Several algorithms can be used to filter DEMs. Most simple and common are:
- Low-pass filter. Also known as moving average filter. Implemented by many (all) tools, for example in SAGA http://www.saga-gis.org/saga_tool_doc/2.3.0/grid_filter_0.html
- Gaussian filter. Implemented by OTB see https://www.orfeo-toolbox.org/CookBook/Applications/app_Smoothing.html or by SAGA see http://www.saga-gis.org/saga_tool_doc/2.3.0/grid_filter_1.html
A less common but very recognized and efficient:
- Mdenoise. It is a despeckle algorithm that can denoise DEMs or meshes. Source code download, install and use see https://personalpages.manchester.ac.uk/staff/neil.mitchell/mdenoise/. Implemented also by SAGA see http://www.saga-gis.org/saga_tool_doc/2.3.0/grid_filter_10.html.
Examples of such filters on an noisy DEM :
mdenoise_n5_t099 
mdenoise_n10t09 
mdenoise_n15_t05 
gaussian_circle_r3 
gaussian_circle_r5 
lowpass_circle_r5 
lowpass_circle_r7
Click on caption to enlarge. r : radius