Transcript Lecture: Image Processing and Interpretation

Lecture 7: Image Processing
and Interpretation
Last time we discussed:
pixel
The cells are sensed one after another along the line.
In the sensor, each cell is associated with a pixel that is
tied to a microelectronic detector
Pixel is a short abbreviation for Picture Element
a pixel being a single point in a graphic image
Each pixel is characterized
by some single value of radiation
(e.g., reflectance) impinging on
a detector that is converted by
the photoelectric effect into electrons
2Q
- see handout, Q is bit of each pixel
two prime approaches in the use of
remote sensing
• 1) standard photo-interpretation of scene
content
• 2) use of digital image processing and
classification techniques that are generally
the mainstay of practical applications of
information extracted from sensor data
sets
To accomplish this, we will utilize just one Landsat TM subscene that covers
the Morro Bay area on the south-central coast of California
?
Image below, brighter portions relate to higher energy levels
Landsat thematic mapper (TM)
Image interpretation
• relies on one or both of these approaches:
– Photointerpretation:the interpreter uses his/her knowledge and
experience of the real world to recognize scene objects
(features, classes, materials) in photolike renditions of the
images acquired by aerial or satellite surveys of the targets
(land; sea; atmospheric; planetary) that depict the targets as
visual scenes with variations of gray-scale tonal or color patterns
(more generally, spatial or spectral variability that mirror the
differences from place to place on the ground)
– machine-processing manipulations (usually computer-based)
that analyze and reprocess the raw data into new visual or
numerical products, which then are interpreted either by
approach 1 or are subjected to appropriate decision-making
algorithms that identify and classify the scene objects into sets of
information
Image Processing
• Computer-Assisted Scene Interpretation (CASI);
also called Image Processing
• The techniques fall into three broad categories:
– Image Restoration and Rectification
– Image Enhancement
– Image Classification
• There are a variety of CASI methods:
Contrast stretching, Band ratioing, Band
transformation, Principal Component Analysis,
Edge Enhancement, Pattern Recognition, and
Unsupervised and Supervised Classification
Classification
• Classification is probably the most informative
means of interpreting remote sensing data
• The output from these methods can be
combined with other computer-based programs.
• The output can itself become input for
organizing and deriving information utilizing what
is known as Geographic Information Systems
(GIS)
For the Morro Bay subscene the various images shown in this Section
were created using the IDRISI software processing program
(it's worthwhile to check their ClarkLabs website (http://www.clarklabs.org/)
[Clark University in Worchester, Mass.]).
The IDRISI program is especially user-friendly to students wishing to
gain experience in image processing
Image Classification
• In classifying features in an image we use
the elements of visual interpretation to
identify homogeneous groups of pixels
which represent various features or land
cover classes of interest. In digital
images it is possible to model this process,
to some extent, by using two methods:
Unsupervised Classifications and
Supervised Classifications.
• Unsupervised Classifications
this is a computerized method without direction from the
analyst in which pixels with similar digital numbers are
grouped together into spectral classes using statistical
procedures such as nearest neighbour and cluster
analysis. The resulting image may then be interpreted
by comparing the clusters produced with maps,
airphotos, and other materials related to the image site.
• Supervised Classification:
• Limitations to Image Classification:
have to be approached with caution because it
is a complex process with many assumptions.
In supervised classifications, training areas may
not have unique spectral characteristics
resulting in incorrect classification.
Unsupervised classifications may require field
checking in order to identify spectral classes if
they cannot be verified by other means (i.e.
maps and airphotos).
Online Reading
• http://hosting.soonet.c
a/eliris/remotesensing
/bl130lec10.html
Image
Type
Color
Levels
Pixel Value
8-bit
imag
e
28 = 256
0-255
16-bit
imag
e
216 = 65536
0-65535
24-bit
imag
e
224 =
1677721
6
0167772
15
Online Reading
•
Image Type
Pixel Value
Color Levels
8-bit image
28 = 256
0-255
16-bit image
216 = 65536
0-65535
24-bit image
224 = 16777216
0-16777215
Image Processing
• Digital Images:remote sensed images can also be
represented in a computer as arrays of pixels (picture
elements), with each pixel corresponding to a digital
number, representing the brightness level of that pixel in
the image. In this case, the data are in a digital format.
