Instructor

Dr. Anthony Vodacek
76/3104
475-7816
vodacek@cis.rit.edu

Aerial Photography

Airphotos, Wisconsin

Airphoto, Minnesota

 

Link to Aerial Photos of Minnesota Forests
 
 

Advanced Technology in Forestry
- satellite imagers, hyperspectral imager, digital camera, lidar and radar
 
 

Global Tropical Forest Inventory Using AVHRR

NDVI - Normalized difference vegetation index

Southeast Asia

 

Central and South America

 

Change Detection Using Landsat (from CCRS)
Landsat TM
    Red - Band 5 (1550-1750 nm) from 1991
    Green - Band 4 (750-900 nm) from 1984
    Blue - Band 5 (1550-1750 nm) from 1984

 

Digital Camera with three spectral channels
    675 nm - Chlorophyll absorption
    698 nm - Red edge
    840 nm - near IR

Classified image using a supervised Gaussian
maximum likelihood classifier

Mosaics

 
 

A paper describing the use of AVIRIS spectral data to improve forest vegetation type mapping
 
 
 

LIDAR

Lidar + NIR channel for 3D visualization

Stand maps - end product for use by foresters

 
 

Imaging Radar,  Airborne and Spaceborne
All images from the Canada Center for Remote Sensing

C and X band Synthetic Aperature Radar, airborne
C-Band 3.75-7.5 cm
X-Band 2.4-3.75 cm

Polarization effects from vegetation
HH penetrates to the ground more
VV is sensitive to vegetation more, for example, grasses versus broadleaf crops
HV is more sensitive to forest canopies because of multiple scattering

Look angle effects

Corner reflectors for brightness calibration

 

Airborne image examples

The above imagery was acquired using the CCRS C-SAR (C-HH) in nadir
mode, with a pixel spacing of 3.89 metres (azimuth) by 4.0 metres (range). In
both cases, the aircraft was flying in a roughly southerly direction (202° on
March 13/91 and 203° on February 8/92) and looking westward. The
overflights were done during the winter to improve the contrast between the
cutovers (which would be snow-covered) and the surrounding standing
forest. The imagery was UTM-registered with cubic convolution
re-sampling to a 5 metre x 5 metre pixel spacing.
 

Radarsat, spaceborne
5.3 GHz, HH polarization
 

Whitecourt, Alberta.  Radarsat 1

The image collected on March 5, 1996 was in position 4 of the Fine Beam Mode or F4.
Data collected in F4 have a nominal incidence angle of 43 to 46 degrees. The
clearcuts are easy to identify on this image. Their dark tone contrasts well with the
brighter returns from the surrounding Boreal forest. The 25 metre pixel spacing, allows
clearcuts (A) to be easily viewed. If you look closely you can see a pumping station and
gas pipelines (B) and a very steep sided valley (C). Some of the roads (D) accessing
the area are also visible.
 

Whitecourt, Alberta.  Radarsat 1

The second image was acquired over the same location, on Feb.12, 1996, in a
Standard Beam Position - 1 or S1. S1 provides SAR data with a nominal incidence
angle of 20 to 27 degrees. The steeper incidence angles of this Standard mode
image provide poorer contrast between the clearcuts and the forest.

Compare the two RADARSAT images for clearcut identification. Both the incidence
angle and the resolution differences (and/or rendered pixel spacing) will change the
interpreter's ability to discriminate clearcuts.  At full enlargement, the fine mode data
provided here has a pixel spacing of 25 metres, while the standard mode is presented
at 56 metres.

ERS-1, Whitecourt Alberta

An ERS-1 image has been provided to allow a comparison between the
two satellites. The ERS-1 satellite, operated by the European Space Agency, provides
C-Band SAR with VV polarization. The image was collected on Jan. 22, 1993.
ERS-1 acquires data with a resolution of 30 metres, but is presented here at
  approximately 40 metre pixel spacing. The nominal incidence angle of ERS-1 is 22 to
24 degrees. Although its nominal resolution and incidence angles are similar to the
Standard beam1 RADARSAT image shown above, the clearcuts are almost impossible to
see. Why? It is quite likely this discrimination difference is the result of the
polarization differences between the two sensors. RADARSAT is a C-HH system
while ERS-1 & 2 are C-VV systems. Other environmental and/or engineering
differences may also play a role.