Suppressing ATEM In-Band Vibration Noise via Neural Network Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Airborne transient electromagnetic in-band vibration noise cannot be effectively suppressed by traditional motion noise suppression methods due to its frequency range overlapping with the useful signal, making spectral separation difficult.
Innovation Solution
The method involves dividing ATEM signals into segments, processing the noise segment to limit bandwidth, training a wavelet neural network using the processed data to predict IBV noise, and subtracting the predicted noise from the useful signal segment to suppress the in-band noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional spectral separation methods (high-pass filtering, polynomial fitting, wavelet transform) are used to suppress motion-induced noise, then low-frequency noise outside the signal band can be removed, but in-band vibration noise within the ATEM detection frequency range cannot be effectively suppressed
Solution Approach 1:
The patent divides the ATEM observation data into multiple time segments, separating the useful signal segment from the pure noise segment. This segmentation allows different processing strategies to be applied to different parts of the data, enabling effective suppression of in-band vibration noise while preserving the useful signal characteristics.
Solution Approach 2:
The patent introduces a wavelet neural network as an intermediary tool to predict and remove in-band vibration noise. The network is trained on processed noise segment data and then applied to predict noise in the useful signal segment, acting as a mediator between the raw data and the cleaned output.
2Measurement precision
If the bandwidth of processed data is limited to be just greater than the bandwidth of IBV noise, then the noise can be characterized more accurately, but information loss may occur
Solution Approach 1:
The patent applies partial action by limiting the bandwidth of processed data to be just greater than the bandwidth of IBV noise. This selective bandwidth limitation focuses processing resources on the relevant noise frequencies while avoiding unnecessary processing of out-of-band frequencies, thereby characterizing noise accurately without excessive information loss.
3Measurement precision
If a wavelet neural network is trained using processed noise segment data, then prediction accuracy for IBV noise is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary action by training the wavelet neural network in advance using processed noise segment data before applying it to the useful signal segment. This pre-training phase allows the network to learn noise characteristics thoroughly, improving prediction accuracy when actually processing the valuable signal data, thereby reducing overall processing time for the critical signal portion.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively suppresses airborne transient electromagnetic in-band vibration noise, improving data quality and laying a foundation for subsequent processing and interpretation.
Implementation Method 1
training a wavelet neural network using the processed data to predict IBV noise
Implementation Method 2
training a wavelet neural network using the processed data to predict IBV noise
Implementation Method 3
the main frequency range of the motion-induced noise is lower than the ATEM detection fundamental frequency and generally does not overlap with the frequency range of the ATEM detection useful signal, such that the suppression can be easily achieved by spectral separation
Implementation Method 4
processing the data of the segment B, limiting the bandwidth of the data of the segment B to be just greater than the bandwidth of the IBV noise
Data Source
AI summary
Disclosed in the present invention is a method for suppressing airborne transient electromagnetic in-band vibration noise, comprising: dividing the data after current turn-off into two segments according to whether the useful signal is attenuated to the system noise level: the segment A is the useful signal segment, and the segment B is the pure noise segment; limiting the bandwidth of the data of the segment B according to the frequency range of the in-band noise, and labeling the result as BL; training a neural network using the BL, utilizing the well trained neural network to predict the in-band vibration noise contained in the data of the segment A, and labeling the prediction result as PNA; and subtracting the PNA from the data of the segment A to suppress the in-band vibration noise contained in the data of the segment A.