Challenge of Infrared Kite Aerial Photography

James S. Aber¹, Susan W. Aber¹ and Brooks Leffler²

Earth Science Department, Emporia State University, Emporia, KS 66801. E-mail to: aberjame@emporia.edu.
P.O. Box 34, Pacific Grove, CA 93950.

This article is published in the Transactions of the Kansas
Academy of Science, vol. 104, no. 1/2, p. 18-27 (2001).

Table of Contents
Introduction Infrared KAP challenge
Infrared KAP rigs Field testing of infrared KAP
Applications of infrared KAP Conclusions
References Related sites

**Table of Contents**
Introduction	Infrared KAP challenge
Infrared KAP rigs	Field testing of infrared KAP
Applications of infrared KAP	Conclusions
References	Related sites

ABSTRACT

Kite aerial photography (KAP) involves the use of large kites to lift compact camera systems for capturing images of ground features. Our goal has been to develop equipment and techniques for infrared kite aerial photography. Attempting to conduct infrared photography from kites has proven to be a considerable challenge for several reasons. Among the most important limitations are those inherent in infrared film and conventional cameras. Infrared film has no ISO speed rating, and camera light meters do not measure infrared energy in the scene. Furthermore, KAP equipment should be relatively compact, rugged, light weight, and inexpensive.

We have designed, built, and field tested two KAP rigs based on the Canon EOS RebelX camera, which is a 35-mm, full-featured, SLR camera. Our first rig is a dual-camera setup, in which the cameras take simultaneous color-visible and color-infrared photographs. This rig has radio control of the camera shutters. Camera tilt and pan positions are set manually prior to each flight. The second rig is for a single camera (color-infrared only); the rig has full radio control for pan, tilt, and shutter. The dual-camera rig weighs 1.5 kg, and the single-camera rig is just over 1 kg. Based on considerable field testing, we have determined appropriate manual light settings under full-sun conditions--1/250 shutter speed and f-9.5 aperture. Our preliminary results suggest that infrared KAP may be valuable for detailed investigations of vegetation, soils, and other environmental features.

INTRODUCTION

Kite aerial photography (KAP) involves the use of large kites to lift compact camera systems for capturing images of ground features. Aber, Harrington, and Nowak (1999) and Aber and others (1999) have reported previously on forestry and geomorphic applications of kite aerial photography in Kansas. KAP has been employed for various scientific investigations in other localities including Arctic Canada (Bigras 1997), Antarctica (Carlson 1997), Arctic Russia (Gawronski and Boyarsky 1997), and Poland (Aber and Galazka 2000).

Kite aerial photographs are typically taken at heights of 50-150 m above the ground, and radio controls are often utilized to operate camera rigs while in flight. KAP is capable of producing high-resolution, large-scale imagery in which dm-sized details are visible for ground features. A great variety of conventional film, video, and digital cameras may be used in many different configurations for KAP. Nearly all imaging systems now in use, however, are based on visible light of the color spectrum (0.4 to 0.7 ľm wavelength).

Aerial infrared photography has long been practiced from manned airplanes and other large platforms, including the space shuttle and unmanned satellites (McDonald 1997). Special photographic films are sensitive to the near-infrared (NIR) portion of the spectrum (0.7 to 0.9 ľm wavelength). Originally employed for camouflage detection, NIR photography has proven especially valuable for environmental studies of surficial vegetation, water resources, soils, and other natural or cultural features (Avery and Berlin 1992).

Infrared film comes in two basic varieties--black-and-white and color-infrared. These films are sensitive to ultraviolet, visible and NIR radiation. Filters are used normally to block ultraviolet and shorter visible (blue) wavelengths from reaching the film. Special camera systems are required for conventional infrared aerial/space photography. Note: photographic film is not sensitive to mid-infrared or thermal-infrared (heat) radiation.

Marzolff and Ries (1997) have developed an infrared photographic system for an unmanned hot-air blimp, which they use for geomorphic investigations. The blimp holds 100 mł of hot air, which is kept warm by periodic blasts from two propane burners that are suspended in a basket beneath the blimp. Control is maintained by two 500-meter tether lines attached to the blimp's nose. This gives the blimp a practical flying limit of 400 m above the ground, depending on wind and temperature conditions. The blimp has a lifting capacity of around 6 kg for the camera and radio-control apparatus. The photographic system employs two, vertically mounted Pentax 35-mm cameras--one for color-visible and the other for color-infrared--for dual photographs of the ground. The success of the blimp system inspired us to adopt these techniques for kite aerial photography.

