Use this procedure to perform neural net classification, which is a supervised classification where the net is trained on a set of input ROIs. The keywords ALPHA, ETA, THETA, NUM_SWEEPS, NUM_LAYERS, and RMS_CRIT are associated with training the net. You can use the optional keyword THRESH to set the minimum activation threshold a class must satisfy in order to be classified.
Syntax
ENVI_DOIT, 'ENVI_NEURAL_NET_DOIT', ACT_TYPE={0 | 1}, ALPHA=value, CLASS_NAMES=string array, DIMS=array, ETA=value, FID=file ID, /IN_MEMORY, LOOKUP=array, NUM_CLASSES=integer, NUM_LAYERS=integer, NUM_SWEEPS=integer, OUT_BNAME=string array, OUT_NAME=string, POS=array [, R_FID=variable], RMS_CRIT=value, ROI_ID=array [, RULE_FID=variable] [, /RULE_IN_MEMORY] [, RULE_OUT_BNAME=string array] [, RULE_OUT_NAME=string], THETA=value [, THRESH=value], /TRAIN
Keywords
ACT_TYPE
Set this keyword to one of the following values to specify the activation function for training the neural net.
- 0: Logistic
- 1: Hyperbolic
ALPHA
Use this keyword to specify the training momentum. ALPHA is a floating-point or double-precision number between 0.0 and 1.0.
CLASS_NAMES
Use this keyword to specify a string array of class names for classification images. The first element (Class 0) is “Unclassified.” The array has num_classes+1 elements.
DIMS
The “dimensions” keyword is a five-element array of long integers that defines the spatial subset (of a file or array) to use for processing. Nearly every time you specify the keyword FID, you must also specify the spatial subset of the corresponding file (even if the entire file, with no spatial subsetting, is to be processed).
- DIMS[0]: A pointer to an open ROI; use only in cases where ROIs define the spatial subset. Otherwise, set to -1L.
- DIMS[1]: The starting sample number. The first x pixel is 0.
- DIMS[2]: The ending sample number
- DIMS[3]: The starting line number. The first y pixel is 0.
- DIMS[4]: The ending line number
To process an entire file (with no spatial subsetting), define DIMS as shown in the following code example. This example assumes you have already opened a file using ENVI_SELECT or ENVI_PICKFILE:
envi_file_query, fid, dims=dims
ETA
Use this keyword to specify the training rate. ETA is a floating-point or double-precision number between 0.0 and 1.0.
FID
The file ID (FID) is a long-integer scalar with a value greater than 0. An invalid FID has a value of -1. The FID is provided as a named variable by any routine used to open or select a file. Often, the FID is returned from the keyword R_FID in the ENVIRasterToFID routine. Files are processed by referring to their FIDs. If you work directly with the file in IDL, the FID is not equivalent to a logical unit number (LUN).
IN_MEMORY
Set this keyword to specify that output should be stored in memory. If you do not set IN_MEMORY, output will be stored on disk and you must specify OUT_NAME (see below).
LOOKUP
Use this keyword to specify an byte array of size [3, num_classes+1], containing color tables for the classification image. Each output class can have a unique RGB triplet, ordered as [r, g , b], and the “Unclassified” class typically uses the RGB triplet [0, 0, 0] for black.
NUM_CLASSES
Use this keyword to specify the number of output classes, which is equal to the number of input ROIs as specified by ROI_PTR.
NUM_LAYERS
Use this keyword to specify the number of hidden layers in the neural net classifier. Typically this value is set to 0, 1, or 2.
NUM_SWEEPS
Use this keyword to specify the maximum number of training sweeps to perform count. The actual number of training sweeps may be less than NUM_SWEEPS if error criteria have been met. See the keyword RMS_CRIT below.
OUT_BNAME
Use this keyword to specify a string array of output band names.
OUT_NAME
Use this keyword to specify a string with the output filename for the resulting data. If you set the keyword IN_MEMORY, you do not need to specify OUT_NAME.
POS
Use this keyword to specify an array of band positions, indicating the band numbers on which to perform the operation. This keyword indicates the spectral subset of bands to use in processing. POS is an array of long integers, ranging from 0 to the number of bands minus 1. Specify bands starting with zero (Band 1=0, Band 2=1, etc.) For example, to process only Bands 3 and 4 of a multi-band file, POS=[2, 3].
POS is typically used with individual files. The example code below illustrates the use of POS for a single file with four bands of data:
pos=[0,1,2,3]
envi_doit, 'envi_stats_doit', dims=dims, fid=fid, pos=pos, $
comp_flag=3, dmin=dmin, dmax=dmax, mean=mean, stdv=stdv, hist=hist
But what if you need to create an output file consisting of data from different bands, each from different files? Library routines such as CF_DOIT and ENVI_LAYER_STACKING_DOIT can accomplish this, but they use the POS keyword differently. Suppose you have four files, test1, test2, test3, and test4, with corresponding FIDs of fid1, fid2, fid3, and fid4, respectively. In the following example, you want Band 3 from test1 in the first position, Band 2 from test2 in the second position, Band 6 from test3 in the third position, and Band 4 from test4 in the fourth position. The code should be as follows:
fid_array = [fid1,fid2,fid3,fid4]
pos=[2,1,5,3]
envi_doit, 'cf_doit', dims=dims, fid=fid_array
out_name='test_composite_file'
R_FID (optional)
ENVI Classic library routines that result in new images also have an R_FID, or “returned FID.” This is simply a named variable containing the file ID to access the processed data. Specifying this keyword saves you the step of opening the new file from disk.
RMS_CRIT
Use this keyword to specify the RMS training error criteria. During training, if the RMS error is less than RMS_CRIT, then the training ends and the image is classified according to the trained neural net.
ROI_ID
Use this keyword to specify an array of ROI IDs returned from a call to ENVI_GET_ROI_IDS. Each ID in the array will use the corresponding ROI to train the neural net.
RULE_FID (optional)
Use this keyword to specify a named variable that contains the file ID for the processed rule image. This file ID can be used to access the processed data.
RULE_IN_MEMORY (optional)
Set this keyword to store output rule images in memory.
RULE_OUT_BNAME (optional)
Use this keyword to specify a string array that contains the output band names for the rule image.
RULE_OUT_NAME (optional)
Use this keyword to specify an output filename for the rule image. If this item is present, the rule image is automatically saved.
THETA
Use this keyword to specify the training threshold contribution. THETA is a floating-point or double-precision number between 0.0 and 1.0.
THRESH (optional)
Use this keyword to specify an minimum activation threshold a class must have in order to be classified.
TRAIN
Set this keyword to train the neural net prior to classification.
API Version
4.2