Use Segmentation Image to create a segmentation image from a classified image. Segmentation partitions a classified image into regions of connected pixels or blobs of pixels that are contained in the same class. Each connected region is given a unique DN value. You can specify the minimum number of pixels that must be contained in a region and you can select whether eight or four neighbors are considered for the connectivity.

  1. From the Toolbox, select Feature Extraction > Segmentation Image. The Segmentation Image Input File dialog appears.
  2. Select the input classification file and any optional spatial subsetting, then click OK. The Segmentation Image Parameters dialog appears.
  3. In the Min Thresh Value field, select the minimum pixel value. In the Max Thresh Value field, select the maximum pixel value.
  4. In the Population Minimum field, enter the minimum number of pixels that must be contained in a region for it to become a segment.
  5. Use the toggle button to designate whether to consider eight or four neighboring pixels for the connectivity.
  6. Select output to File or Memory.
  7. Click OK. ENVI adds the resulting output to the Layer Manager, with each segment having a unique DN value.