A median filter circuit and method

Abstract

The invention provides a median filtering circuit and a method thereof. The median filtering circuit comprises a control processing unit, a timestamp processing unit, a first queuing processing unit, a second queuing processing unit and a median data output selector, wherein the control processing unit generates control signals of the entire median filtering circuit, and finishes control of the entire median filtering circuit. The timestamp processing unit receives input data and outputs data which is stamped with effective timestamps under the control signals output by the control processing unit. The first queuing processing unit and the second queuing processing unit carry out median filtering arithmetic process on the data which is stamped with effective timestamps under the control signals output by the control processing unit and output median data. The median data output selector selects the median data of the first queuing processing unit or the second queuing processing unit as effective median data to output under the control signals output by the control processing unit. Through utilization of the timestamp processing unit, output time is ensured, one-dimensional quick median filtering is realized through two queuing processing units, the algorithm and control are relatively simple so as to adapt to the median filtering in a large window.

Description

technical field [0001] The invention relates to the field of median filtering, in particular to a median filtering circuit and method applied to weak signal measurement. Background technique [0002] When measuring weak signals, due to various interferences, the measurement results will produce singular data, and the median filter can solve this problem well. Median filtering is a typical non-linear filtering technique. It continuously samples a measured parameter n times (generally, n is an odd number), and then arranges the n sampling values ​​according to their size, and takes the median value of this sampling value (if n is an even number, the average of the two sorted intermediate values ​​is the current sampling value). The median filter can effectively overcome pulse interference such as fluctuations caused by accidental factors or bit errors caused by unstable samplers. [0003] The median filter is now mostly used in the field of two-dimensional image processing, ...

AI Technical Summary

Problems solved by technology

In the prior art, the two-dimensional median filtering method uses the fast queuing algorithm to replace the traditional sorting serial algorithm based on the bubble sorting method, but there are still the following disadvantages: on the one hand, the window length is too short to adapt to large windows. Median filtering; on the other hand, for a median filtering window of length n, n+1 serially connected processing units are required, the algorithm and control are complex, the calculation is large, time-consuming, and the output time is uncertain

Images