplified UDP protocol only retains the source port, destination port, length and checksum fields, and removes optional fields. The custom application layer protocol adds a dual verification mechanism of overall frame header verification and data segment verification. For long-distance transmission scenarios, a simplified version of the MQTT protocol is used, with the keep-alive mechanism enabled and the redundancy functions for retained messages and will messages disabled.

6. The eddy current data compression and transmission optimization method according to claim 4, characterized in that, The method for calculating the defect feature matching degree is as follows: Feature extraction is performed on the preprocessed data corresponding to the compressed data to obtain the feature vector to be matched; the same feature extraction process is performed on the sample data in the standard defect sample library to obtain the standard feature vector; feature extraction includes frequency domain feature, time-frequency domain feature or impedance feature extraction, and the cosine similarity between the feature vector to be matched and the standard feature vector is calculated, and the cosine similarity is mapped to a defect feature matching degree of 0-100%.

7. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The method for collecting transmission link metrics and identifying link status is as follows: Short-distance transmission scenarios: Excellent state: U < 30%, L < 0.1%, D < 3ms; Good state: 30% ≤ U < 60%, 0.1% ≤ L < 1%, 3ms ≤ D < 8ms; Medium state: 60% ≤ U < 80%, 1% ≤ L < 5%, 8ms ≤ D < 20ms; Poor state: U ≥ 80%, L ≥ 5%, D ≥ 20ms Mid-range transmission scenario: Excellent state: U<35%, L<0.2%, D<5ms; Good state: 35%≤U<65%, 0.2%≤L<1.5%, 5ms≤D<15ms; Medium state: 65%≤U<85%, 1.5%≤L<6%, 15ms≤D<30ms; Poor state: U≥85%, L≥6%, D≥30ms; Long-distance transmission scenarios: Excellent state: U<40%, L<0.5%, D<10ms; Good state: 40%≤U<70%, 0.5%≤L<2%, 10ms≤D<30ms; Medium state: 70%≤U<90%, 2%≤L<8%, 30ms≤D<60ms; Poor state: U≥90%, L≥8%, D≥60ms; Where U represents bandwidth utilization, L represents packet loss rate, and D represents transmission delay.

8. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The frame structure of fixed header + variable data segment + parity bit is as follows: The fixed header includes a 2-byte device ID, a 1-byte priority identifier, a 1-byte data type identifier, a 2-byte compression parameter index, a 2-byte variable data segment length, and a 2-byte timestamp; The variable data segment length is 64-1024 bytes, storing the compressed eddy current data body; High-priority data uses 4-byte CRC-32 checksum, while medium-priority and low-priority data use 2-byte CRC-16 checksum.

9. The eddy current data compression and transmission optimization method according to claim 1, characterized in that, The specific selection of standard compression algorithms is as follows: Low sampling + real-time compression scenario: The standard compression algorithm is compressed sensing, the sparse basis of compressed sensing is the Sym8 wavelet basis, the observation matrix is ​​a binary random matrix, the sampling rate coefficient α = 0.1-0.3, and the reconstruction algorithm is orthogonal matching pursuit. High-precision + low-distortion scenarios: The standard compression algorithm is an improved wavelet transform, with Sym25 as the wavelet basis and 8-10 decomposition levels. An improved soft thresholding strategy is employed, with the threshold λ = σ × , where σ is the noise standard deviation and N is the data length; High compression ratio + complex scenarios: Standard compression algorithms are CS+WT concatenated algorithm, K-SVD+CS algorithm or lightweight autoencoder deep learning algorithm.

Images