These types of digital images are referred to as raster
images in which the pixels are arranged in rows and
columns
Data visualization
The images that we view are visual representations
of the digital output from the sensor
 8-bit gray shade image is the case when the
sensor output is converted to one of 256 gray
shades (0 to 255)
 24-bit color does the same except in shades or
red, green, and blue
Image Processing: Pixel Values
• Pixel Values: The magnitude of the
electromagnetic energy captured in a digital
image is represented by positive digital
numbers.
• The digital numbers are in the form of binary
digits (or 'bits') which vary from 0 to a selected
power of 2
Image Type
8-bit image
16-bit image
24-bit image
Pixel Value
28 = 256
216 = 65536
224 = 16777216
Color Levels
0-255
0-65535
0-16777215
Image Processing: Image
Resolution
• Image Resolution: the resolution of a
digital image is dependant on the range in
magnitude (i.e. range in brightness) of the
pixel value. With a 2-bit image the
maximum range in brightness is 22 = 4
values ranging from 0 to 3, resulting in a
low resolution image. In an 8-bit image the
maximum range in brightness is 28 = 256
values ranging from 0 to 255, which is a
higher resolution image
2-bit Image
(4 grey levels)
8-bit Image
(256 grey levels)
Image Processing Procedures
• Image Restoration: most recorded
images are subject to distortion due to
noise which degrades the image. Two of
the more common errors that occur in
multi-spectral imagery are striping (or
banding) and line dropouts
Image Processing Procedures
• Dropped Lines are errors that occur in
the sensor response and/or data recording
and transmission which loses a row of
pixels in the image.
•
Image Enhancement
• One of the strengths of image processing
is that it gives us the ability to enhance
the view of an area by manipulating the
pixel values, thus making it easier for
visual interpretation.
• There are several techniques which we
can use to enhance an image, such as
Contrast Stretching and Spatial
Filtering.
Image Enhancement
• Image Histogram: For every digital image the pixel value
represents the magnitude of an observed characteristic such as brightness level. An
image histogram is a graphical representation of the brightness values that
comprise an image. The brightness values (i.e. 0-255) are displayed along the xaxis of the graph. The frequency of occurrence of each of these values in the
image is shown on the y-axis.
8-bit image
(0 - 255 brightness levels)
Image Histogram
x-axis = 0 to 255
y-axis = number of pixels
Image Enhancement
• Contrast Stretching: Quite often the useful data in a
digital image populates only a small portion of the available range of
digital values (commonly 8 bits or 256 levels). Contrast
enhancement involves changing the original values so that more of
the available range is used, this then increases the contrast between
features and their backgrounds. There are several types of contrast
enhancements which can be subdivided into Linear and NonLinear procedures.
Image Enhancement
•
Linear Contrast Stretch: This involves identifying lower and upper bounds
from the histogram (usually the minimum and maximum brightness values
in the image) and applying a transformation to stretch this range to fill the
full range.
•
Equalized Contrast Stretch: This stretch assigns more display values
(range) to the frequently occurring portions of the histogram. In this way, the
detail in these areas will be better enhanced relative to those areas of the
original histogram where values occur less frequently.
Linear Stretch Example:
Before Linear Stretch
After Linear Stretch
The linear contrast stretch enhances the contrast in the image with light toned
areas appearing lighter and dark areas appearing darker, making
visual interpretation much easier.
This example illustrates the increase in contrast in an image before (left) and after (right)
a linear contrast stretch.
Spatial Filtering
• Spatial filters are designed to highlight or suppress
features in an image based on their spatial frequency.
The spatial frequency is related to the textural
characteristics of an image. Rapid variations in
brightness levels ('roughness') reflect a high spatial
frequency; 'smooth' areas with little variation in
brightness level or tone are characterized by a low
spatial frequency. Spatial filters are used to suppress
'noise' in an image, or to highlight specific image
characteristics.
• Low-pass Filters
• High-pass Filters
• Directional Filters
• etc
Spatial Filtering
• Low-pass Filters: These are used to emphasize large
homogenous areas of similar tone and reduce the smaller
detail. Low frequency areas are retained in the image resulting
in a smoother appearance to the image.
