Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Using median filters for temperature noise removal

Not open for further replies.


Advanced Member level 4
Jul 16, 2001
Reaction score
Trophy points
Activity points
signal noise filtering

Hi all
For long time I've used FIR or IIR filters (depend on application) for noise filtering. Last time I've found very interesting article ( and there is median filters proposed for temperature noise removal. I am thinking to use median filter in next application. Does anybody have experience to could compare with standard Low Pass filters?


error in link

I am getting a 404 error for your link. Try posting it again. Here is a link **broken link removed** that works and has many examples of how it affects the image.

In general, it is a way of filtering images by using the median instead of the average level of the surrounding pixel levels. This preserves the detail of the image. The averaging method low pass filters the image which removes detail.

Hi flatulent
I am sorry for link, I'he typed it manualy. Here is proper link **broken link removed** Thanks for your information, I have know this application before, but I have never thought to use median filter in temperature or another parameter measurements.
I could imagine how could mediane filter work in 2D but linear application 8O


good idea

I read your link and now understand the question. This looks like a very good way of removing non-gaussian noise, like shot or impulse noise.

Although not described in the article there are several other ways to improve things. Oversampling will spread out the ADC quantization noise so that low pass filtering in DSP will be able to improve the SNR and then decimate the samples to get the data rate lower.

Oversampling will also make the impulse noise occur on a smaller fraction of the samples. You could also look for these impulses and replace them with the median of the adjacent samples.

This last idea was used for decades in an analog form in short wave receivers for removing impulse noise effects from the audio output. They used slew rate limiting and clipping compared to the previous levels for this.

Not open for further replies.

Part and Inventory Search

Welcome to