CHARACTERISTICS OF REMOTELY SENSED IMAGERY Spatial Resolution
There are a number of ways in which images can differ. One set of important differences relate to the various resolutions that images express. Resolution: the fineness of detail that can be distinguished in an image. Detail, however, can refer to several different dimensions. RESOLUTION?
HOW CAN IMAGES DIFFER?
How is the scale of a photograph determined? FFFFFFFFFF lllllllllll AAAAAAAAAAAAAAAA HOW HIGH CAN WE GO?
Altitude 681 km Low Earth Orbit satellites
MODIS: Yellow River Delta 2009 250 m spatial resolution @ 705 km MODIS: Ice and Clouds in the Bering Strait Low Earth Orbit satellites
AVHRR 1 km resolution @ a nominal altitude of 833 km Low Earth Orbit satellites
MIDDLE EARTH ORBITS Also
GOES image, spatial resolution varies from 1 to 13+ km @ 35,800 km above the Earth Geostationary satellites
SPATIAL RESOLUTION
The fineness of detail visible in an image. For photographic film, a function of the grain size and the camera lens (line pairs/mm) For sensors, a function of the sensor s characteristics (IFOV) [Angle A] For any sensor atmospheric conditions, the viewing angle (perpendicular or oblique), the height of the sensor s platform, and the stability of the platform all affect the resolution. (View the science education interactive footprint demo.) SPATIAL RESOLUTION
Contrast enhancement: Something you ve worked on in your labs. Kutztown University in Pennsylvania where we can discern details far smaller than the 0.5m pixel size. GeoEye-1 APPARENTLY COARSER OR FINER
Other factors that influence the spatial resolution include: Characteristics of the features being observed (high contrast vs low contrast e.g., lines in the GeoEye image) Relation of the feature geometry to the sensor geometry (see also mixed pixels) How the data is displayed View the what a sensor sees interactive demo. SPATIAL RESOLUTION
Image processing can increase our ability to discern features
There are several terms used in describing the spatial resolution of data: Ground Sampling Distance (GSD): for digital imagery, the ground distance represented by the width of a pixel [B] Ground Resolving Distance (GRD): for photographic film, the size of the smallest object expected to be detected Both are a function of the scale at which the image is displayed. SPATIAL RESOLUTION
The true resolution of the sensor [Collection GSD] may differ from the pixel size [Product GSD] (data can be resampled and presented as being coarser or finer than it inherently is). 10 m resolution, 10 m pixel size 30 m resolution, 10 m pixel size 80 m resolution, 10 m pixel size DATA DISPLAY
IKONOS image resampled to 40m 4 meter resolution IKONOS image with 40 meter grid overlaid. DATA DISPLAY
If we assume that within each pixel there will be a unique set of Digital Numbers (DN) that we can match to a single spectral reflectance curve, mixed pixels are obviously a significant problem. MIXED PIXELS
Higher spatial resolution can reduce the number of mixed pixels, but at a cost. MIXED PIXELS
With high spatial resolution comes aliasing, where a non-existent pattern appears because of sample spacing. With crops, spatial sampling on the scale of a meter combines with the row spacing to alias pseudo-rows that are tens of meters wide. The apparent rows are false. Visual clues to this are seen by comparing the apparent row spacing with the size of roads and homes. The imagery was taken with a nominal ground resolution of approximately 2 meters. The spacing of the rows is less. SPATIAL RESOLUTION AND ALIASING
Combining higher spatial resolution data with higher spectral resolution data is called pan sharpening Quickbird data (50cm panchromatic, 2m multispectral) PAN SHARPENING
There are several methods by which pan sharpened images are produced. One common method is HSI, which stands for "Hue Saturation Intensity" (ESRI s description). The lower resolution RGB image is upsampled (i.e., resampled to match the higher spatial resolution pixel size) and converted to HSI space. The panchromatic band is then matched and substituted for the Intensity band. The HSI image is converted back to RGB space. PAN SHARPENING
High-resolution panchromatic images Low-resolution natural colour images: QuickBird (spatial resolution 0.6 m) (spatial resolution 2.4 m) PAN SHARPENING
PICK YOUR SPATIAL RESOLUTION
The highest resolution data isn t always necessary, or even the best to use. PICKING THE RIGHT SPATIAL RESOLUTION
CAN T SEE THE FOREST FOR THE TREES
Basic elements Coarser-resolution imagery Finer-resolution imagery Primary Tone Colour Spatial arrangements of tone Size Shape Texture Pattern Height Shadow Secondary Tertiary Degree of complexity Site Association Quaternary Identifying forests and urban areas vs collections of individual trees and houses, streets, lawns, etc. ELEMENTS OF IMAGE INTERPRETATION
SPATIAL / SPECTRAL RESOLUTIONS
Fine/high spatial resolution > small IFOV Small IFOV > reduction in energy detected > less radiometric resolution increase the amount of energy detected (radiometric resolution) without reducing spatial resolution broaden the wavelength range detected for a particular channel or band (decrease the spectral resolution) reduce spatial resolution improved radiometric and/or spectral resolution. SPATIAL / RADIOMETRIC RESOLUTION
MULTI-RESOLUTIONS FOR INTEGRATED STUDIES.
Spatial resolution has steadily increased over time. Higher resolution data comes with added costs: Size of files More intelligent analysis of the data required Smaller areas sampled (footprints of sensors) But also added benefits: Many more application areas (esp. urban) SUMMARY