Image calibration and analysis toolbox - a free software suite for objectively measuring reflectance, colour and pattern

Abstract

Quantitative measurements of colour, pattern and morphology are vital to a growing range of disciplines. Digital cameras are readily available and already widely used for making these measurements, having numerous advantages over other techniques, such as spectrometry. However, off-the-shelf consumer cameras are designed to produce images for human viewing, meaning that their uncalibrated photographs cannot be used for making reliable, quantitative measurements. Many studies still fail to appreciate this, and of those scientists who are aware of such issues, many are hindered by a lack of usable tools for making objective measurements from photographs.We have developed an image processing toolbox that generates images that are linear with respect to radiance from the RAW files of numerous camera brands and can combine image channels from multispectral cameras, including additional ultraviolet photographs. Images are then normalised using one or more grey standards to control for lighting conditions. This enables objective measures of reflectance and colour using a wide range of consumer cameras. Furthermore, if the camera's spectral sensitivities are known, the software can convert images to correspond to the visual system (cone-catch values) of a wide range of animals, enabling human and non-human visual systems to be modelled. The toolbox also provides image analysis tools that can extract luminance (lightness), colour and pattern information. Furthermore, all processing is performed on 32-bit floating point images rather than commonly used 8-bit images. This increases precision and reduces the likelihood of data loss through rounding error or saturation of pixels, while also facilitating the measurement of objects with shiny or fluorescent properties.All cameras tested using this software were found to demonstrate a linear response within each image and across a range of exposure times. Cone-catch mapping functions were highly robust, converting images to several animal visual systems and yielding data that agreed closely with spectrometer-based estimates.Our imaging toolbox is freely available as an addition to the open source ImageJ software. We believe that it will considerably enhance the appropriate use of digital cameras across multiple areas of biology, in particular researchers aiming to quantify animal and plant visual signals.

Keywords: animal coloration; camera calibration; colour measurement; colour vision; communication; cone‐catch quanta; image processing; pattern analysis; signalling; spectrometer.

Figure 1

Multispectral image preparation example. This diagram outlines the important steps taken in order to create a normalised, aligned stack from a scene photographed in both visible and UV light. Quantitative measurements can be made from the resulting image, or it can be converted to animal cone‐catch quanta if the camera's sensitivity functions are known. The final step shows regions of interest being selected. The user can select whether the regions labelled ‘p’ for petal in this instance should all be measured individually, or whether they should be combined prior to automated measurement, enabling easy within and/or between region and photograph measurements. Channels are named with the lower case prefix denoting the filter type (e.g. ‘v’ for visible and ‘u’ for UV), and the suffix the camera's channel (e.g. R, G or B). Many more combinations can be created with additional filters.

Figure 2

Example of camera linear responses from a Nikon D7000. Expected reflectance values normalised between exposures (x‐axis) are plotted against mean observed pixel values measured for each standard (y‐axis). This normalisation allows us to test for within and between photograph linearity across a wide range of apertures and shutter speeds while preserving the actual pixel value on the y‐axis, allowing us to detect any hardware‐based nonlinearity at specific pixel values. Error bars show mean ±1 standard deviation; point brightness is scaled with aperture.

Figure 3

Spectral sensitivities of duplicate camera set‐ups. Solid lines and dashed lines show the spectral sensitivities of two different set‐ups that use the same equipment and – in the case of the D7000s – have undergone the same full‐spectrum conversion. The sensitivities are almost identical between all set‐ups, suggesting that these spectral sensitivity functions could be used for other identical set‐ups.

Figure 4

Colour sample measurements. This set of 48 colour pastels was used to compare camera and spectroradiometer measurements of cone‐catch quanta. Image ‘a’ is a normal human‐visible photograph, and image ‘b’ is a false colour combination of green, blue and UV channels (vG, vB and uR). Both images were square‐root‐transformed from 32‐bit linear, normalised images to display correctly on low‐dynamic range media. Standard colour charts have poor UV reflectance presumably to reduce colour fading with age, while these pastels have varied reflectance spectra in UV. For example, the pastels that appear blue in image ‘b’ have UV peaks relative to green and blue.

Figure 5

Blue tit cone‐catch quanta estimates for a range of pastel colours, comparing spectrometer and camera estimates. Cone‐catch estimates from a Canon 7D with two filters and five channels (i.e. vR, vG, vB, uB and uR) are shown in the left column, and estimates from the same camera with four filters and seven channels are shown in the right column (i.e. rR, gR, gG, bG, bB, uB and uR). Shaded areas show standard error. Cone‐catch estimates with two filters are exceptionally good; however, the additional channels provide some greater spectral partitioning in the medium‐wave range, improving the quality of fit most for this channel. Two of the pastel colours were fluorescent – absorbing short wavelengths and emitting them in the long‐wave range. These two points highlight the ability of the toolbox to work with values outside of the 0–100% reflectance range.