Noise band-limiting filtering method applied to FFT data

Abstract

The invention discloses a noise band-limiting filtering method applied to FFT data. Different filtering treatments are performed according to differences of filtering processing bands. The method comprises the steps of filtering between maximal values of limiting values, specifically comprising performing analysis filtering on signals in an amplitude section of more than a reference value dRef and lower than a maximum decrease delta in FFT signal data to be processed, and using average noise to substitute the data value in the section for processing; average noise upper band filtering, specifically comprising analyzing signals in a section of more than a set multiple of an average noise multiple factor theta in the FFT signal data to be processed, and using the average noise to substitute the data value in the section for processing; and average noise upper limit filtering, specifically comprising analyzing signals in a section of more than a set multiple of an average noise lower limit factor beta and lower than the reference value dRef in the FFT signal data to be processed, and using the average noise to substitute the data value in the section for processing. Useless signals or unconcerned signals in the known preselected band are effectively filtered.

Description

technical field [0001] The invention relates to the technical field of digital signal processing, in particular to a noise band-limited filtering method for FFT data. Background technique [0002] In the field of high-speed acquisition and processing, due to the high sampling rate of high-speed A / D, it brings great difficulties to the back-end hardware design and real-time data analysis and processing. After the sampling data is processed by FFT transformation, many times due to large bandwidth problems, signal leakage problems, signal aliasing problems, in-band self-excitation problems, etc., the effect after FFT is not ideal, some unwanted signals or irrelevant signals It will appear in the useful signal band, and the front-end hardware and programmable logic processing parts are difficult to solve under high-speed conditions, which affects the data processing effect. [0003] Usually, when the hardware environment is not ideal, especially in high-speed applications, it i...

AI Technical Summary

Problems solved by technology

After the sampling data is processed by FFT transformation, many times due to large bandwidth problems, signal leakage problems, signal aliasing problems, in-band self-excitation problems, etc., the effect after FFT is not ideal, some unwanted signals or irrelevant signals It will appear in the useful signal band, and the front-end hardware and programmable logic processing parts are difficult to solve under high-speed conditions, which affects the data processing effect

[0003] Usually, when the hardware environment is not ideal, especially in high-speed applications, it is generally difficult to completely improve the signal effect through hardware optimization.

At the same time, the improvement of the hardware environment needs to increase labor costs, time expansion, and hardware resource consumption to a certain extent, and in many cases, due to different environments, the results of one debugging do not meet the general requirements of multiple environments.

Images