“The Geography Discipline of the United States Geological ...



DRAFT

May 2, 2007

A PROPOSAL FOR USGS TO IMPLEMENT THE RECOMMENDATIONS OF THE 1999 ASPRS PANEL CONCERNING AIRBORNE CAMERA CALIBRATION

==================================

BACKGROUND:

The US Geological Survey (USGS) has the responsibility for calibrating airborne mapping cameras for the national aerial mapping community. It is noted that the USGS regards airborne film-based cameras as one of their most important sources of geo-spatial data.

“The Geography Discipline of the United States Geological Survey (USGS) has the mission to provide basic cartographic information for the United States. Primary sources for these data are airborne film mapping cameras. Since 1973, the USGS in Reston, VA. has been responsible for calibrating these cameras for the aerial mapping community.… “ [ ]

Calibrations are currently accomplished at the Reston facility by laboratory methods in a well-controlled environment. Results of the laboratory methods include estimates of the values of the elements of camera interior orientation and lens resolution.

It has been 8 years since a review panel, appointed by the ASPRS at the request of the USGS, made recommendations to the USGS regarding camera calibration. Of the four recommendations, only the third has not been implemented. The third recommendation was “Undertake development and implementation of operationally based [in situ] methods of calibration for aerial applications combining cameras, GPS and INS sensors.”

As an example of the metric characteristics of the film-based camera subject to an in situ calibration, reference is made to results obtained by the US Air Force working with the USQ-28 geodetic subsystem using the Fairchild KC-6A film camera in a RC-135 aircraft (D. Brown, 1969). The photography was flown at night and a series of illuminated ground control targets were used. The positions of the exposure stations were obtained by three ground-based stellar cameras and using star background for control. The internal consistency of the photogrammetric procedure approached 1:1,000,000. Three ray intersections on ground control exhibited geo-spatial accuracies of 1:300,000 in the horizontal and 1:100,000 in elevation. These results clearly demonstrated the spatial accuracy potential of the air-borne, film-based cameras.

The Global Positioning System (GPS) has made it possible to determine the individual photo exposure station coordinates to an accuracy of several centimeters. When GPS imagery is collected over a suitably targeted and controlled range, environmental influences on the elements of interior orientation will be included in the resulting calibration. This approach is termed in situ calibration and results in a more accurate estimate of the camera’s interior orientation, It also provides a lower cost of operational resources, and is more convenient for operators.

To broaden in situ calibration experience, a series of flight tests under more conventional circumstances than those of the USQ-28 program, have been conducted with airborne film-based cameras (Merchant, Schenk, Habib and Yoon, 2004). In this report, the consistent improvements offered by in situ methods of calibration when compared to the laboratory method are reported. The report describes different cameras (Zeiss LMK/15-23, Wild RC30 15/23), different calibration ranges (low, mid and high altitude ranges/targets), and different aircraft configurations (open port at 1260 meters and at 3070 meters above ground, and windowed port at 5817 meters above ground [NOAA Cessna Citation, see Figure 1.]). All results exhibited significant differences from those provided by the laboratory calibrations conducted on the same cameras.

Figure 1. The NOAA Citation Preparing for High-Altitude Calibration Over a Madison, Ohio Test and Calibration Range

This paper discusses in situ calibration procedures and spatial accuracies compared to those produced by laboratory methods and also typical calibration range requirements. To conclude, it is proposed that the in situ method be adopted as an alternative to the currently required laboratory method of calibration.

THE IN SITU METHOD OF CALIBRATION:

Preparations for this method of calibration include instrumentation of the aircraft with camera and GPS equipment. After the GPS antenna has been installed, the spatial offset coordinate values between the camera entrance node and the GPS phase center are measured. These measurements are in a system parallel to the camera coordinate system and can readily be accomplished in the field to an accuracy of a few millimeters.

Recent research has indicated that the design of the targeted range can be based on simple crossroad patterns with targets radiating from the center at about 12.5 degrees as measured at the planned exposure stations. Based on synthetic data, one photo flying in the direction of a road, a second flying in the direction 45 degrees from the first, and taken over the intersection of roads will provide sufficient data to calibrate the camera with all functional correlations suppressed below a practical level of significance. The average functional correlation is typically 1.0e-004.

In practice, to increase the density of image locations, four photos may be required. It is important to note that since a simple crossroad pattern is sufficient for range geometry, suitable calibration ranges could readily be located on all but in mountainous regions.

Figure 2. provides an indication of a possible calibration range located 12 miles SSW of the Harrisburg, PA airport. Depending on the typical flight heights above ground, details such as target size and distances along roadways would need to be determined before a final calibration site selection could be made. In this case (Figure 2.), a typical flight height would be about 3000 feet.

Figure 2. Typical calibration Range Site a Few Miles Southwest of Harrisburg, PA (from Google Map)

COMPARATIVE SPATIAL ACCURACIES:

For purposes of spatial accuracy assessment, GPS produced coordinates of each exposure station can provide a standard of comparison to the photogrammetrically computed spatial values.

Comparisons were made for the ODOT Zeiss LMK 15/23 film camera. The comparisons were made between spatial coordinates of exterior orientation computed by single photo resections and those provided by GPS. One comparison, based on camera interior orientation, was made from a laboratory calibration and a second from an in situ calibration (see Figure 3.). This comparison is intended to represent the procedure for comparing the photogrammetric results to a standard of accuracy provided by GPS. Consistency among similar results was referenced above.