Linear Stretched Image
Low-pass Filter Image
Spatial Filtering
• High-pass Filters: allow high frequency areas
to pass with the resulting image having greater
detail resulting in a sharpened image
Linear Contrast Stretch
Hi-pass Filter
Spatial Filtering
• Directional Filters:are designed to enhance linear features
such as roads, streams, faults, etc.The filters can be designed to enhance
features which are oriented in specific directions, making these useful for
radar imagery and for geological applications. Directional filters are also
known as edge detection filters.
Edge Detection
Lakes & Streams
Edge Detection
Fractures & Shoreline
Image Ratios
• It is possible to divide the digital numbers of one
image band by those of another image band to create a
third image. Ratio images may be used to remove the
influence of light and shadow on a ridge due to the sun
angle. It is also possible to calculate certain indices
which can enhance vegetation or geology
Sensor
Image
Ratio
EM Spectrum
Application
Landsat TM Bands 3/2
red/green
Soils
Landsat TM Bands 4/3
PhotoIR/red
Biomass
Landsat TM Bands 7/5
SWIR/NIR
Clay Minerals/Rock Alteration
For example:
Normalized Difference Vegetation Index (NDVI):
a commonly use vegetation index which uses the red and infrared bands
of the EM spectrum.
Image Ratio example: NDVI
NDVI image of Canada.
Green/Yellow/Brown
represent decreasing
magnitude of the
vegetation index.
Principle Components Analysis
• Different bands in multispectral images like those from Landsat
TM have similar visual appearances since reflectances for the
same surface cover types are almost equal. Principle Components
Analysis is a statistical procedure designed to reduce the data
redundancy and put as much information from the image bands
into fewest number of components. The intent of the procedure is to
produce an image which is easier to interpret than the original.
Data Visualization
Contrast enhancement or stretch reassigns the DN range that
corresponds to the 256 gray shades
Top row of images are ETM+ data with no enhancement and bottom row
consists of linear contrast stretches of the image DNs to the full 0-255 gray
shades
Data Visualization
Ability to quickly discern features is improved by using 3-band color mixes
Image below assigns blue to band 2, green to band 4, and red to band 7
Vegetation is green
Surface water is blue
Playa is gray and white
(Playas are dry lakebeds)
Color display
Rely on display hardware to convert between DN and gray levels
 Digital Numbers (DNs) are image data
 Grey Levels (GLs) are numerical display values
 Look-Up Tables (LUTs) map DNs —> GLs and change image
brightness, contrast and colors
 Actual displayed colors depend on the color response characteristics of
the display system
Data Visualization
Changing the color assignment to red, green, and blue does not alter the surface
material only appearance of the image
All images below show only combinations of bands 2, 4, and 7 of ETM+
Data Visualization
Other band combinations of the same data set bring out different features
(or in some cases lack there of)
All images below show only combinations of bands 2, 4, and 7 of ETM+
Multispectral display - CIR
• Visualize spectral content with 3band color composites
• Example: color infrared (CIR)
– red channel assigned to near IR
sensor band
– green channel assigned to red
sensor band
– blue channel assigned to green
sensor band
• vegetation appears red,
soil appears yellow - grey,
water appears blue - black
Image formats
File formats
File formats play an important role in that many are automatically
recognized in image processing packages
 Makes life very easy
 Raw data typically have no header information
 GeoTIFF is a variant of TIFF that includes geolocation information in
header (http://remotesensing.org/geotiff/geotiff.html)
 HDF or Hierarchical Data Format (http://hdf.ncsa.uiuc.edu/) is a selfdocumenting format
 All metadata needed to read image file contained within the image file
 First developed for web sites in the 1980s
 Allows for variable length subfiles
 EOS-HDF is NASA version (http://hdf.ncsa.uiuc.edu/hdfeos.html)
 NITF
 National Imagery Transmission Format
(http://remotesensing.org/gdal/frmt_nitf.html)
 Department of Defense
Data processing levels
Recently, operational processing of remote sensing data has led to multiple
processing levels
 “Standard” types of preprocessing
 Radiometric calibration
 Geometric calibration
 Noise removal
 Formatting
 Generic description
 Level 0: raw, unprocessed sensor data
 Level 1: radiometric (1R or 1B) or geometric processing (1G)
 Level 2: derived product, e.g. vegetation index