You can fill bad values in digital elevation models (DEMs) with values calculated using a surface fitting technique. Bad pixels often occur in radar shadows of DEMs generated from interferometric radar. The locations of the bad pixels are identified in the DEM image by entering the DN value of the bad pixels or a range of bad values. You can also make a mask of the bad pixels and use it to locate the pixels that need to be replaced. The surface fitting uses Delaunay triangulation to fill the bad pixels with triangles calculated from the surrounding good elevation values.

If there are large holes in your dataset, you may see the triangles used to fill them.

  1. From the Toolbox, select Terrain > Replace Bad Values. The DEM Input File dialog appears.
  2. Select a DEM file and perform optional spatial subsetting and/or masking, then click OK. The Replace Bad Data Parameters dialog appears.
  3. Enter a bad pixel value in the Bad Value field, or a range of bad pixel values in the Min and Max fields.

    4. You do not need to enter values if you selected a mask band that contains the locations of the bad pixels.

  4. Select output to File or Memory.
  5. Click OK. ENVI adds the resulting output to the Layer Manager.