CIE International Commission on Illumination

CRI Colour Rendering Index

LED Light emitting diode

NHMUK Natural History Museum, London, UK

RGB Red, Green, Blue colour space

UV Ultraviolet

VIS visible (for humans)

For this paper, spectral measurements were carried out and documented using a Blue Spec Cube spectroradiometer from JETI (Jena, Germany). This was done in a similar way to that described by Brehm (2017). The focus is on the measurement of light sources for the photography and digitization of insects (and other objects) in natural history museums. Measurements were carried out between 2019 and 2024 in Jena, Berlin and London. For a better understanding and classification of the results, other light sources used in the museums (workrooms, exhibitions) were also measured, as good light seems desirable in all working areas of a museum. The widespread ‘Colour Rendering Index’ (CRI) is used to assess the quality of light. It reaches its maximum of 100 or just below in daylight and daylight-like sources. The CRI index is based on the agreement in the representation of eight or 16 standard colours, correctly referred to as CIE R_a index and CIE R_e index, respectively. I focus on the stricter index with the higher quality standard (i.e. the CIE R_e index with 16 standard colours).

Quantitative measurements were also carried out in all three devices to check the uniformity of the lighting. Uniformity of illumination is of obvious importance; in cinema projectors, for example, the ST 431-1:2006 standard stipulates that the peripheral areas must have at least 75% of the brightness of the centre. If images are to be evaluated quantitatively, all image areas should be evenly illuminated in order to prevent measurement artifacts. For the evaluation, one measurement was taken in the centre and four measurements in the respective corners. A total of nine measurements were carried out in the light hemisphere due to the increase in brightness observed at the edges. Suppl. material 1 shows the measurement points and the measurement results in detail.

Various full-frame mirrorless cameras from Sony are used for photography, but products of similar quality can also be used for the applications described. For photographing individual insects, I use a Sony Alpha 7 camera (24.3 megapixels), which was modified according to the specifications of Prutchi (2017), i.e. the visible pass/IR cut filter directly in front of the sensor was removed by a specialist company. The lens used is an EL Nikkor 80 mm f/5.6 (quartz glass). As this lens cannot be focused by itself, a bellows device is used as an intermediary. The bellows (dating from the 1970s, made by Leitz, Wetzlar) is marked with different levels of magnification. Photographs are taken at the following scales: 1:1, 1:1.25, 1;1.5, 1:1.75, 1:2, 1:2.5 and 1 to 2.74 (maximum range). For even larger moths, the bellows is replaced by an intermediate ring (Leica R Macro Adapter). With this an image scale of 1:4.74 is achieved. Suppl. material 1 provides an overview of the image scales and the light measurements at each stage in graphic form. The distance between the filter wheel (see below) and the object ranges between 9.4 cm (scale 1:1) and 41.0 cm (scale 1:4.74) (details in Table 2). A filter wheel was mounted in front of the lens using an adapter, which can be operated using an electronic control (Lab View) with a Windows tablet. For normal images (VIS range), a visible pass/IR cut filter is again used, which only allows radiation in the range between 400 and 780 nm to pass through. For UV photography, on the other hand, a Baader U filter is used (Prutchi 2017), which only allows UV radiation in the range between approx. 320–400 nm to pass through. The sensitivity of the camera sensor was measured with a monochromator (Optronic Laboratories OL750/421S). This was done in steps of 2 nm in the ranges between 325–399 nm (UV range with a deuterium lamp) and between 400–800 nm (VIS range with a halogen lamp). In the resulting photographs, the sum of the brightness of all pixels in the image was determined. This was done for all three colour channels in the VIS range and only for the red channel in the UV range – as this is by far the most sensitive colour channel. I use a Sony alpha 7R mark IV camera (61.0 megapixels) in combination with a 60 mm Leica Macro Elmarit f/2.8 to photograph whole insect drawers. For boxes lying inside the drawers (size a quarter of the drawer) I use a 100 mm Leica Apo Macro Elmarit f/2.8. Both high-class lenses have a manual focus and were manufactured in the 1980s.