INFRARED KAP CHALLENGE

Attempting to conduct infrared photography from kites has proven to be a considerable challenge for several reasons. KAP techniques demand relatively small, rugged, and light-weight components to be practical. Furthermore, KAP is normally conducted with a much lower budget compared to conventional aerial photography from manned platforms or to the unmanned hot-air blimp of Marzolff and Ries (1997). These conditions place severe constraints on the types of camera systems that may be utilized for KAP. Among the most important limitations are those inherent in infrared film and conventional cameras.

Infrared film does not carry an ISO speed rating. Film speed is designated only for the visible portion of the spectrum and so does not apply to infrared-sensitive film.
Camera light meters do not measure the amount of NIR radiation present in the scene, so light values and automatic camera settings are useless.

The amount of near-infrared radiation in a scene may vary substantially depending on many factors such as sun angle, amount of photosynthetically active vegetation, presence of water bodies, and camera orientation. Time of day and seasonal conditions also have a strong influence on the amount of solar NIR energy that is reflected from the ground toward the camera. These factors mean that infrared photography is a trial-and-error undertaking in which appropriate lighting conditions and camera settings (shutter speed, f-stop) must be determined through experience. Furthermore, each camera/lens/filter system behaves differently in terms of infrared radiation. Thus, a particular camera/lens/filter combination must be tested under various field conditions in order to arrive at empirical results for best settings.

Marzolff (pers. comm. 1998) has tested various exposure settings for color-infrared film for hot-air blimp aerial photography with a Pentax camera and 50-mm lens. She found that under conditions of bright sunlight, color-infrared film could be treated as ISO 200 speed for vertical, low-height airphotos. Her empirical results are presented in Table 1.

**Table 1. Manual compensation for color-infrared film for default value of ISO 100. Based on Marzolff (pers. comm. 1998).**
Lighting conditions	Exposure correction
Very bright sun, mid-day--clean, dry atmosphere	ISO 200 (+1 f-stop)
Light, but not bright sun--hazy, humid, dusty air	ISO 160 (+˝ f-stop)
Slightly overcast, indirect light--thin clouds	ISO 100 (no correction)
Pale, diffuse light--early morning or late afternoon	ISO 80 (-˝ f-stop)
Overcast, indirect, rather dark--heavy clouds	ISO 50 (-1 f-stop)

To be useful for NIR photography, a camera must be capable of manual adjustments for shutter speed and f-stop, and the lens must accept filters. Auto bracketing of exposures is another useful function. Few, if any, point-and-shoot cameras have these capabilities. On this basis, a full-featured, single-lens-reflex (SLR) camera is necessary to attempt infrared kite aerial photography. Digital infrared cameras are available, but at much higher cost ($10,000).

INFRARED KAP RIGS, CAMERA AND FILM

The authors have designed, built, and field tested two KAP rigs for infrared photography. Both rigs employ the Canon EOS RebelX camera, which is a 35-mm, full-featured, SLR camera. This camera was selected because its plastic body makes it quite light. The camera is equiped with a plastic zoom lens (35-80 mm focal length), which is also relatively light. The two rigs are described in more detail as follows.

Dual-camera KAP rig. Two cameras are mounted bottom to bottom and linked to a single shutter release, which is radio controlled. One camera is loaded with color-visible film and the other takes color-infrared film. Given the weight of two cameras, all other components were minimized to save on total rig weight. Thus, pan and tilt positions of the rig are adjusted manually on the ground and cannot be changed while the rig is in flight. This rig is relatively heavy at 1.5 kg (fig. 1).
Single-camera KAP rig. One camera is mounted in a radio-controlled mechanism that allows for horizontal pan (0-360°) and vertical tilt (depression angle 15° to vertical). The shutter is also triggered by radio control. Manual light settings for color-infrared photography are adjusted on the ground prior to each flight. This rig weighs just over 1 kg (fig. 2).

Figure 1. Dual-camera rig for kite aerial photography. A (left) - camera rig supported by a cable-and-pulley suspension system that is attached to the kite line. B (right) - closeup view; one camera carries color-visible film, the other takes color-infrared film. Cameras are positioned vertically in these views. Click on the small images to see full-sized versions.

Figure 2. Closeup view of the single-camera rig for color-infrared photography. Radio-controlled servos operate camera pan, tilt, and shutter. Click on the small image to see a full-sized version.

