Camera Calibration

Camera calibration (see Chapter 3) significantly improves the accuracy of photogrammetric analyses. Chandler et al. (2005) found that radial lens distortion errors effectively constrain the accuracies achievable, making accurate modelling of lens distortion an important issue for the use of consumer-grade digital cameras. Investigations into the temporal stability of a digital compact camera by Wackrow et al. (2007) confirmed the relative importance of inaccurate lens distortion parameters as compared to internal geometry variations, which were found to be remarkably low over a one-year period.

Radius [mm]

FIGURE 6-14 Typical radial distortions for two SLR lenses, resulting from the calibration described in Table 6-4. Both lenses show slight pincushion distortion near the image center (radius < 5 mm) and increasing barrel distortion toward the edges.

Radius [mm]

FIGURE 6-14 Typical radial distortions for two SLR lenses, resulting from the calibration described in Table 6-4. Both lenses show slight pincushion distortion near the image center (radius < 5 mm) and increasing barrel distortion toward the edges.

TABLE 6-4 Interior orientation parameters for two DSLRs focussed to infinity.

Interior orientation parameter

EOS 350D with Canon EF 20/2.8

EOS 300D with Sigma 23/1. 8 EX DG

Focal length

20.2207 mm

26.B299 mm

Principal point X

-0.100B26 mm

0.035143 mm

Principal point Y

-0.009645 mm

0.171121 mm

Lens distortion parameter A1

-2.12E-04

-1.12E-04

Lens distortion parameter A2

4.0921 E-07

1 89E-07

Calibrated by IM, M. Koch, and M. Kahler using the test field of Berlin's University of Applied Science and Pictran photogrammetry software.

Calibrated by IM, M. Koch, and M. Kahler using the test field of Berlin's University of Applied Science and Pictran photogrammetry software.

For non-metric cameras, calibration reports are not provided by the manufacturer, but methods of camera calibration applicable to digital consumer-grade cameras have evolved rapidly over the last decades (Wackrow, 2008). Various calibration software, both commercial and non-commercial, exists as well as prefabricated 3D test fields with photogrammetric target points. However, these test fields are designed for close-range photogrammetry and usually are too small to be suitable for calibrating images focused at infinity (as in the SFAP case). For calibration of SFAP cameras, a larger test field is required, for example with high-precision targets mounted on the cornered walls and courtyard of a building. For small-format aerial photogrammetry, the commonly used alternative to test-field calibration is camera self-calibration during the actual project, where the elements of interior orientation are determined at the same time as the object points coordinates. The quality of the results, however, is highly dependent on the number, precision, and distribution of the ground control points involved (see Table 3-1).

Test-field calibration and field self-calibration procedures use the same concepts and methods as those outlined for object point reconstructions. The focal length, the position of the principle point and one to several lens distortion parameters are determined with iterative adjustment algorithms. Figure 6-14 and Table 6-4 show typical results for a consumer-grade DSLR camera.

6.7. SUMMARY

Any camera designed primarily for hand-held use on the ground may be adapted for small-format aerial photography. In spite of a tremendous range in cost and quality of such cameras, they all have certain basic components—lens, diaphragm, shutter, and an image sensor (film or electronic detector) within a light-proof box. Traditional cameras record photographs in the light-sensitive chemicals of the film emulsion. Spectral range is ~0.3-0.9 mm wavelength (near-ultraviolet, visible, near-infrared). Films of 35-mm and 70-mm formats are most commonly utilized for analog SFAP.

Digital cameras have come to dominate the market in the early twenty-first century, as cost has declined and quality has improved rapidly. Digital image sensors employ an array of tiny semiconductors to detect light intensity; two main types are the charge-coupled device (CCD) and the complementary metal oxide semiconductor (CMOS).

Three main controls of image exposure are the lens aperture (/-stop), shutter speed, and detector sensitivity (ISO rating). These factors are related by reciprocity, such that changing one variable by one stop can be matched exactly by changing another variable in the opposite direction by one stop. In most cameras for SFAP, automatic light settings are utilized. However, this may create problems for scenes with mixed illumination or highly contrasting bright and dark features. The geometrical and technical characteristics of the camera lens and sensor may have important influence on image distortions and artifacts.

Color-infrared photography normally utilizes the green, red, and NIR portions of the spectrum that are color coded, respectively, as blue, green, and red in the resulting false-color image. Both film and digital cameras may be utilized for color-infrared SFAP, although availability and processing of color-infrared film have become quite limited in recent years. The primary applications for color-infrared photography include vegetation types, soils, and water bodies.

Photogrammetric analysis of SFAP involves additional challenges. Among the important considerations are the camera lens, image sensor, file format, camera type, and camera calibration. In general DSLR cameras with single-focus lenses are most suitable for photogrammetric purposes and should be operated without image stabilization or dust removal functions.

Champion Flash Photography

Champion Flash Photography

Here Is How You Can Use Flash Wisely! A Hands-on Guide On Flash Photography For Camera Friendly People!. Learn Flash Photography Essentials By Following Simple Tips.

Get My Free Ebook


Post a comment