Signature Files

Producer Field Guide

HGD_Product
Producer Field Guide
HGD_Portfolio_Suite
Producer

A signature is a set of data that defines a training sample, feature space object (AOI), or cluster. The signature is used in a classification process. Each classification decision rule (algorithm) requires some signature attributes as input—these are stored in the signature file (.sig). Signatures in ERDAS IMAGINE can be parametric or nonparametric.

The following attributes are standard for all signatures (parametric and nonparametric):

  • name—identifies the signature and is used as the class name in the output thematic raster layer. The default signature name is Class <number>.
  • color—the color for the signature and the color for the class in the output thematic raster layer. This color is also used with other signature visualization functions, such as alarms, masking, ellipses, and so forth.
  • value—the output class value for the signature. The output class value does not necessarily need to be the class number of the signature. This value should be a positive integer.
  • order—the order to process the signatures for order-dependent processes, such as signature alarms and parallelepiped classifications.
  • parallelepiped limits—the limits used in the parallelepiped classification.

Parametric Signature

A parametric signature is based on statistical parameters (for example, mean and covariance matrix) of the pixels that are in the training sample or cluster. A parametric signature includes the following attributes in addition to the standard attributes for signatures:

  • number of bands in the input image (as processed by the training program)
  • minimum and maximum data file value in each band for each sample or cluster (minimum vector and maximum vector)
  • mean data file value in each band for each sample or cluster (mean vector)
  • covariance matrix for each sample or cluster
  • number of pixels in the sample or cluster

Nonparametric Signature

A nonparametric signature is based on an AOI that you define in the feature space image for the image file being classified. A nonparametric classifier uses a set of nonparametric signatures to assign pixels to a class based on their location, either inside or outside the area in the feature space image.

The format of the .sig file is described in Help. Information on these statistics can be found in Math Topics.