edu.umn.cs.spatialHadoop.nasa

Class RecoverHoles

java.lang.Object

edu.umn.cs.spatialHadoop.nasa.RecoverHoles

public class RecoverHoles
extends java.lang.Object

Recovers missing data from a set of images by substituting missing points from neighboring images based on time.

Author:

Ahmed Eldawy

Constructor Summary

Constructors

Constructor and Description
RecoverHoles()

Method Summary

Methods

Modifier and Type	Method and Description
static void	addDate(Path dir)
static void	main(java.lang.String[] args)
static void	recoverInterpolationDir(Path dir) Recover all images in the given directory by performing an interpolation on each missing point from the two nearest points on each side on its horizontal line (the nearest left and right points).
static void	recoverNearest(Path dir) Recover all images in the given dir.

Methods inherited from class java.lang.Object

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

Constructor Detail

RecoverHoles

public RecoverHoles()

Method Detail

recoverNearest

public static void recoverNearest(Path dir)
                           throws java.io.IOException

Recover all images in the given dir. The passed directory contains a set of images each corresponding to a specific date. The algorithm works in iterations. In the first iteration, each image is used to recover holes in the image that directly follows it. For example, the valid points in the image of day 1 is used to recover missing points in the image of day 2. In the second iteration, each image is used to recover missing points in the preceding day. All odd numbered iterations work like the first iteration while all even numbered iterations work like the second one. The algorithm works like this until either all images are recovered or we run n iterations where n is the total number of all images (time points).

Parameters:

dir -

Throws:

java.io.IOException

recoverInterpolationDir

public static void recoverInterpolationDir(Path dir)
                                    throws java.io.IOException

Recover all images in the given directory by performing an interpolation on each missing point from the two nearest points on each side on its horizontal line (the nearest left and right points). To determine which points should be interpolated (e.g., under cloud) and which points should remain blank (e.g., in sea), we first overlay all images on top of each other. If a point is missing from all images, it indicates with very high probability that it should remain blank.

Parameters:

dir -

Throws:

java.io.IOException

addDate

public static void addDate(Path dir)
                    throws java.io.IOException

Throws:

java.io.IOException

main

public static void main(java.lang.String[] args)
                 throws java.io.IOException

Parameters:

args -

Throws:

java.io.IOException