Moths and other insects are photographed from above (also with the other devices), with a camera that can be raised and lowered on a repro stand (Kaiser RTX tripod). The camera used is a modified Sony Alpha 7 camera (24.3 megapixels) (see above). The basic body of the light cylinder was commissioned from a locksmith. It consists of a cylindrical body made of sheet zinc with a door with a diameter of 45 cm and a height of 35 cm (Fig. 1a, b). In the metal lid (thickness 3 mm) there is a circular hole (diameter 8 cm) for the camera and lens. The lid also serves as a heat sink for the lamps. Eight Lumitronix LinearZ 280-26 elements with Nichia daylight LEDs (length 28 cm) were professionally glued to the underside of the metal lid as lighting and are supplied with power from a suitable power supply unit. Unfortunately, this article from Lumitronix is no longer available but there are suitable alternatives (see below). A polyoxymethylene disc (6 mm thick) suspended under the lid at a distance of 5 cm is used for colour-neutral light diffusion (Fig. 1d) in order to increase the proportion of scattered light and avoid cast shadows. Objects can be moved up and down in the cylinder by hand using a laboratory lifting platform (stainless steel) (Fig. 1c). A scale (10 mm) is moved into the image independently of the object (Fig. 1e, f). Focusing is done manually by first bringing the scale into focus. The insect is then moved into focus by rotating the lifting platform. In this method, the objects to be photographed are arranged by size beforehand in order to enable the most efficient and quickest possible workflow. Before photography, the moths are placed on plates made of roughened aluminum (see Box 2: background). These plates contain an insect needle with a short pipette tip attached to its head (Fig. 1c, e). This is filled with wax so that the objects can be easily fixed both with the pointed side of the needle (photograph of the upperside) and with a needle head (photograph of the underside). The purpose of this is to achieve the greatest possible distance from the background so that it remains completely out of focus. Three xenon flash units (two Neewer NW570 and one NW635) are used for UV photography (Fig. 1c), which indirectly illuminate the object via the ceiling. The cover was removed from the units to allow UV radiation to escape during the flash (spectrum: Fig. 4f). Photographs are always taken with the door closed to ensure uniform lighting.

Figure 1. 

Light cylinder for VIS and UV photography: a. Reduced overview; b. Overview with open door (Scale: 10 cm); c. Structure of the lifting platform and three UV flash units in the cylinder. d. Interior view of the lighting on the ceiling; e. Moth with scale; f. Moth underside during photography of the proboscis length (or photography of labels) on polyethylene plate.

The light box for insect drawers (Fig. 2) was built from an aluminum profile system (15 mm thick) into which (previously used) reflective, scattering aluminum plates with a honeycomb-like surface structure were built, as used in conventional lamp construction. The box has external dimensions of 63 × 53 × 83 cm (WxDxH) and is therefore designed for the photography of insect drawers in standard format (51 × 42 cm). The roof consists of a 3 mm thick metal plate, which also serves as a heat sink for the LEDs. As in the light cylinder, Lumitronix LinearZ 280-26 elements (see above) are also used (eight units), which are arranged all around the edges (6 units) and two parallel in between (Fig. 2c). A hole measuring 20 × 16 cm (WxD) was sawn out in the roof. The reproduction tripod standing on top was equipped with a self-made base made of 30 mm thick aluminum profiles and enables photography through this base. The camera used is a Sony Alpha 7R mark IV (61.0 megapixels), which is attached to a repro stand 124 cm above the insect drawer. The insect drawer can be pushed in and taken out from below with the front panel being attached (Fig. 2a). A scale (10 mm) is placed next to the drawer using a small tripod before the picture is taken. The printed inventory number of the drawer is also included.

Figure 2. 

Light box for insect drawers. a. Overview with front panel attached (when photographed); b. Overview without front panel; c. Inside view of the lighting on the ceiling; d. Inside view of the photographed insect drawer with scale holder.

As a cost-effective, lightweight and easy-to-make alternative to the light cylinder, the possibility of working with polystyrene hemispheres was tested. These are available in sizes with a diameter of up to 50 cm as decorative material (Rayher no. 3306400). Polystyrene has good reflective properties, including in the UV range. However, the material is translucent, so that on the one hand light escapes from the hemisphere and on the other hand light can enter from outside. Therefore, the hemisphere was covered with aluminum foil on the outside (Fig. 3). In the middle at the top there is a circular opening in which a metal ring is embedded (Meetory baking ring with a diameter of 8 cm). An LED strip was glued to the outside of this metal ring (Lumitronix Lumiflex 700 pro Sunlike), which is powered by a power supply unit. To make handling insects easier, a door was cut into one side with a hot wire, which can be raised and lowered using a pull rope (Fig. 3a, b). The height was adjusted to the working distance of 36 cm between the camera and the sample using two squared timbers wrapped in aluminum foil (height 7 cm). The light hemisphere was optimized for the photography of ethanol samples. The insects lie in a liquid, so that direct lighting from above leads to strong reflections on the surface (Fig. 8c). The reflections were reduced by covering the direct radiation of the LED strip with aluminum foil (Fig. 3c). The insects were distributed in a rectangular white ceramic bowl (size 25 × 17 cm, Meeden), covered in 70% ethanol and photographed. The area in the bowl in which the photograph is to be taken was marked in the proportions of the sensor (19.5 × 13 cm). As in the light box, a Sony Alpha 7R Mark IV is used as the camera, which is attached to a repro stand (see details above).

