 Hello, this is Hans von Neckar, Senior Lecturer at IC Delft Institute for Water Education. In this video we are going to look at the window operators of PCRuster in QGIS, and these are also called focal operations. They work in a moving window of a certain length, so we always have to specify an input dataset and the length of the window that we want to use. Let's first apply the window average tool to our DTM. So I choose DTM as an input raster and it will then calculate the average of self values in a specified square neighborhood. And let's make that 150 meters. Here I save it. Let's call it smooth DTM because it will smooth the values by taking the average in the window. Then I run it and here we see the result and it indeed looks smoother. Note that the spatial resolution doesn't change. Next tool that I'm going to demonstrate is the window diversity tool and I'll apply that to our BuildG layer with classes because it needs a discrete input, a Boolean nominal ordinal. So I choose BuildG and I choose a window length of 300. And it will calculate the amount of classes, the diversity of classes that it finds in that window. When I run it, move it to the top, I get this result and I can style it with palleted unique values and here we see the amount of classes that it finds in that moving window and then assign to the center pixel. It is very useful if you're interested in environments with different habitats and how many habitat types you find in a certain area that you define. Let's apply the window high pass to the DTM values. This is a high pass filter. So to increase the frequency within a specified square neighborhood. Also going to make that 300 and call it DTM high pass and we see the result and let's look at the histogram of this result. You can apply any other QGIS tool to this data because it's GDAL supported format. So click right and go to properties and there it histogram, compute histogram and I see the frequency histogram for this layer and the effect of a high pass filter. Next I'm going to demonstrate the window majority operation and I apply that to our classes file build G and I'm going to look at an area of 300 meters again and it will give the most occurring class in the 300 meter window that I specified and let's style it to see the result in a better way, palleted unique values because it assigns the most occurring class number of build G within the 300 meter neighborhood and we see that class 0 is most occurring and then the red class and the blue class. There's also the window maximum operation that calculates the maximum value in the window and the window minimum operation which is the minimum value and window total calculates the total the sum of values that it finds in the window and window operations always assign the result to the center pixel and you lose always the boundaries because there's no information on the boundaries so those will result in no data.