Image classification methods in the remote sensing

Digital image classification assigns pixels to classes. Each pixel is treated as a set of values in several spectral bands, derived from the brightness of the same area on the earth's surface. By comparing pixels to one another, and to pixels of known identity, it is possible to assemble groups of similar pixels into classes that are associated with the informational categories of interest to clients. Because these classes form regions on an image, after classification, the digital image is presented as a mosaic of uniform parcels, each identified by a color or symbol. These classes are, in theory, homogeneous-- pixels within classes are spectrally more similar to one another than they are to pixels in other classes. In practice, of course, each class will display some variation, because category exhibits natural variation. Image classification forms one of the most important tools for examination of digital images--sometimes to produce a final product, other times as one of several analytical procedures applied to derive information from an image.

Image classification requires the analyst to matchinformational classes to spectral classes. Informational classes are the categories of interest to the ultimate users of the data. Informational classes are (for example) the different kinds of forest, or the different kinds of land use that convey information to planners, managers, and scientists who will use information derived from remotely sensed data. These classes convey the information that we wish to derive from the data--they are the object of our analysis. Unfortunately, remotely sensed images do not directly convey informational classes-- we can derive them only indirectly, using the brightnesses that compose each image. For example, the image cannot directly show geological units, but rather only the differences in topography, vegetation, soil color, shadow, and other factors that lead the analyst to conclude that certain geological conditions exist in specific areas.

In contrast, Spectral classes are groups of pixels that are uniform with respect to brightnesses in their several spectral channels. The analyst defines spectral classes within remotely sensed data; then must define links between spectral classes on the image and informational classes that are of interest to the client. In this manner, image classification proceeds by matching spectral categories to informational categories. If the match can be made with confidence, then the information is likely to be reliable. If spectral and informational categories do not correspond, then the image is unlikely to be a useful source for that particular application.

Informational classes are typically composed of numerous spectral subclasses--spectrally distinct groups of pixels that together may be assembled to form an informational class. In digital classification, we must often treat spectral subclasses as distinct units during classification, but then display several spectral classes under a single symbol for the final image or map to be used by planners or administrators (who are, after all, interested only in the informational categories, not the intermediate steps required to generate them).

Overview of Process of Supervised Image Classification

1. select prospective training areas (training data polygons)
2. evaluate the training areas on the basis of
   • visible features
   • normalilty
   • separability
   • partitioning
3. make decisions on which training areas to use
4. classify the scene
5. evaluate the accuracy using an error (contingency) matrix

Unsupervised classification can be defined as the identification of natural groups, or structures, within multispectral data. The notion of the existence of natural, inherent groupings of spectral values within a scene may not be intuitively obvious, but it can be demonstrated that remotely sensed images are usually composed of spectral classes that internally are reasonably uniform in respect to brightnesses in several spectral channels. Unsupervised classification is the definition, identification, labeling, and mapping of these natural classes.

read more: 

Image processing - Classification

https://www.facebook.com/AssistanceSig
https://www.facebook.com/groups/317846591656057/

DIFFERENTIAL GPS

1.1 INTRODUCTION
Satellite navigation systems can provide far higher  accuracy than any other current long and medium
range navigation system. Specifically, in the case of GPS, differential techniques have been developed
which can provide accuracies comparable with current  landing systems. The aim of this chapter is to
provide an overview of current DGPS techniques and flight applications. Due to the existence of a copious
literature on GPS basic principles and applications, they will not be deeply covered in this dissertation.
Only a brief review of GPS fundamental characteristics is presented in Annex A, with an emphasis on
aspects relevant to the scope of this dissertation.

Differential GPS (DGPS) was developed to meet  the needs of positioning and distance-measuring
applications that required higher accuracies than stand-alone Precise Positioning Service (PPS) or Standard
Positioning service (SPS) GPS could deliver. DGPS involves the use of a control or reference receiver at a
known location to measure the systematic GPS errors; and, by taking advantage of the spatial correlation of
the errors, the errors can then be removed from the measurement taken by moving or remote receivers
located in the same general vicinity. There have been a wide variety of implementations described for
affecting such a DGPS system. It is the intent in this chapter to characterize various DGPS systems and
compare their strengths and weaknesses in flight applications. Two general categories of differential GPS
systems can be identified: those that rely primarily upon the code measurements and those that rely primarily
upon the carrier phase measurements. Using carrier phase, high accuracy can be obtained (centimeter level),
but the solution suffers from integer ambiguity and cycle slips. Whenever a cycle slip occurs, it must be
corrected for, and the integer ambiguity must be re-calculated. The pseudo-range solution is more robust,
but less accurate (2 to 5 m). It does not suffer from cycle slips and therefore there is no need for
re-initialization.


1.2 DGPS CONCEPT
A typical DGPS architecture is shown in Figure 1-1. The system consists of a Reference Receiver (RR)
located at a known location that has been previously surveyed, and one or more DGPS User Receivers
(UR). The RR antenna, differential correction processing system, and datalink equipment (if used) are
collectively called the Reference Station (RS). Both the UR and the RR data can be collected and stored
for later processing, or sent to the desired location in real time via the datalink. DGPS is based on the
principle that receivers in the same vicinity will simultaneously experience common errors on a particular
satellite ranging signal. In general, the UR (mobile receivers) use measurements from the RR to remove
the common errors. In order to accomplish this, the UR must simultaneously use a subset or the same set
of satellites as the reference station. The DGPS positioning equations are formulated so that the common
errors cancel.

The common errors include signal path delays through the atmosphere, and satellite clock and ephemeris
errors. For PPS users, the common satellite errors are residual system errors that are normally present in
the PVT (Position, Velocity, and Time) solution. For SPS users, the common satellite errors (typically
affected by larger ionospheric propagation errors than SPS) also included the intentionally added errors
from Selective Availability (SA), which have been removed with the current US-DoD policy. Errors that
are unique to each receiver, such as receiver measurement noise and multipath, cannot be removed without
additional recursive processing (by the reference receiver, user receiver, or both) to provide an averaged,
smoothed, or filtered solution [1]. Greater receiver noise and multipath errors are present in SPS DGPS
solutions.
Various DGPS techniques are employed depending on the accuracy desired, where the data processing is
to be performed, and whether real-time results are required. If real-time results are required then a datalink
is also required. For applications without a real-time requirement, the data can be collected and processed
later. The accuracy requirements usually dictate which measurements are used and what algorithms are
employed. Under normal conditions, DGPS accuracy is largely independent of whether SPS or PPS is
being used (although, as mentioned before, greater receiver noise and multipath errors are present in SPS
DGPS). When SA was on, real-time PPS DGPS had a lower data rate than SPS DGPS because the rate of
change of the nominal system errors was slower than the rate of change of SA. In any case, the user and
the Reference Station must be using the same service (either PPS or SPS).
The clock and frequency biases for a particular satellite will appear the same to all users since these
parameters are unaffected by signal propagation or distance from the satellite. The pseudorange and deltarange (Doppler) measurements will be different for different users because they will be at different
locations and have different relative velocities with respect to the satellite, but the satellite clock and
frequency bias will be common error components of those measurements. The signal propagation delay is
truly a common error for receivers in the same location, but as the distance between receivers increases,...
Read more

You find this page useful; thanks for making a donation by clicking one ads
Vous trouvez cette publication utile; merci de faire un don en cliquant un lien publicitaire