Figure 3. 

Light hemisphere. a. Overview with door open (lifted with pull rope); b. Overview with door closed (Scale 10 cm); c. Interior view with metal ring and LEDs; d. Interior view with metal ring hidden to avoid strong reflections on liquid; e. View of the photographed ceramic bowl with insects from ethanol sample; f. Hemispheres in different formats: diameter 30 cm and 50 cm.

Table 1 gives an overview of the measurements carried out in the three museums in Jena, Berlin and London. The complete measurements are documented in Suppl. material 2 (pdf) and Suppl. material 3 (xlsx). Figure 4 shows an overview of typical spectral patterns of different light sources with reference to daylight. LEDs (Fig. 4a–e) are typically characterized by a primary blue peak. Depending on the LED, the emitted spectrum is expanded with the help of certain chemical substances in the LED. In the case of bluish cold white LEDs, the blue peak is very pronounced (Fig. 4d, e), whereas in other LEDs it is moderately to significantly reduced (Fig. 4b, c). Daylight LEDs (Fig. 4a) are most similar to natural light in terms of spectral composition (CIE R_e = 98), while the CIE R_e values of the other LED lamps tested are between 45 and 82. The lowest CIE R_e values are achieved by the Supra Quartz Scanner and the Leica Z16 microscope because the spectrum deviates quite significantly from daylight – in particular due to an overemphasis on blue and an underemphasis on turquoise and red. Flash light could not be measured effectively due to its short pulse time. Instead, a continuous Xenon arc lamp was measured (Fig. 4f). It shows a balanced spectrum with a value of R_e = 87. Devices with fluorescent tubes (mercury vapour) show the typical spectral pattern with some clearly recognizable characteristic peaks that are typical of mercury vapour lamps (Brehm 2017). Here, too, an almost continuous spectrum is achieved by adding certain chemical compounds (Fig. 4g, h). The CIE R_e values for the tested devices for insect photography reach 65 to 84. The resulting photographs reflect the differences in quality of the light used (Fig. 5); in particular, the photograph taken in the SatScan device is significantly inferior in quality to the other examples.

Figure 4. 

Overview of typical spectral patterns of different light sources with reference to daylight (coloured). a–e. LED light sources; f. Xenon arc lamp; g, h. Mercury vapour lamps. The R_e value is given in each case, which is > 98 in daylight. See Suppl. materials 2, 3 for detailed results.

The light cylinder ensures shadow-free illumination in the VIS range (Fig. 5a–d) with very high light quality. In the UV images, slight shadows can be seen due to the use of three flash units, especially in very large moths. With different image scales, the amount of incident light and also the distribution within the image section change. Very homogeneous light conditions are achieved at all image scales. The largest deviations (3.7% for irradiance and 2.7% in relation to the brightness of the pixels) were achieved at an image scale of 1:2. At larger image scales up to 1:1, the image section is smaller, which leads to greater homogeneity (only 1.1% for irradiance and 1.0% in relation to the brightness of the pixels). At smaller image scales, the distance to the light source increases, and with it the proportion of light scattered in the light cylinder, so that very small deviations are measured here too (2.3% for irradiance and 1.7% in relation to the brightness of the pixels). Detailed results are documented in Suppl. material 1. In routine operation, the system makes it possible to photograph around 20–25 specimens per hour (VIS+UV). Naming, further processing of the image files and data management, require additional working time.

Figure 5. 

Comparison of photographs taken in with different light sources. The images were standardized in terms of white balance and exposure (white reflectance standard set to RGB 241/241/241). They were saved in tif format but otherwise not modified. a. Shot in daylight (overcast sky); b. Shot in light box; c. Shot with SatScan (device in Berlin); d. Shot with Xenon flash. Note that monitors cannot reproduce all stored colour information of the Adobe RGB colour space and general limitations when reproducing images on displays and in print. See Suppl. material 4 for more detailed results and corresponding colour charts.

The light box enables evenly bright and shadow-free illumination in the insect drawer (Fig. 7) with very high light quality (R_e = 97, see above). The homogeneity of the lighting is very high overall – the maximum deviation of the five measured irradiance values is 4.7%. In routine operation, the system makes it possible to photograph around 25 drawers per hour, provided the boxes are available within reach next to the system. The image size of 61 megapixels allows a resolution of around 150 pixels per centimeter (ppcm) and thus detailed insights into a collection. Visible parts of small-print labels can be easily recognized. Naming, further processing of the image files and data management require additional working time.

