 Hello, welcome back on my YouTube channel. In today's video we are going to look at filtering point cloud data. Let's check which attributes are available for this point cloud. We can find that in the layer properties. In the layer properties dialog, click on statistics. There you'll find a table with the attribute statistics. It gives us the list of attributes, the minimum, maximum, mean and standard deviation. And for the classification it gives us the statistics on the count, the number of points and the percentage of points in that class. Now we'll use these attributes for filtering points out of the data set. Let's first use the attributes for styling the layer. Go to the layer styling panel and there change RGB to classification. And this will show the points with their classes. And here the table also shows the percentage again. Now I'm going to create a layer with only the building points from the point cloud. In the processing toolbox under point cloud extraction I use filter. I use our last file as the input layer. And there I go to filter and I choose classification equals. And then from values I choose building by double clicking the value. I click test and there it says that it's valid. And then I save the result and I call it buildings. After calculating the result it will convert it to the COP C format, the cloud optimized point cloud and then it will show the result. If I hide the classification layer here we can see the resulting building points. I can also filter on height. Let's go back to filter and there I'm going to look for Z larger than 25 meters. So this is a valid expression, click OK and I call this one high. This expression applies to all the points, not only the buildings and will return the ones that are higher than 25 meters. And here we see all the points that are higher than 25 meters. Maybe we are not interested in buildings but in urban green. Let's see if we can use a filter to get the vegetation out of the point cloud. I'm going to use the return number. The return number refers to the number of returns from a single emitted pulse of the LiDAR system. If the return number is larger than one this could be vegetation. Save the file as vegetation and run the filter. After calculation close the dialog and check the result. When I zoom in I see that many trees have been detected but also a lot of side catch mostly the sides of buildings. Let's see if we can refine our filter to reduce the amount of building points that are included in this result. I'm going to use again return number larger than one and classification is not buildings. I test the expression and it works. Click OK and I save the result. I call it vegetation too. Let's run it. Click close and check the result. We see that now less building points have been included so we can still further refine this for example by using elevation criteria or other types of filtering. It's also important to note that many of these point cloud processing algorithms have filter options which you can find under the advanced parameters. That means when you execute such a tool you can also include a filter so the tool is only applied on the filter criteria. In this video you've learned how to apply filters to point clouds.