The radio controls are standard model-airplane components that are modified to activate servos on the KAP rigs, and rechargeable NiCd batteries are utilized. After, working with several types of film for color-visible photography, we have come to rely mainly on Kodak Ektachrome and Kodachrome 200 slide films. They have excellent, near-natural reproduction of colors. Kodak Ektachrome EIR film is the only color-infrared slide film that is readily available (in the U.S.) in 35-mm format. It can be found or ordered from camera and film specialty stores. For color-infrared photography, we employ a yellow (Kodak Y2) filter to eliminate ultraviolet and blue light; a conventional UV-haze filter is used for color-visible photographs.

In the field, Ektachrome EIR film can be treated much like conventional slide film. Both should be protected from excessive heat (> 95°F or 35°C) and direct sunlight. All Ektachrome films can be processed with the E-6 method, which is available in most photo stores; turn-around time may be as little as one day. Kodachrome requires special processing that is done in only a few photo labs, and turn-around time is usually a week or more (for mailing).

FIELD TESTING OF INFRARED KAP RIGS

Beginning in 1999, we have conducted numerous field tests of these two KAP rigs for color-infrared photography. Flying rigs that weigh 1-1˝ kg generally requires a kite of 2.8 to 3.3 m² (30-36 ft²) lifting surface area for moderate wind speed of 16-24 km/h (10-15 mph). Smaller kites can be used with stronger wind. For lighter wind, we use a pair of two large kites flying from the same main line. Our experience with each rig is described in more detail below.

Dual-camera rig. The primary advantage of this system is acquisition of paired color-visible and color-infrared photographs (non stereo). The Canon Rebel cameras have sensitive shutter controls, so it has proved difficult to obtain simultaneous exposures with this rig. Sometimes one shutter will trigger and the other does not. In other cases, both shutters trigger for multiple exposures. With practice, we have gained the right "touch" on radio controls to avoid this problem most of the time.
A major drawback to operation is that camera tilt and pan are fixed for each flight. In order to change the camera position, it is necessary to bring down the KAP rig and then relaunch the rig. It is also necessary to bring the rig down in order to change the manual light settings for the infrared camera. Repeated launch and retrieval of the KAP rig is time consuming and quite tiring for kite handlers.
Single-camera rig. The primary advantage of this system is radio-control of camera pan and tilt positions while the rig is in flight. The shutter is more robust than for the dual-camera rig. It is still necessary to bring the KAP rig down in order to change manual light settings for infrared photography. In general, the lighter weight of this rig makes it somewhat easier to handle during launch, flight, and retrieval. The main disadvantage is that only one kind of film can be used per flight. Thus, acquiring a pair of color-visible and color-infrared images of the same view is nearly impossible.

We have experimented with various camera light settings under different background lighting conditions ranging from full sun, to hazy, to partly overcast sky. We began by using the empirical results of Marzolff (see Tab. 1), but found that these light settings resulted in underexposed color-infrared photographs. We surmise that the Canon zoom lens absorbs more light energy than does the Pentax fixed-length lens of Marzolff and Ries (1997).

The results for our Canon cameras suggest that color-infrared film can be treated as ISO 100 speed for full-sun conditions. We have obtained best results with camera settings of 1/250 shutter speed and f-11 to f-8 aperture (fig. 3). We now routinely utilize a shutter speed of 1/250 and f-9.5 for all color-infrared photography. These settings have produced uniformly good results under full-sun, mid-day conditions; we rarely conduct infrared KAP under other lighting conditions.

Figure 3. Color-visible (A - above) and color-infrared (B - below). Oblique view, Emporia State University campus. A portion of the football field, dormitories, parking lots, cars, grass, and deciduous trees. Photosynthetically active vegetation is relatively dark (green) in the visible image and bright (red, pink) in the infrared image. Taken under full-sun conditions, shutter speed 1/250, and aperture f-8. September, 1999. Click on the small images to see full-sized versions.

APPLICATIONS OF INFRARED KAP

After more than a year of designing, building, and testing KAP rigs for infrared photography, we have achieved success in acquiring scientifically useful images. Our primary application is for investigations of forest canopy conditions in northeastern Kansas. In this regard, our objective is to document the meter-scale heterogeneity of forest canopy. This resolution is considerably higher than Landsat Thematic Mapper (TM) imagery, which is nominally 30 m x 30 m. Mature hardwood forest displays a rough canopy, at the scale of meters, which contains many shadows and variations in brightness (fig. 4).