The light hemisphere also enables shadow-free illumination (Fig. 8) with very high R_e values. The homogeneity of the illumination is high in the inner area of the hemisphere (diameter 16 cm – i.e., the area in which the bowl with samples is located). The maximum deviation of the five measured irradiance values is 4.7%. In an area with a larger diameter (32 cm), the brightness increases significantly towards the edge, with a deviation of 20% – but this range is not relevant in this specific example. The pure time for photographing a sample is about two minutes per sample (30 per hour). More working time must be taken into account for preparing the samples, as well as for naming, further processing of the image files and data management.

Measuring the quality of light sources has, on average, revealed a significantly worse light quality than I expected. It could be argued that further problems can occur during subsequent colour management, e.g. due to the possible incorrect and distorted display of colours in monitors. Suitable lighting can certainly not prevent such errors, but is not a source of error itself. Apparently, too little attention has been paid to the quality of light sources in photography and digitization. For example, Mantle et al. (2012) and Blagoderov et al. (2014) present a device for mass digitization with the Satscan, without mentioning the quality of the light source. Hardly any light source tested, other than the daylight LEDs, achieves the CRI value of more than 90 required by the Federal Agencies Digital Guidelines Initiative (2023). The light quality of the mercury vapour lamps reaches a maximum of R_e = 84, but in one case only 65 (SatScan in London). Although mercury vapour emits photons over a wide spectral range, it is not possible to achieve a spectrum without pronounced peaks. Due to environmental aspects, it can be observed that mercury vapour lamps are being replaced by LEDs in many places. This is also desirable in all areas where the recognition of colours plays a role, such as work rooms in museums where taxonomists work. Here, for example, the ceiling lighting at the NHMUK in London only achieves R_e values less than 70. This is all the more regrettable as some work rooms have no windows at all, which in my opinion also has a negative impact on general work in these rooms.

The measured LEDs vary considerably in terms of light quality. Satisfactory values are achieved in the exhibition area of museums, where LEDs are used in which the blue component is significantly reduced for conservation reasons and which therefore appear ‘warm white’. Here, the focus is not on perfect colour reproduction – as is the case with street lighting, for example – but on protecting sensitive objects. Surprisingly poor R_e values are achieved by the Supra Quartz Scanner at the NHMUK in London and the Leica Z16 microscope at the Museum of Natural History in Berlin. In both cases, these are LEDs with a strong blue peak and a pronounced deviation from the daylight spectrum. It is difficult to understand why LEDs of this quality, which are unsuitable for colour reproduction, are used in otherwise expensive devices. Flash units perform very well in terms of colour quality with CIE R_e values of almost 90 and can be used without any problems. Extremely good CRI values are only achieved by the daylight LEDs presented here. R_e values of >90 or >95 are now available on the market for a number of LEDs according to manufacturer information. In more recent works such as by Steinke et al. (2024), lighting systems with daylight LEDs are described. Daylight LEDs were still a niche product around 2020, but they are now easy and inexpensive to obtain. However, many products are only on the market for a short period of time. For example, the daylight LEDs used in the light cylinder are no longer available in this form; however, they can easily be replaced by LED strips of equivalent quality (e.g. from Lumitronix). The results suggest that daylight LEDs should in principle be used in all applications in standardized photography and digitization. I recommend not relying on manufacturer information alone, but to carry out spectral measurements if in doubt. Converting existing devices to other LEDs should generally be possible.

All three lighting devices serve their respective purpose, although each has a different basic shape (cylinder, cube, hemisphere). The basic principle is to illuminate from many sides, which is made possible by scattering the light from the inside. The aim of the lighting is to obtain as high a proportion of scattered light as possible, with perfect conditions occurring outdoors on a cloudy day. This ensures that there are no unwanted shadows that can lead to problems in image evaluation. The devices presented do not, however, enable reflection-free photography, but in the light cylinder and in the light hemisphere these “specular” reflections are greatly reduced by the use of diffusers (Fig. 8). The light box is not (yet) optimized in terms of reflections, because direct reflections hardly occur in Lepidoptera due to the fine scales on their bodies. However, if other groups of insects are to be photographed, a diffuser could also be mounted under the ceiling so that, for example, on the shiny surfaces of beetles, no individual LEDs can be seen, but instead a more homogeneous, brighter surface.

