org.gridfour.lsop

Class LsOptimalPredictor12

java.lang.Object

org.gridfour.lsop.LsOptimalPredictor12

public class LsOptimalPredictor12
extends Object

Applies the methods of Lewis and Smith (1994) using optimal predictors for data compression.

The method used for this class is based on the PDF document: Lewis, M. and Smith, D. H. (1994). "Optimal Predictors for the Data Compression of Digital Elevation Models using the Method of Lagrange Multipliers"

The floating-point arithmetic operations for the encode method in this class are performed using the Java strictfp specification. This design choice is essential to the correct operation of this module. Because one of the goals for Gridfour is to facilitate portability to other development environments, it is essential that the math operations performed here be reliably reproducible across platforms and programming languages.

Constructor Summary

Constructors

Constructor and Description
LsOptimalPredictor12()

Method Summary

Modifier and Type	Method and Description
double[]	computeCoefficients(int nRows, int nColumns, int[] values) Computes the coefficients for an optimal predictor.
LsOptimalPredictorResult	encode(int nRows, int nColumns, int[] values) Encode the specified raster using the Smith-Lewis Optimal Predictor 12-coefficient variation
double	getMeanAbsError()
double	getMeanError() Gets the mean error for the predictor.
double	getPercentZeroIntegerResiduals() Gets the percent of the integer coded residuals that have a zero value.
double	getRootMeanSquaredError() Gets the root mean squared error (RMSE) for the predictor.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

LsOptimalPredictor12

public LsOptimalPredictor12()

Method Detail

encode

public LsOptimalPredictorResult encode(int nRows,
                                       int nColumns,
                                       int[] values)

Encode the specified raster using the Smith-Lewis Optimal Predictor 12-coefficient variation

Parameters:

nRows - the number of rows in the raster

nColumns - the number of columns in the raster

values - the values for the raster given as an array of integers in row-major order.

Returns:

if successful, a valid results instance; otherwise, a null.

computeCoefficients

public double[] computeCoefficients(int nRows,
                                    int nColumns,
                                    int[] values)

Computes the coefficients for an optimal predictor. In the unusual case that no solution is available, or that the input data size is inadequate, a null value will be returned. The layout of the coefficients is as shown below

    row i:      u3   u0   S(i,j)
    row i-1:    u4   u1   u2
    row i-2:    u5   u6   u7
 

Parameters:

nRows - a value of 6 or greater

nColumns - a value of 6 or greater

values - an array of values given in row-major order.

Returns:

if successful, an valid array; otherwise, null.

getMeanError

public double getMeanError()

Gets the mean error for the predictor. The error is defined as the the value minus the predictor computed using single-precision floating-point values (4-byte floats)).

Returns:

a valid floating point value

getRootMeanSquaredError

public double getRootMeanSquaredError()

Gets the root mean squared error (RMSE) for the predictor.

Returns:

a valid floating point value

getPercentZeroIntegerResiduals

public double getPercentZeroIntegerResiduals()

Gets the percent of the integer coded residuals that have a zero value.

Returns:

a value in the range zero to 100.

getMeanAbsError

public double getMeanAbsError()