Spectral dataset of natural objects' reflectance from the Southern cone of South America

Abstract

The reflection in natural objects mediates an important fraction of the light reaching animal photoreceptors. Knowledge of the spectral properties of natural objects is increasingly valuable for different research fields. Measured datasets of natural objects' reflectance can offer insights into fundamental and applied research questions, contributing to investigations from coloration and color vision to color analysis and representation. Thus, datasets of natural objects' reflectance across different locations are crucial to assessing the universality and variability of physical visual inputs in diverse environments. However, the Southern Hemisphere is notably underrepresented in publicly available datasets of natural objects. To address this gap, we present a spectral dataset of natural objects' reflectance from the Southern cone of South America, specifically Northwestern Argentina. Our dataset encompasses 532 samples representing diverse natural objects such as barks, flowers, fruits, leaves, plant fruits, stones, and animal specimens, including birds, beetles, and butterflies. By openly sharing this dataset, as a publicly available online resource, we aim to facilitate research across various disciplines, from evolutionary biology to industrial applications.

Fig. 1

Experimental setup for the acquisition of reflectance data from natural objects. (a) Lateral view of the illumination cabin and the spectroradiometer, (b) Top view, depicting the variable distance (with two possibilities, d1 = 70 cm and d2 = 40 cm) between the spectroradiometer and the sample.

Fig. 2

Experimental setup top view for birds and butterflies exhibiting iridescent effects on part or all of their body surface. Due to the sample size, in these cases, the distance between the spectroradiometer and the sample was fixed at 70 cm.

Fig. 3

Examples of natural objects and their corresponding spectral reflectance data collection. (a) Specimen of Citrus aurantifolia, an object with a relatively simple coloration pattern. (b) Citrus aurantifolia viewed through the SpectraScan PR-715 viewer. The white outline delineates the patch to be measured, and the black and gray dots (M1, M2, and M3) represent different measurements at the same patch. (c) Specimen of Viola wittrockiana, an object with a relatively complex coloration pattern. (d–f) Viola wittrockiana viewed through the SpectraScan PR-715 viewer. The ellipses (P1, P2, and P3) delineate the patches (white, light purple, and dark purple, respectively) to be measured, and the black and gray dots (M1, M2, and M3) represent different measurements at each patch.

Fig. 4

Workflow sequence for spectral reflectance data collection of natural objects.

Fig. 5

Examples of three general cases of file name structure (when different identity information was available, (a–c) and of the four exceptions to those general rules (d–g). (a) The case of plants and birds, for which we had a taxonomic identification at the species level. (b) The case of butterflies and beetles, for which we had a taxonomic identification at the family level. (c) The case of stones, using a general identification of the rock or the mineral. (d) File name structure for objects that were collected and measured more than one time (i.e., repeated objects). (d) File name structure for leaves that corresponded to the same plant species but with flowers of different colors. (f) File name structure for butterflies and birds that had iridescent colors. (g) File name structure for bird species that were represented for both male and female specimens.

Fig. 6

(a) Instrument error in arbitrary units for the SpectraScan PR-715 spectroradiometer. (b) PTFE reflectance as given by the manufacturer (OPTRONIC® Laboratories).

Fig. 7

The area under the curve (AUC) values of samples’ spectra organized by category. Each panel contains the values for each sequential measurement. We found no significant differences among the three measurements [All (mixed-effects): F(1.61, 839) = 1.32, p = 0.27; Barks (rm-anova): F(1.61, 30.63) = 0.57, p = 0.54; Beetles (rm-anova): F(1.94, 15.5) = 1.17, p = 0.34; Birds (rm-anova): F(1.66, 172.3) = 0.13, p = 0.84; Butterflies (rm-anova): F(1.37, 71.47) = 1.38, p = 0.25; Flowers (mixed-effects): F(1.69, 180) = 0.93, p = 0.38; Fruits (mixed-effects): F(1.18, 33.1) = 3.89, p = 0.052; Leaves (mixed-effects): F(1.17, 131) = 0.35, p = 0.58; Plant fruits (rm-anova): F(1.69, 23.6) = 1.9, p = 0.18; Stones (rm-anova): F(1.38, 20.7) = 1.51, p = 0.24; Vegetables (mixed-effects): F(1.13, 8.49) = 0.2, p = 0.7].

Fig. 8

(a) Distribution of CIE1931 xy chromaticity of all reflectance samples collected in Argentina. Chromaticities were computed under the equal energy white light. (b) Chromaticity distribution of reflectance samples in Japan. In both panels, the black plus shows the chromaticity of the equal energy white, and the magenta line shows the CIE daylight locus from 4,500 K to 20,000 K. Upper-right inserted figures show the chromaticity gamut and the filled region shows the overlapping region with the gamut of the other country.