 In this video, we'll take a look at the image analysis functions and how you can use them to apply spatial filters to your imagery in ArcGIS. We'll begin in ArcMap, where I've got an 8-band WorldView 2 image displayed as a color infrared composite. You can use geo-processing tools to apply spatial filters, but the spatial filters available to you in the image analysis tools are both faster and easier to use. The first step is to select the raster layer we want to apply the spatial filters to, and then we'll scroll down and under Processing, choose the appropriate filter. Let's start with a Shop and Filter. We'll select Shop and from the drop-down menu, and then click on the Filter button to apply the filter and generate a new temporary raster layer. Most filters produce a new layer with the exact same number of bands as the original. Most image analysis functions will produce a new raster layer in a matter of seconds, but that output is temporary. If we want to create a permanent raster, under the Processing menu in the image analysis tools, we'll click on the Export Raster Data button and then choose the appropriate location and file name for our raster. One of the most useful ways to evaluate our output is to use the Swipe tool in the Effects toolbar. Using the Swipe tool, we can swipe back and forth between our raster layers to examine the impact of the spatial filter. The Shop and Filter created a new data set with crisper features in the original. The edges of the buildings are more easily discernible, and we see increased texture in the tree canopy. Shop and Filter emphasize the spatial properties of the imagery over the spectral ones. Now let's remove our Shop and Layer and apply a Blur filter. As we would expect, the Blur filter has the opposite effect of the Shop and Filter. It de-emphasizes the spatial properties of the imagery in favor of the spectral ones. The building edges appear much less crisp, and the tree canopy is a more uniform texture. When performing image classification, Shop and Filters can be useful for using object-based techniques, and we need the texture to differentiate between, say, different vegetation types. Blur Filters are useful for using traditional pixel-based approaches, where the spectral uniform signature of a feature is very important for its identification. Many filters were output-similar results. For example, a 5x5 smoothing filter on the Worldview 2 data set appears very similar to the output from the Blur filter. Determining the filters you want to apply to your image depends on how you plan to use them for additional image analysis tasks. It's as much trial-and-error as understanding image filter theory and operations. Up until now we've been working with High Resolution Worldview 2 imagery. Let's switch gears a bit and work with a moderate resolution Landsat image. We're going to apply a Laplacian 3x3 filter. A Laplacian filter is essentially an edge detection filter. It'll emphasize those areas where there's high contrast between pixels. Let's drag the filtered output below our Landsat image and then use the Swipe tool to evaluate our output. The Laplacian filter does an excellent job of helping us understand the landscape heterogeneity in this Landsat scene. There exists a clear gradient from west to east where near the lake we have a complex landscape of agriculture and urbanized features. As we move towards the mountains covered by clouds, we see that the Laplacian filter shows a more uniform approach indicating undisturbed forest patches. We can use spatial filters for data exploration, manual image interpretation, or automated feature extraction.