OTDR Event Detection Using Wavelet Scalogram Peak Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OTDR devices face a trade-off between improving the Signal-to-Noise (SN) ratio and maintaining distance resolution, making it difficult to accurately detect event occurrence locations such as defects or open ends in optical fibers.
Innovation Solution
The proposed solution involves analyzing OTDR waveforms using wavelet transform to generate a scalogram, calculating noise threshold values, and extracting peaks from the scalogram to create a peak graph, which is then used to identify events in the optical fiber.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a narrower window width of moving average is used, then distance resolution is improved, but noise influence increases and SN ratio deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the signal processing into multiple stages: first applying moving average with a specific window width to reduce noise, then performing wavelet transform to analyze the smoothed signal at multiple scales. This multi-stage approach separates noise reduction from feature extraction, allowing each stage to optimize for its specific function without compromising the other.
Solution Approach 2:
The patent transitions from one-dimensional time-domain analysis to two-dimensional time-scale analysis using wavelet transform. By adding the scale dimension, the system can analyze signals at different resolutions simultaneously, enabling noise suppression while preserving distance resolution through multi-scale feature extraction rather than being constrained to a single window width.
2Object-affected harmful factors
If a wider window width of moving average is used, then noise influence is reduced and SN ratio is improved, but distance resolution deteriorates
Solution Approach 1:
The patent employs dynamic analysis through wavelet transform, which provides variable resolution across different scales. At coarse scales, the transform captures overall trends with noise suppression, while at fine scales, it resolves detailed features with high distance resolution. This dynamic multi-scale approach adapts the analysis granularity to the specific features being detected, overcoming the static limitation of fixed window width.
Solution Approach 2:
The wavelet transform decomposes the signal into different frequency bands and scales, segmenting the analysis into multiple resolution levels. This allows the system to apply appropriate analysis granularity to different parts of the signal - coarser analysis for noise-prone regions and finer analysis for feature-critical regions - thereby simultaneously achieving noise reduction and distance resolution.
3Reliability
If moving average with fixed window width is used, then either SN ratio or distance resolution can be improved, but both cannot be improved simultaneously
Solution Approach 1:
The patent adds the scale dimension through wavelet transform, moving from fixed one-dimensional moving average to multi-dimensional time-scale analysis. This dimensional expansion enables the system to achieve both high SN ratio and distance resolution by analyzing the signal at multiple scales simultaneously, where each scale contributes different aspects of the signal characteristics.
Solution Approach 2:
The patent changes the analysis parameter from fixed window width to variable scale through wavelet transform. By using scales that vary by a factor of 2^(1/2), the system dynamically adjusts the effective analysis window size according to the frequency content, achieving optimal balance between noise suppression and resolution for different signal features rather than being constrained by a single fixed parameter.
Data Source
AI summary
An object of the present disclosure is to improve an SN ratio while maintaining a distance resolution and to accurately detect an event occurrence location. An event detection device according to the present disclosure includes: an OTDR waveform acquisition unit that acquires an OTDR waveform of an optical fiber to be measured; a feature quantity extraction unit that performs wavelet transform on the OTDR waveform and generates a scalogram with each wavelet coefficient as a feature quantity; a peak extraction unit that calculates a noise threshold value from the OTDR waveform, and extracts a peak from the feature quantity on the scalogram based on the noise threshold value to generate a peak graph; and an event identification unit that identifies an event in the optical fiber to be measured from the peak graph.