DIEST

Description

DIEST combines multiple layers of spatial data to derive a favorability model for predicting a geological event (mineral potential, landslide, etc.). Available algorithms include probability estimation, certainty factor estimation, and fuzzy membership estimation. DIGRP must be run before DIEST.

Parameters

diest(file, unicon, datfil, algo, dboc)

Name	Type	Caption	Length	Value range
FILE *	str	Input file name	1 -
UNICON *	List[int]	Uni-condition region IDs channel	1 - 1
DATFIL *	str	ASCII data file name	1 -
ALGO *	str	Estimation algorithm	1 - 16	See Description Default: CF
DBOC *	List[int]	Output estimated favorability map channel	1 - 1

* Required parameter

Parameter descriptions

FILE

Specifies the name of the PCIDSK file containing the uni-condition regions channel.

UNICON

Specifies the channel containing the uni-condition region ID(s), produced by DIGRP.

DATFIL

Specifies the name of the ASCII data file that contains the data layer statistics, as produced by DIGRP.

ALGO

Specifies the estimation algorithm to be used.

DBOC

Specifies the output channel to receive the estimated favorability map (which is scaled to [0, 255]).

Details

DIEST provides various algorithms for estimating the favorability of occurrence, by combining evidence from multiple input layers. All the statistics of input layers are gathered and saved in an ASCII data file by DIGRP in a preprocessing step. DIEST does not need to read data from the original input channels.

The favorability measure is estimated for each unique condition region rather than for each pixel. Because the number of unique condition regions is much smaller than the number of pixels in the image, this process significantly reduces computation.

The computed favorabilities for all the uni-condition regions are saved in a file specified by REPORT.

In addition, a raster image of favorabilities is generated where each pixel contains the favorability measure mapped to the range of 0-255 (the lowest favorability is mapped to 0 and the highest is mapped to 255). The mapping function is similar to histogram equalization; you may expect approximately the same number of pixels at each gray level. For example, to obtain the top 10% most favorable areas, threshold the output image at 255*90%. The pixels above this threshold should fall in the top 10% category.

Example

The following example assumes that the first 7 channels of land.pix are the data layers to be integrated. Segment 2 contains a bitmap of the known occurrence areas. Segment 3 contains a bitmap of the surveyed areas. First, by running DIGRP, the data layers are grouped into a uni-condition image and saved in channel 17. Some statistics of uni-condition regions are saved in the ASCII file land.dat. Then, by running DIEST, a favorability map is generated using the Certainty Factor algorithm (CF) and saved in channel 10.

from pci.digrp import digrp
from pci.diest import diest

file='land.pix'
dbic=[1,2,3,4,5,6,7]
kosn=[2]
sasn=[3]
unicon=[17]
datfil='land.dat'

digrp(file, dbic, kosn, sasn, unicon, datfil)

algo='CF'
dboc=[10]

diest(file, unicon, datfil, algo, dboc)

Algorithm

For the Fuzzy Logic Gamma operator, the Gamma value must be 0 < Gamma < 1. Otherwise, a default Gamma value of 0.95 is used.