Figure 4. Color-visible (A - above) and color-infrared (B - below). Oblique view, Falley Boy Scout Reservation, near Oskaloosa, northeastern Kansas. Deciduous oak-hickory forest with lake in the left background; small clearing for an archery range, shower house, and septic lagoon. Photosynthetically active vegetation is relatively dark (green) in the visible image and bright (red, pink) in the infrared image. Taken under partly overcast conditions, shutter speed 1/250, and aperture f-11. June, 1999. Click on the small images to see full-sized versions.

The normalized difference vegetation index (NDVI) and other vegetation indices are derived from Landsat and other satellite imagery. These indices are based mainly on the red to infrared ratio and are utilized widely for regional and global measurements of vegetation cover and vigor (Lymburner et al. 2000). The coarse resolution of such data sources means that meter-scale variations in forest canopy are lost. Our intention is to determine the degree to which variations in canopy roughness may affect satellite-derived NDVI values. Our preliminary results suggest that rough canopies may depress NDVI values in comparison to forests with smooth canopies. Further study is underway to confirm this situation. Infrared KAP could have many other applications for detailed environmental studies of vegetation, soils, water bodies, and other surficial phenomena.

CONCLUSIONS

Conducting infrared aerial photography from kites presents several technical challenges. The successful KAP rig must be relatively light-weight, compact, rugged, and carry a camera that is capable of acquiring infrared photographs. We have developed two KAP rigs that meet this challenge. A dual-camera rig is used to acquire pairs of color-visible and color-infrared images. Its use is limited by its relatively heavy weight (1.5 kg) and lack of pan/tilt controls while in flight. A single-camera rig has full radio control of camera pan, tilt, and shutter while in flight. Its lighter weight (1 kg) makes it easier to handle during flight. Many field trials have been conducted to arrive at empirical results for best camera settings for color-infrared photographs--1/250 shutter speed and f-11 to f-8 aperture. Infrared kite aerial photography appears to hold promise for detailed investigations of forest canopy as well as other types of environmental studies.

Return to beginning.

REFERENCES

Aber, J.S., and D. Galazka. 2000. Potential of kite aerial photography for Quaternary investigations in Poland. Geological Quarterly 44, p. 33-38.

Aber, J.S., J. Harrington, and M.C. Nowak. 1999. Kite aerial photography for forest assessment at Fort Leavenworth, Kansas. Proceedings--From image to information, p. 484-490 (on CD-ROM). American Society of Photogrammetry and Remote Sensing.

Aber, J.S., R. Sobieski, D.A. Distler, and M.C. Nowak. 1999. Kite aerial photography for environmental site investigations in Kansas. Kansas Academy Science, Transactions 102(1-2):57-67. See abstract.
Avery, T.E., and G.L. Berlin. 1992. Fundamentals of remote sensing and airphoto interpretation. Macmillan Publishing, New York, 472 p.
Bigras, C. 1997. Kite aerial photography of the Axel Heiberg Island fossil forest. American Society of Photogrammetry and Remote Sensing, Proceedings of the First North American Symposium on Small Format Aerial Photography, p. 147-153.
Carlson, J. 1997. Kiteflying in the freezer. The Aerial Eye 3, p. 6-7.
Gawronski, J.H.G., and P.V. Boyarsky, eds., 1997. Northbound with Barents: Russian-Dutch integrated archaeologic research on the Archipelago Novaya Zemlya. Uitgeverij Jan Mets, Amsterdam, 255 p.
Lymburner, L., P.J. Beggs, and C.R. Jacobson. 2000. Estimation of canopy-average surface-specific leaf area using Landsat TM data. Photogrametric Engineering & Remote Sensing 66, p. 183-191.
Marzolff, I., and J.B. Ries. 1997. 35-mm photography taken from a hot-air blimp: Monitoring processes of land degradation in northern Spain. The first North American symposium on small format aerial photography, p. 91-101. Edited by Bauer et al., American Society for Photogrammetry and Remote Sensing.
McDonald, R.A. 1997. Potential new applications for declassified early satellite reconnaissance imagery. In McDonald, R.A. (ed.), Corona between the Sun & the Earth: The first NRO reconnaissance eye in space, p. 245-254. American Society of Photogrammetry and Remote Sensing, Bethesda, Maryland.