Scribd
Upload
52 views21 pages

12.6 Advanced Analysis and Classification

This document provides information on various remote sensing analysis techniques including: 1. Spatial filters which are used to enhance imagery by uncovering patterns or restore imagery by reducing gaps and artifacts. 2. Color space transformations and pan sharpening which are used to produce clear, reproducible color representations of imagery. 3. Spectral indices which are algorithms used to analyze imagery and extract environmental information using specific wavelength combinations. 4. Image classification techniques including unsupervised classification which uses software to group similar pixel values into classes, and supervised classification where user-defined training samples are used to classify pixels into classes.

Uploaded by

Jose Luis Flores
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
52 views21 pages

12.6 Advanced Analysis and Classification

This document provides information on various remote sensing analysis techniques including: 1. Spatial filters which are used to enhance imagery by uncovering patterns or restore imagery by reducing gaps and artifacts. 2. Color space transformations and pan sharpening which are used to produce clear, reproducible color representations of imagery. 3. Spectral indices which are algorithms used to analyze imagery and extract environmental information using specific wavelength combinations. 4. Image classification techniques including unsupervised classification which uses software to group similar pixel values into classes, and supervised classification where user-defined training samples are used to classify pixels into classes.

Uploaded by

Jose Luis Flores
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

IMAGINERY ANALYSIS

Advanced analysis and


classification
Spatial Filters

Spatial filters generally serve two purposes when applied to remotely


sensed data: 1) enhance imagery or 2) restore imagery. When it comes
to enhancing imagery, spatial filters can help uncover patterns and
processes. Spatial filters are useful for both manual image
interpretation and
automated feature
extraction. Spatial
filters can also help
to restore imagery
that has either gaps
or artifacts.
Color space transformation
• A color space is a specific organization of colors.
• In combination with physical device profiling, it allows for
reproducible representations of color, in both analog and digital
representations.
• A color space may be
arbitrary, with
particular colors
assigned to a set of
physical color swatches
and corresponding
assigned color names
or numbers.
Pan sharpening
• A fusion process of panchromatic images usually from the same sensor
finding a clear image.
Spectral Indices

Over the years numerous spectral indices have been introduced by the
scientific community to solve complex environmental (or other) issues.
Advanced Vegetation Index (AVI)
Bare Soil Index (BSI)
Shadow Index (SI)
Normalized Difference Vegetation Index
(NDVI)
Normalized Difference Water Index (NDWI)
Normalized Difference Snow Index (NDSI)
Normalized Difference Glacier Index (NDGI)
Normalized Difference Moisture Index
(NDMI)
Normalized Burned Ratio Index (NBRI)
Normalized Pigment Chlorophyll Ratio Index
(NPCRI)
Image classification
• Image classification refers to the task of extracting information classes
from a multiband raster image. The resulting raster from image
classification can be used to create thematic maps.
Unsupervised image classification
Unsupervised classification is where the outcomes (groupings of pixels
with common characteristics) are based on the software analysis of an
image without the
user providing sample classes. The computer uses techniques to
determine which
pixels are related and
groups them into
Classes
Supervised image
classification
In supervised classification, you select
training samples and classify your
image based on your chosen samples.
Your training samples are key because
they will determine which class each
pixel inherits in your overall image.
Unsupervised image classification
Unsupervised classification is where the outcomes (groupings of pixels
with common characteristics) are based on the software analysis of an
image without the
user providing sample classes. The computer uses techniques to
determine which
pixels are related and
groups them into
Classes
Unsupervised image classification
Start the tool
Load the image
Preprocessing of downloaded files
Create a band dataset
Band processing - clustering
Run classification
• SCP 6 has two algorithms for performing unsupervised classification,
K-Means and Isodata. To execute the classification, choose the
method and specify the number of classes, the rest of the parameters
will be left with their default values.
Run classification

K-Means classification 7 clases

Isodata classification 11 classes


Supervised image classification
Start the tool
Creating regions of interest
• The supervised classification requires that regions of interest be
entered for each land cover class, these are polygons that are plotted
on homogeneous areas that overlap pixels belonging to the same
land cover class.
Once the ROIs are created using training areas, the polygons and spectral signatures used for the
classification of the ground cover of the image are saved. ROIs can be created in two ways:

Using incremental region algorithm (automatically)


Creating Polygons (manually).
Creating regions of interest
Once the training areas have been entered,
we choose the algorithm to use that gives us
the best result in the classification, we can
also define a threshold, which helps to
improve the classification results, especially
when the firms have some similarity.
Identify overlapping signatures
Some signatures are orange and
others are not, this is because in
these, there is overlap with other
signatures, in the Color column
(Overlap MC_ID-C_ID) you can see
which overlap exists, in this case it
is important to see the standard
deviation of each of the firms,
since the objective is to have a
smaller standard deviation for each
firm, if it is high, it indicates that
the ROI is not homogeneous and
we have to modify it.
Reviewing spectral distances
Another thing that we can
review is the spectral
distances between the
signatures, with which we
can evaluate the
separation between the
ROIs, in case there is
similarity it is better to
delete some or change the
threshold.
Run
Before proceeding with the final result of the classification we can
preview, edit the colors that most closely resemble the ground cover
that will be shown in the final classification raster.

You might also like