A Feature Extraction and Classification Method of Optical Fiber Vibration Signal Based on dwt-dfpa-gbdt
A vibration signal and optical fiber vibration technology, which is applied in the directions of instruments, calculations, character and pattern recognition, etc., can solve the problems of poor generalization ability of the model, missing values, outliers, unbalanced sample data, etc., to reduce the dimension of feature vectors, The effect of high classification accuracy and high classification accuracy
Active Publication Date: 2020-12-15
BEIJING UNIV OF POSTS & TELECOMM
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the generalization ability of the model is poor, especially some models are very sensitive to problems such as missing values, outliers, and imbalanced sample data.
In addition, overfitting may occur during model training
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0049] specific implementation plan
[0050] The present invention will be described in further detail below through examples of implementation.
[0051] The data set selected in this implementation case includes the optical fiber vibration signals in four situations: knocking, climbing, vehicle passing and natural state. The number of acquisitions of each type of vibration signal is 50 times, and the acquisition frequency is 2KHz, corresponding to four types of vibrations Signal, a total of 200 sets of experimental data. Divide each group of data into 10 segments, divide 1 to 5 segments into one sample, 2 to 6 segments into one sample, and so on. Each group of signals can get 6 samples, and 50 sets of data can get 300 samples. Therefore, the total number of samples in the data set is 1200, 960 of which are randomly selected as training samples, and the remaining 240 are used as test samples.
[0052] The overall flow of the optical fiber vibration signal feature extraction ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The present invention relates to the fiber optic vibration signal feature extraction and classification method based on DWT-DFPA-GBDT, is a kind of method that fiber optic vibration signal is carried out feature extraction and classification, belongs to the field of signal processing and machine learning, is characterized in that adopting following steps: ( 1) determine the time domain feature of the original vibration signal; (2) determine the frequency domain feature of the original vibration signal; (3) determine the wavelet domain feature of the original vibration signal; (4) calculate the Mahalanobis distance; (5) obtain the feature vector by Sorting of separability; (6) Determine the loss function; (7) Determine the fitting function; (8) Determine the strong learning classifier of the current iteration. The present invention realizes feature extraction in three different domains of time domain, frequency domain and wavelet domain, constructs a complete vibration signal feature vector, combines GBDT and DWT, and provides a guarantee for the field of vibration signal feature extraction and classification A method for reducing model complexity based on classification accuracy.
Description
technical field [0001] The invention relates to the fields of signal processing and machine learning, and mainly relates to a method for feature extraction and classification of optical fiber vibration signals. Background technique [0002] At present, the problem of feature extraction and classification of optical fiber vibration signals is mainly realized by using traditional machine learning algorithms. The general processing flow is to first denoise the signal, decompose it, extract useful features, and finally train the model according to the extracted features to achieve classification. For the vibration signal feature extraction problem, the commonly used feature extraction mainly includes the extraction of time domain and frequency domain features. Due to the randomness of optical fiber vibration signals, the features extracted in the time domain and frequency domain may not be able to fully represent the vibration signal. The case of complex distribution features. ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00
CPCG06F2218/06G06F2218/08G06F2218/12
Inventor 王松胡燕祝刘娜熊之野
Owner BEIJING UNIV OF POSTS & TELECOMM