REMOTE SENSING MIDTERM UNIT 4

Single Image Enhancement Purpose

To improve the visual interpretability of a band by increasing the apparent distinction between the features of the scene

Constant stretching

improve visual distinction between features in an image by increasing the range of pixel values which enhances the difference between light and dark areas making details more visible (stretch to bit range such as 0 - 8).
Do if the DN values in a band don’t vary over the entire possible range

Linear Stretching vs Histogram Stretching

Linear: Assign the minimum and maximum value of your bit number to the DN ranges min / max and spread them equally which maintains the original shape of the histogram

Histogram: Redistribute the values based on the images histogram to balance brightness levels (alters histogram shape by redistributing intensities for a more balanced contrast)

Why you’d do histogram stretching > linear stretching

Helps if areas in shadow or brightness lose detail - repress the concentration of pixels in the raw data

Band-Ratioing

The process behind Indices. Pixel values of one spectral band are divided by the pixel values of a second band (Band A / Band B)

Pros of Band Ratioing (and picture explanation)

Enhances specific features while reducing the effects of illumination differences (shadows and topography)

Helps differentiate land cover types are difficult to identify in one band

NDVI formula and when to use

(NIR - R) / (NIR + R)

Use when calculating vegetation changes over a temporal or spatial extent

normalizes for atmosphere

most general vegetation application

DVI

NIR - R

absolute difference in veg

not normalized

SRI

Simple ratio index

high biomass

not normalized

NIR/R

SAVI

Soil Adjusted Vegetation Index

((NIR - R) / (NIR + R + L)) (L + 1)

Minimize soil noise

L is a correction factor between 0 (high veg cover) and 1 (low veg cover)

Why are indices used in remote sensing (4)

1. they are simple

2. they are computed directions w/o assumptions regarding land cover class, soil types, or climate conditions (no prior knowledge , widely applicable)

3. Provide percise and continuous measures of spectral and temporal variability

4. Empirically correlated w many environmental variables

Principle Component Analysis

Remove inter-band correlation (two bands that convey similar information) in multispectral (and hyperspectral) data.

Creates a subset of “components” that can substitute bands.

Data is mapped along the two axes which explain the most variance, draw line of best fit, make that new axis, and repeat until most of the variability is explained by the components