Home | Contact ST  
Follow ST

Feature Article

Capturing Accurate Colors Underwater With Consumer Digital Cameras

By Derya Akkaynak and Roger T. Hanlon

Comparison of the spectral composition of light underwater (black curves) and on land (red curves). Underwater profiles were measured at specific dive sites at depths shallower than 10 meters in Urla, Turkey; Woods Hole, Massachusetts; and Kona, Hawaii. Spectra are normalized by their values at 560 nanometers for comparison.

In 1926, when Charles Martin and Dr. William Longley captured the first color photograph underwater, they had to use highly explosive magnesium flash powder to provide enough light to the hypersensitized glass plates for an exposure time of one-twentieth of a second. In those days, Autochrome glass plates were the only means of obtaining color photographs with exposure times around 12 seconds.

Before Martin, who was the first chief of National Geographic's Photo Laboratory, developed a hypersensitizing solution to coat the Autochrome plates, it was not reasonable to take color photographs of anything other than still subjects, and it was practically impossible to take photos underwater. But the 2,400-flashbulb-strong explosion provided enough light at 4 meters below the sea surface that the first underwater color photograph—that of a hogfish—was successfully captured in the Dry Tortugas.

In 2012, there is no need to discharge a pound of magnesium flash powder to provide light underwater. The sensors of our commercial off-the-shelf (COTS) digital cameras are sensitive enough that detailed, colorful photographs can be captured using just the ambient daylight at shallow waters. For deeper and darker waters, there are many configurations of strobe systems that can be purchased, or tailored one-of-a-kind designs can be contracted to companies for any set of requirements. Similarly, watertight housings are available for many makes and models of cameras, and new ones are being designed continually as new camera models are introduced to the public.

While digital cameras are highly advanced, they do not capture the world in a standard way so their outputs must be calibrated to obtain device-independent color. For scientific or industrial purposes, it is critical that photographs contain accurate colors that can be reproduced by others with no knowledge of the original imaging system.

Underwater Color Calibration
Every COTS digital camera captures images in its own RGB (red, green and blue) color space. Transforming these images to a device-independent color space is done through a transformation matrix. This matrix, however, is not unique, and the accuracy of the resulting colors in the image will depend on the colors chosen to obtain the matrix. For land photography, the transformation from the camera color space to human color space is usually done through the use of photographic calibration targets, such as a Macbeth ColorChecker. For underwater applications, the use of targets designed for land photography is not appropriate for two reasons. First, these targets contain color patches selected to represent the range of colors humans see on land on a daily basis and, thus, are more diverse and saturated than the colors found in most underwater habitats. Second, the transformation process for land-based photos frequently uses the spectra of standard illuminants, which do not represent underwater ambient light fields well, such as the International Commission on Illumination (CIE) D-series of illuminants for approximating the natural daylight spectra (D65 is used as the baseline in many applications).

One solution is to create habitat-specific color charts, or habitat charts. This requires collecting reflectance data from a variety of substrates found in a given underwater habitat with a spectrometer, as well as representative light profiles. The ideal way to collect spectral data in an underwater scene is by using a hyperspectral imager, which records a spectrum for every pixel in an image within a certain range of the electromagnetic spectrum. However, hyperspectral imagers are expensive and therefore not available to many research laboratories.

In their absence, spectrometers can be used. This way, images captured by COTS digital cameras can be transformed from the camera space to a device-independent space (such as standard RGB) using colors and light spectra that are representative of that particular habitat. Each habitat has different optical properties of water and varying colors amid the plants, animals and abiotic entities of each substrate. Therefore, unique habitat charts should be built for each dive site.

Raw Images
When the intensity of pixels in an image constitutes scientific data, the analysis should start with raw images rather than the post-processed files (e.g., JPG or TIFF). Post-processed files will not only be irreversibly altered in features like white-balance, color and contrast but may also be compressed in a lossy fashion. Such changes that happen without user control cannot be quantified and will compromise data quality.

Every COTS digital camera has its own RGB color space because each sensor has unique spectral sensitivities. This difference is usually substantial between different makes of cameras and between different models of the same make. Even two copies of the same camera make and model produced at the same factory are likely to have slightly different spectral sensitivity curves.

Transforming raw images recorded by COTS digital cameras to a device-independent color space is usually done through a three-by-three transformation matrix that relates the camera color space to the human color space. While cameras have built-in software that perform this transformation, they do so without the knowledge of the colors in the scene, the ambient light conditions or the specific sensor in the camera. To continue this article please click here.

Derya Akkaynak is a Ph.D. candidate at the MIT WHOI Joint Program in Applied Ocean Science & Engineering. She holds bachelor's and master's degrees in aerospace engineering from Middle East Technical University and the Massachusetts Institute of Technology, respectively. She is an American Academy of Underwater Sciences diver and Professional Association of Diving Instructors divemaster.

Roger T. Hanlon is a senior scientist at the Marine Biological Laboratory's Marine Resources Center in Woods Hole, Massachusetts. He uses digital imagery and spectrometry to quantify camouflage and signaling in cephalopods and fish. He holds a Ph.D. from the University of Miami and performed postdoctoral research at Cambridge University.

-back to top-

-back to to Features Index-

Sea Technology is read worldwide in more than 110 countries by management, engineers, scientists and technical personnel working in industry, government and educational research institutions. Readers are involved with oceanographic research, fisheries management, offshore oil and gas exploration and production, undersea defense including antisubmarine warfare, ocean mining and commercial diving.