Figure 3. Photogrammetrically Resected Positions Compared to GPS

When using the laboratory calibration, substantial differences in elevation existed between elevations determined by photogrammetric resections and by GPS (see Figure 4.). For this example, when using the laboratory calibration, differences in elevation in terms proportional to the flight height approached 1:500. When an in situ calibration was used in the photogrammetric resection, differences in spatial coordinates when compared to those produced by GPS, reduced to 1:20,000.

Figure 4. Typical Spatial Coordinate Comparisons to GPS

CALIBRATION SOFTWARE:

The programs used in the above example were IM for image measurement and BASC for calibration (Merchant, Schenk, Habib and Yoon, 2004a). The results of the in situ calibration are provided in Table 1.

===============================================================

CAMERA: Jena LMK 2000 / Serial No. 272296C

LENS: Jena Lamegon PI/D / Serial No. 7390595 / Focal Length 151.898 mm

MAGAZINE: No. 1; Filter yellow

PHOTOGRAMMETRIC SCANNER: Z/I Imaging, PhotoScan 2002, #1101-873

Date of Calibration Photography; September 22, 2005 / Flown at 1500 feet agl

Calibration Range: Madison Range, London, Ohio

Math Model: USGS version of the SMAC model

The calibration was accomplished by using aerial photography flown in a Partenavia Observer aircraft, scanned at 21 micron pixel size using the system operated by the Ohio Department of Transportation (ODOT), Aerial Engineering. Image measurements were made by using the program IM (USGS). Calibration results were produced by using the program BASC (USGS).

Results of Calibration:

Number of Photographs observed: 5

Number of Observed images 232

RMSE of image observations: 0.0064 (mm)

Temperature: At lens = ranges as function of time 26.0 to 30.0 C

At film plane = same as at lens +0.2 C

Calibrated Focal Length: 151.898mm; StdDev. = 0.0039 mm

Principal Point: x0 = -0.005 mm ; StdDev. = 0.0066 mm

y0 = -0.022 mm StdDev. = 0.0067 mm

Distortion Parameters:

K0: 0.0000000000e+000

K1: -7.5008723051e-009

K2: -1.1928112069e-014

K3: 0.0000000000e+000

P1: 1.8218994636e-007

P2: -1.7315949589e-007

P3: 0.0000000000e+000

StdDev(K0): 0.0000000000e+000

StdDev(K1): 4.3254489777e-009

StdDev(K2): 1.7159635558e-013

StdDev(K3): 0.0000000000e+000

StdDev(P1): 1.1360294197e-007

StdDev(P2): 1.1447494436e-007

StdDev(P3): 0.0000000000e+000

==========================================================

TABLE 1. Typical Report of in situ Camera Calibration

Results of the calibration include estimates of all significant elements of interior orientation with associated error estimates. Error estimates are computed by means of the general law of error propagation.

Note that the in situ methods also include the influences of platen unflatness, filter wedge effect (decentering model derived by Brown from earlier work by Washer(1956)), and the possible influences of the image scanner employed as a film to image digitizer.

An additional parameter that characterizes the operational camera system is resolution. In the laboratory, lens resolution is measured based on the MIL Std 150 3-bar images on the reticules of the individual collimators. Under stable and environmentally controlled conditions, using extremely fine grain emulsions, resolutions generally exceed AWARs of 100 lines/mm. Under operational conditions, subject to camera mount image stabilization and using typical emulsions such as AGFA 150 or EKC Aerocolor, operational resolutions are reduced to 26 lines/mm (Topo Photo Inc., 1995).

Fiducial mark coordinates are provided with the original camera calibration (not in situ). As part of the image measurement process, the agreement of the observations with the fiducial coordinates is a measure of the quality of the fiducial data. Statistically significant departures from the published coordinates may be used to determine if it may be necessary to re-calibrate the fiducial marks.

CONCLUSION:

The issue of film-based, mapping camera calibration is of central interest to the mapping community. To implement the in situ methods of aerial camera calibration it remains to develop acceptable specifications for design of targeted calibration fields and specifications for flights and image coverage. In addition, a training of operational personnel may be required.

Given the facts that the in situ method of camera calibration:

• provides a more spatially accurate result

• can be accomplished without removing the camera from the aircraft

• can be accomplished for less user resources

Therefore, in the interest of better serving the mapping community, it is proposed that the USGS accept the airborne (in situ) method of film based camera calibration as an alternative to the current labortory methods used in the Optical Science Laboratory of the USGS.

REFERENCES:

Brown, D. C. (1969), “Advanced Methods for the Calibration of Metric Cameras”, DBA Report presented at the Symposium on Computational Photogrammetry at SUNY, Syracuse University, January, 1969.

Merchant, D, C., A. Schenk, A. Habib, T. Yoon, (2004), “USGS/OSU Progress with Digital Camera In Situ Calibration Methods”, XXth Congress, International Society for Photogrammetry and Remote Sensing, Istanbul, Turkey, July 2004

Merchant, D., A Schenk, A. Habib, T. Yoon, M. Ghanma, (2004a), “Camera Calibration Development for IN SITU Applications”, Final Report for U S Geological Survey, March 23, 2004

Topo Photo, Inc. (1995), “Testing Calibrations in Application to Airborne GPS Controlled Photogrammetry”, Final Report for USGS, National Mapping Division, April 25, 1995

Washer, F. E. (1956), “Sources of Error in Various Methods of Airplane Camera Calibration”, Photogrammetric Engineering, September, 1956

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download