Color-Infrared Kite
Aerial Photography

J.S. and S.W. Aber
(2019)

Table of contents
Introduction Film CIR
Digital CIR Comparison
References

Introduction

Human color vision is among our most important natural senses. The human eye is receptive to three primary colors, namely blue, green, and red, which in various combinations make up all possible colors. The visible range of light wavelengths is approximately 0.4 Ám (400 nm) to 0.7 Ám (700 nm). The near-infrared (NIR) lies just beyond human vision in the range 0.7 to 1.5 Ám.

Spectrum of ultraviolet (UV), visible (VIS), and near-infrared (IR) radiation. Primary visible colors are blue, green, and red. Wavelengths given in nanometers (nm). Image adapted from Wikimedia Commons.

Note: this discussion deals only with near-infrared. Thermal (heat) infrared is much longer wavelengths and requires completely different imaging techniques.

Film infrared photography

The first near-infrared photographs were taken and published by R.W. Wood in 1910 (Finney 2007). He discovered that active, green vegetation is strongly reflective for NIR, and this phenomenon is known as the "Wood Effect." By the 1930s, black-and-white infrared film was perfected and made available commercially. The need for aerial reconnaissance in World War II led to development of color-infrared film designed for camouflage detection. Since then, color-infrared photography has been applied for many types of aerial and space imagery dealing with vegetation, water, soils, and other natural resources.

Traditional color-infrared (CIR) photography is based on green, red, and NIR layers in the film emusion, and blue/UV light is excluded by use of a yellow filter. As developed into a visible picture, green is shifted to blue, red becomes green, and NIR appears red. This shifting of spectral bands is known as false-color imagery.

Analog (film) photographs of the Emporia State University campus, Kansas. Color-visible (A) and color-infrared (B) show dormitory buildings, parking lots, trees, and part of the football stadium. Active vegetation appears in pink, red, and maroon colors in the CIR image. Images acquired with a dual-camera rig.

During the late 1990s we developed CIR kite aerial photography based on film cameras. This required considerable experimentation and testing, as CIR film carries no ISO rating and camera light meters do not measure near-infrared. By the turn of this century, we had determined appropriate camera light settings based on estimates of infrared radiation in the landscape (Aber et al. 2001). We utilized this approach extensively across the United States and overseas in Estonia. However, CIR film became increasingly expensive, most photo labs stopped processing this film, and we gave up analog CIR kite aerial photography in 2004.

CIR view over the Nature Conservancy marsh at Cheyenne Bottoms, Kansas. Black dots are cattle grazing on wet meadow. Ektachrome EIR 35-mm film taken with an SLR camera and yellow filter. Our last analog color-infrared KAP, July 2004.

Digital color-infrared photography

As film-based CIR photography phased out, digital CIR photography became more attractive in the first decade of this century. All digital camera are sensitive to NIR and, so, are equiped with an infrared-blocking filter. There are two ways to overcome this situation.

  1. Specially built cameras and lenses designed specifically for CIR imagery. Such cameras are available for applications in agriculture and environmental studies. These cameras tend to be rather expensive and may not be easily adapted for KAP. See Tetracam ADC (Aber et al. 2009).

    Overview of the Rowley River (right) and its tributaries in the saltwater marsh complex, Plum Island Ecosystems Long Term Ecological Research site, Massachusetts (August 2009). Active vegetation appears in pink, red, and maroon colors, and water is black. Note the strongly curved horizon and slightly fuzzy appearance of the image.

    Starting in 2008, we tested the Tetracam ADC at several sites in Kansas, Pennsylvania, and Massachusetts. Because of its relatively high cost we flew it only under ideal conditions, but the image quality was disappointing, so we quit using this camera the following year.

  2. Certain consumer cameras may be modified for CIR imagery. The infrared-blocking filter is removed, and a filter is inserted that passes blue, green, and near-infrared wavelengths. The blue and green channels are recorded normally, and near-infrared is shifted to the red band. This results in B/G/NIR false-color imagery. One such camera is the Sony Alpha 6000, a mirrorless interchangeable-lens camera (Aber et al. 2018).

    Color-visible (left) and color-infrared (right) views of a pond surrounded by trees and grass, Peter Pan Park, Emporia, Kansas. Active vegetation appears in shades of orange, and water is dark blue. Deciduous trees and grass are bright orange, and conifers are dark orange.

Comparison of CIR false-color formats

Traditional CIR photos depict active vegetation in red, pink, and maroon colors. This has been the norm since World War II based on G/R/NIR images, and generations of photointerpreters have learned to visualize and analyze this false-color format. The false-color of vegetation is explained by reflectivity of active, emergent vegetation. Blue and red wavelengths are absorbed by leaves for photosynthesis, green is reflected weakly (what we see), and near-infrared is reflected strongly.

Spectrum of near-UV, visible (B/G/R), and near-IR sunlight reflected from common objects. Note sugar beet, grass, and tree curves showing weak green reflection, blue and red absorption, and strong NIR reflection. Dead vegetation or fallow fields, however, do not have this spectral signature. Adapted from Short (1982, fig. 3-5B).

For traditional G/R/NIR false-color imagery, the strong NIR reflection appears in red, and the weak green reflection is shifted to blue. Red + blue = maroon in the resulting image. B/G/NIR imagery, on the other hand, depicts the strong NIR reflection in red, and the weak green reflection remains green. Red + green = orange in the resulting image.

False-color combinations for G/R/NIR and B/G/NIR imagery and the typical appearances of active vegetation and clear water.
Type of imagery False color Vegetation Clear water
G/R/NIR
G = blue
R = green
NIR = red
Red + blue
= maroon
Blue excluded
= near black
B/G/NIR
B = blue
G = green
NIR = red
Red + green
= orange
Blue allowed
= dark blue
Based on Aber et al. (2018).

At first glance, the orange color of vegetation appears almost bizarre, but it is no more strange than the pink-maroon color of vegetation in traditional CIR imagery. G/R/NIR color-infrared imagery has been the norm for conventional aerial photography and space-based imagery. This is because blue light is strongly scattered in the atmosphere. Blue scattering is minimal for low-height KAP, however, so the B/G/NIR format may be utilized.

Color-infrared photos tend to enhance special lighting effects, for example sun glint and the hot spot. Sun glint is a mirror-like reflection from water bodies, metal and glass structures, etc. The hot spot is the position on the ground at the antisolar point in direct alignment with the camera and sun. It appears bright because of shadow hiding (Hapke et al. 1996).

Sun glint (left) from a lake surface shows wave pattern clearly. A hot spot (right) appears left of scene center in prairie vegetation. Hot spots are most noticeable over relatively homogeneous ground cover. Both photos taken with the modified Sony camera.

Taking color-infrared photographs with a consumer-grade camera modified for B/G/NIR imagery does introduce some artifacts. The light meter and automatic exposure setting are intended only for visible light and may not respond appropriately for visbile plus near-infrared. Likewise, lenses are designed for visible light; the longer wavelengths of near-infrared may not refract or focus properly (Finney 2007).

In oblique views looking toward the sun, bright pink-white streaks may appear in some images taken with the modified Sony camera. This is presumably due to internal lens reflections.

References


Text and images © J.S. and S.W. Aber

Return to KAP cameras or KAP home.
Last update: February 2019.