What is a Digital Surface Model (DSM)?

A Digital Surface Model represents the world as seen from above, including rooftops, tree canopies, vehicles, and every other object on the landscape. While a DTM strips away everything to show bare earth, a DSM captures the actual surface you'd encounter if you descended from the sky — the top of whatever exists at each location.

How DSMs Are Created

Creating a DSM is simpler than creating a DTM because no classification is required. The algorithm takes the first return (or highest point) from each LiDAR pulse and interpolates these points into a continuous raster grid. Every cell contains the elevation of whatever surface the laser hit first — whether that's a building roof, a tree crown, or the ground in open areas.

The simplicity of DSM generation means fewer parameters to tune and fewer opportunities for classification errors. The trade-off is that a DSM alone tells you less than DTM and DSM together — you see the surface, but you don't know what's beneath it.

DSM vs DTM: The Power of Both

The real value emerges when you have both models. Subtract the DTM from the DSM, and you get a normalized surface model (nDSM) showing the height of everything above ground. A building that sits on a hillside might have a DSM elevation of 150 meters and a DTM elevation of 140 meters — the 10-meter difference is the actual building height. This simple calculation works for every tree, structure, and object in your dataset.

DSMs with LidarVisor

LidarVisor generates both DSM and DTM from your point cloud. Compare them to extract building heights, tree heights, and any other vertical features — or export each surface separately for use in your GIS workflows.