9.2. Technical Description
This section describes the major components and the basic design
concepts of imaging system sensors. The different types of sensors can
be categorized by the three-dimensional image data acquisition
modes. The 3D image data cube is defined by spatial,
x and y, and
wavelength,
dimensions (Figure 9-1).
One approach to collecting an image would be to sample the two
spatial dimensions while temporally sampling the
dimension. An example of this can be found in a framing 2D
array. A single image band from a framing array can be simulated
by the plate, b, whose thickness,
, is determined by the spectral bandpass.
Scanning along the spectral dimension, the single-band image
spans along the
axis. The pushbroom method utilizes, among others, an
entrance slit and an imaging spectrometer with a dispersive element
to allow a 2D detector to sample the
and the x
dimensions simultaneously. The image along the second
spatial dimension,
y, is typically generated by scanning
the 1D spectral image, which is described by plate,
c, of finite width,
along the y dimension. Alternatively,
the line scanner employs a scanning
mirror to project the image of a single detector along a line on
the ground. This is shown as column a in the
figure.
Figure 9-1. The different data-acquisition modes (``Hyperspectral prism-grating-prism imaging spectrograph'', Mauri Aikio, Doctoral Dissertation, University of Oulu, Finland, p.21, June 2001);
9.2.1. Line Scanner
This instrument type consists of a single detector or a 1D focal plane that acquires the image using a constant angular sampling in the across-track direction and constant spatial sampling in the along-track direction. The along-track sampling is provided by the movement in the platform's position. (Figure 9-2) shows a basic design scheme for this type of instrument.
Of all the types of instruments, this is the simplest imaging system. These sensors employ a spinning, scanning, mirror to project the image of the detector along a line on the ground perpendicular to the aircraft or satellite ground track. By sampling the signal from the detector, the across-track image lines can be formed. During the rotation of the scan mirror, the sensor platform advances slightly, and consecutive rotations of the mirror sweep out consecutive lines on the ground, which are sampled to form the across-track lines that make up the image. The angular extent of the image across-track is referred to as the field of view (FOV) of the imager. The angular extent of the individual detector element is called the instantaneous field of view (IFOV). The projection of the detector onto the ground is referred to as the ground instantaneous field of view (GIFOV).
Conventional line scanners use square detectors that are sampled along the track on pixel centers and the ideal ground track advances one GIFOV per rotation to sample contiguous lines with each rotation. Thus every point on the ground is imaged, or sampled, without gaps and without overlaps. Some systems employ oversampling to improve the spatial resolution of the reconstructed image.
Unlike most non-framing imagers, line scanners have a unique set of geometric distortions caused by the way the image is sampled and by the motion of the sensing platform during imaging. Airborne platforms are often not stabilized so that the orientation of the aircraft can change from one line to the next, and in other cases from pixel to pixel within a line.
9.2.2. Framing Array
This instrument requires a 2D focal plane that acquires the entire spatial image ``instantaneously''. This approach uses a 2D array of sensors and a shutter to control the dwell time. Since the forward motion of the sensor will blur the image, exposure times with this type of sensor must be restricted to less than the time it takes to move the sensor one GIFOV. A major advantage to this type of approach is in terms of geometric fidelity. Since the entire image is acquired simultaneously, distortions due to within-frame sensor motion are essentially eliminated. Figure 9-3 shows a simple scheme for the framing array.
9.2.3. Pushbroom Scanner
This is a 1D or 2D focal plane that acquires the image using constant spatial sampling in the across-track direction and constant spatial sampling in the along-track direction. The along-track sampling is provided by the movement in the platform's position. Figure 9-4 shows a concept diagram for a pushbroom instrument.
The pushbroom sensors represent a further step toward increasing the dwell time to allow system designers to make signal-to-noise or resolution trade-offs. These sensors use linear array detectors to collect entire lines of data simultaneously. Multiple spectral bands are collected, with multiple linear arrays filtered for the bands of interest. Using this approach, an individual detector element only needs to sample at one across-track location. This provides increased integration time for the sensor.
In order to generate an image, this system will image in one spatial dimension, across-track direction, and in the spectral dimension. This is typically generated by scanning or moving the imaging system's field of view relative to the scene.
9.2.4. Imaging Spectrometer
The emergence of new detector technologies has made possible the development of remote scanners that are capable of capturing image data in a large number of spectral channels. This is often referred to as hyper-spectral imaging. A typical hyper-spectral imaging system will record data over hundreds of spectral channels. Given the spectral detail available in imaging spectrometer data, this allows us to accurately characterize the pixel's spectral reflectance over the visible and reflected infrared region. Important spectral absorption and other diagnostic features can be identified through analysis of fine spectral features in individual pixels. The large number of narrow channels and its contiguous nature also contributes to a significant degree of redundancy in imaging spectrometry data. Data compression methods can be used to exploit this feature, thereby reducing the data volume significantly. Figure 9-5 highlights the main concepts of a spectrometer.
