Spectral Image Acquisition
In this chapter, we review benchtop imaging spectroscopy acquisition methods and address the process of dark current calibration. We commence by turning our attention to the structure and spectral responses of spectral cameras and sensors. We do this by e
- PDF / 1,064,128 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 99 Downloads / 195 Views
Spectral Image Acquisition
Throughout the book, we make very little, if any, differentiation between the use of spectra in the visible, ultraviolet and near-infrared ranges. This is because applications of imaging spectroscopy for scene understanding using benchtop hyperspectral and multispectral cameras are not necessarily constrained to the range visible to the human eye. Moreover, we will often refer to the electromagnetic radiation in terms of its wavelength. In particular we will focus on wavelengths spanning from the ultraviolet (UV) to the infrared in its A and B bands (IR-A and IR-B). Here, we use the naming convention of the International Commission on Illumination (CIE) in preference over that proposed by the International Organization for Standardization (ISO). In Table 2.1, we summarise the two main subdivisions of the spectrum, i.e. those corresponding to the CIE and the ISO. As mentioned earlier, a hyperspectral or multispectral image comprises a set of wavelength indexed bands sampled over a spectral range. In Fig. 2.1, we show the electromagnetic spectrum, divided following the CIE standards, and illustrate how a landscape appears at different bands. As illustrated in Fig. 1.2, the appearance of the scene shows noticeable changes with respect to wavelength. Since every pixel in the image accounts for a set of wavelength resolved measurements, the acquisition process across the spectral range greatly depends on the camera used to capture the image. Furthermore, the sensor itself may require calibration to account for noise and bias. In this section, we provide an overview of the existing spectral imaging technologies, reduction of the camera sensor noise and rectification of the image spectra with respect to the illumination.
2.1 Spectral Cameras and Sensors Although imaging spectroscopy has been available as a remote sensing technology since the 1960s, until recently, commercial spectral imaging systems were mainly airborne ones which could not be used for ground-based image acquisition. Furthermore, spectral imaging has often only been available to a limited number of A. Robles-Kelly, C.P. Huynh, Imaging Spectroscopy for Scene Analysis, Advances in Computer Vision and Pattern Recognition, DOI 10.1007/978-1-4471-4652-0_2, © Springer-Verlag London 2013
9
10
2
Spectral Image Acquisition
Table 2.1 Spectral subdivisions according to the CIE and the ISO Ultraviolet ISOa
UV-C (100–280 nm)
UV-B (280–315 nm)
UV-A (315–400 nm)
CIEb
UV-C (100–280 nm)
UV-B (280–315 nm)
UV-A1 (315–340 nm)
UV-A2 (340–400 nm)
Visible ISO
VIS (400–780 nm)
CIE
VIS (400–700 nm)
Infrared ISOc
Near-Infrared (780 nm–3 µm)
CIE
IR-A (700 nm–1.400 µm)
a See
Mid-Infrared (3–50 µm) IR-B (1.4–3 µm)
Far-Infrared (50 µm–1 mm)
IR-C (3 µm–1 mm)
ISO-DIS-21348 and, for the infrared range, see ISO 20473:2007
b See
134/1 TC 6-26 report: Standardization of the Terms UV-A1, UV-A2 and UV-B
c See
ISO 20473:2007
Fig. 2.1 The ultraviolet, visible and infrared spectrum as related to the wavelength resolved bands correspond
Data Loading...
