ABSTRACT
Plate 1: Optical measurement of oxygen in corals: a Sample place for the coral, the lagoon of Heron Island, Capricorn Islands, Great Barrier Reef, Australia. b Set-up with the coral in the glass container placed on top of the LED frame while the measuring camera was looking at the cut coral surface from below. c Close up of the set-up with a cut coral porite placed on top of a transparent planar oxygen optode, which was fixed at the bottom of a glass container, filled with lagoon sea water. The blue light is the excitation light, coming from blue LEDs arranged in a frame below the glass;, the orange-red light corresponds to the excited luminescence of the optode. The fixed white tube (upper right part) served for flushing the water surface with a constant air stream to aerate the water. d Measured 2D oxygen distribution (view to cut coral surface) given in % air saturation. The oxygen image is blended into a grayscale image from the coral structure. The high oxygen values were generated by endolithic cells, which live within the coral skeleton. The production was triggered by a weak illumination through the oxygen sensor, which simulated the normal light situation of these cells at daylight in the reef. Images courtesy of Dr. Holst, PCO AG, Kelheim, Germany. (See also Fig. 1.10, p. 10.)
a
b
Plate 2: Images of the planet Venus: a Images in the ultraviolet (top left) show patterns at the very top of Venus’ main sulfuric acid haze layer while images in the near infrared (bottom right) show the cloud patterns several km below the visible cloud tops. b Topographic map of the planet Venus shows the elevation in a color scheme as it is usually applied for maps (from blue over green to brownish colors). This image was computed from a mosaic of Magellan radar images that have been taken in the years 1990 to 1994. Source: http//www.jpl.nasa.gov. (See also Fig. 1.15, p. 15.)
Plate 3: Synthetic color image of a 30.2× 21.3 km sector of the tropical rain forest in west Brazil. Three SAR images taken with different wave lengths (lower three images: Left: X-band, Middle: Cband, Right: L-band) have been composed into a color image. Pristine rain forest appears in pink colors while clear areas for agricultural usage are greenish and bluish. A heavy rain storm appears in red and yellow colors since it scatters the shorter wavelength micro waves. Image taken with the imaging radar-C/X-band aperture radar (SIR-C/X-SAR) on April 10, 1994 on board the space shuttle Endeavour (source: image p-46575 published in https://www.jplinfo.jpl.nasa.gov). (See also Fig. 1.18, p. 18.)
Plate 4: Images of the Topex/Poseidon mission derived from radar altimetry and passive microwave radiometry. a dynamical topography (highs and lows of the ocean currents) shown as a deviation from an area of constant gravitational energy. b significant wave height, c water vapor content of the atmosphere in g/cm2 measured by passive micro wave radiometry, d wind speed determined from the strength of the backscatter. All images have been averaged over a period of 10 days around the 22 December 1995 (source: https://www.jplinfo.jpl.nasa.gov). (See also Fig. 1.17, p. 17.)
a
b
c
d
Plate 5: a Artist’s impression of an X-ray telescope, the so-called Wolters telescope, the primary X-ray instrument on the German ROSAT satellite. Because of the low index of refraction, only telescopes with grazing incident rays can be built. Four double-reflecting pairs of concentric mirrors are used to increase the aperture and thus brightness of the image. b X-ray image of the moon, with 15’ (1/4 °) diameter a small object in the sky as compared to c Explosion cloud of the supernova Vela with a diameter of about 4 °as observed with the German X-ray satellite ROSAT. Scientists of the Max Planck Institute for Extraterrestrial Physics (MPE) discovered six fragments of the explosion marked from A to F. The intensity of the X-ray radiation in the range from 0.1 to 2.4 keV is shown in pseudo colors from light blue over yellow, red to white and covers an intensity range of 500. The bright white circular object at the upper right edge of the explosion cloud is the remains of another supernova explosion which lies behind the VELA explosion cloud and has no connection to it. d Enlarged parts of figure c of the fragments A through F at the edge of the explosion cloud. (See also Fig. 1.19, p. 20.)
Plate 6: A series of 32 confocal images of the retina. The depth of the scan increases from left to right and top to bottom. From Zinser, 1995. (See also Fig. 1.23, p. 24.)
a b
Plate 7: Reconstruction of the topography of the retina from the focus series in Fig. 1.23. a Depth map: deeper lying structures as the exit of the optical nerve are coded brighter. b Reconstructed reflection image showing all parts of the image sharp despite significant depth changes. From Zinser, 1995. (See also Fig. 1.24, p. 24.)
a b c d
e f g h
i
Plate 8: a - h Part of a depth scan using confocal microscopy across a female human nucleus. Visible are chromosomes X and 7 (green). For differentiation between chromosomes X and 7, a substructure of chromosome 7 has been additionally colored with a red dye. The depth increment between the individual 2-D images is 0.3 µm, the image sector is 30× 30 µm. i 3-D reconstruction. The image shows the inactive chromosome X in red, the active chromosome X in yellow, chromosome 7 in blue, and its centromeres in magenta. The shape of the whole nucleus has been modeled as an ellipsoid. From Eils, 1995. (See also Fig. 1.25, p. 26.)
a b
c
Plate 9: Fluorescent dyes can also be used to make flows visible: a visualization of a mixing process; b and c two consecutive 15× 15× 3mm3 volumetric images consisting of 50 layers of 256× 256 pixels taken 100ms apart. The voxel size is 60× 60× 60µm3. From Maas [94]. (See also Fig. 1.29, p. 29.)
Plate 10: Flow field determined by the least squares matching technique from volumetric image sequences. The flow vectors are superimposed to the concentration field. From Maas, 1995. (See also Fig. 1.30, p. 29.)
Plate 11: Screenshot of the graphical interface of a modern image processing software (heurisko®) featuring multi-window image display, flexible image “inspectors” for interactive graphical evaluation of images, and a powerful set of image operators, which can easily be combined to user-defined operators adapted to specific image processing tasks. This software package is not limited to 2-D image processing with 8 bit depth (256 gray scales) but can also handle multi-channel images of various data types, multi-scale images (pyramids), volumetric images, and image sequences. (See also Fig. 2.11, p. 47.)
a b
Plate 12: A computer generated (rendered) 3-D scene illustrating the complexity of the interactions between illumination and objects. a flat shading (uniformly radiating objects), the ideal world for image processing. b Gouraud shading (matte Lambertian surface) with multiple point light sources. Now the radiance of the objects is no longer uniform. It depends on the angle between the surface normal and the direction to the light sources. Furthermore, other objects that are between a light source and the object obstruct light of this light source to reach the object resulting in shadows. (See also Fig. 3.18, p. 86.)
a b c d
Plate 13: Color wedge (a ) produced by additive color mixing from the b green, c red, and d blue color wedges according to Eq. (3.78). (See also Fig. 3.25, p. 94.)
a b c
d e f
g h
Plate 14: Illustration of color ratio imaging used with the shape from reflection technique to measure the slope of water surfaces. It is demonstrated here with a calibration target, a spherical lens. a original color image; b intensity of the original image; watch the intensity fall off towards the edge of the lens (higher slopes); c color image normalized by b; d - f : Red, green, and blue channels of the original color image shown in discrete intensity steps; g and h : x, y positions in the telecentric illumination system computed after Eq. (3.80). These quantities are according to Eq. (3.77) directly proportional to the x and y components of the surface slope. (See also Fig. 3.26, p. 95.)
a
b
Plate 15: Color bar used as a standard video signal: a The four horizontal stripes contain the following components of the signal from top to bottom: red channel only, green channel only, blue channel only, red, green, and blue channel together. b Color bar signal shown as a PAL composite video signal digitized with a gray scale frame grabber that does not suppress the color carrier frequency and thus shows the colors as high frequency patterns. Notice that the white stripe is free of the high frequency patterns since it contains no color (the U and V components are zero). Compare also with Fig. 5.26.