In all devices, light is not allowed to leak out in large quantities (energy efficiency and uniformity) or from the outside to the inside, for example when the devices are operated in a room with windows. Another approach is taken by Steinke et al. (2024), for example – here the lighting is positioned freely in the room. This is particularly useful for applications that require a lot of space, and in this case practically the entire room is a large ‘box’. One advantage of this method is the freedom of movement without restrictions. A disadvantage is that it takes up an entire room, requires a relatively large amount of light and energy, and one is unable to carry out UV photography for safety reasons. The light cylinder has proven itself in photographing individual moths over the past four years. The workflow described with manual focusing, an independently adjustable scale and the insertion of moths on individual aluminum pads, works well. I have produced many thousands of standardized images in this way. I had previously photographed with a Kaiser reproduction system in which light was shone on the object from the left and right. With this method, two shadows were always visible, which is no longer the case with the new methods described here. Shadowless images make subsequent image processing (e.g. for colour plates) much easier, as isolating objects is efficient and precise. Further images, in addition to the images shown in Fig. 6, can be found, for example, in a book that covers all of Germany’s butterflies (Settele et al. in press). Since 2020, thousands of representatives of the Geometridae, Arctiinae, Sphingidae and Saturniidae have been photographed for research projects in Ecuador and Peru. The standardized photographs allow the reliable quantitative evaluation of traits such as brightness and contrast for each individual as well as automated recording of data on body size – and this for a spectral range that also includes UV. Working in the light cylinder is somewhat difficult due to the limited space (diameter 45 cm). Future cylinder models should therefore be dimensioned somewhat larger. The light box was only recently put into operation in combination with a new Sony alpha 7R IV, so there is no long-term experience with it yet. The results are very promising, however. Initial tests have shown that up to 30 boxes can be photographed per hour if they are placed in readiness next to the device. A major advantage of the method is the high throughput and also the manageable costs – especially compared to solutions with medium format cameras and giant sensors. The image size of 61 megapixels allows details on labels to be recognized – ten years ago, such values were still outside the consumer-grade cameras described by Holovatchov et al. (2014). In fact, the effective resolution with the camera described is close to 61 megapixels. If an even higher resolution needs to be achieved (e.g. with very small insects), it is also possible to take four images per box. In the Phyletic Museum Jena, for example, there is a system with four individual boxes per box that can be photographed individually in this way. The Satscan devices in use in some large museums produce very large images (Mantle et al. 2012). However, only a maximum of around 10 scans per hour can be produced here, at a much higher purchase price and with an image quality that shows major weaknesses, especially in the reproduction of red tones (Fig. 5c). Cheap webcams with a very high nominal resolution are not comparable and it is hopeless to expect a comparable performance to that of full-format sensors.

Figure 6. 

Photographs taken in the light cylinder: a. Male and b. female of Gonepteryx rhamni (VIS and UV); c. Male and d) female of Lycaena alciphron gordius (each in the VIS and UV range). In the UV range, patterns become visible in the males of both species through UV photography that cannot be perceived by humans.

Figure 7. 

a. Photograph of complete insect drawer (Inventory number PMJ Hex-1262 from Brettschneider collection). Scale bar: 1 cm, taken in light box with Sony alpha 7R mark IV with 61 megapixels; b. Inset with details of single labeled butterfly specimens.

Figure 8. 

a. Photograph of ethanol sample taken in light hemisphere with indirect LED lighting. Scale bar: 1 cm, taken with Sony Alpha 7R IV at 61 Megapixels; b. Enlarged section with beetles; c. Comparison with direct lighting.

Photographing entire drawers is useful in order to begin recording collections where they do not yet exist. This appears to be an important option, especially in smaller museums, because the resources for individual digitization are not available. Here, it is often the case that hardly any changes occur in historical collections over many years, so that the images represent the current status for a long time. Another advantage is that negative changes are documented, for example due to pest infestation or theft. For parts of the collection, however, where a lot of work and sorting takes place, photographing entire boxes is less useful. Individual recording is particularly suitable for such parts of the collection. The light hemispheres are a simple, inexpensive alternative that anyone can build. Their appearance, however, has a more “DIY” character, but this does not have to be a problem. The hemispheres, like the light cylinder, are relatively cramped. For some objects, the height of the hemisphere itself is not sufficient, so the sphere can be placed on a base, which can be made of polystyrene or wood wrapped in aluminum foil, for example. Shielding the direct LED radiation using aluminum adhesive tape enables low-reflection photography of wet samples. Further work will have to show whether, and to what extent, it is possible to identify material based on the photographs. The high resolution achieved in any case allows the taxa to be identified by order, but usually at a much higher taxonomic level.

UV photography enables the documentation of colour patterns in insects and other objects that go well beyond the usual RGB colour space. This is sometimes impressive, as in the case of male brimstone butterflies (Pieridae: Coliadinae) and many copper butterfly (Lycaenidae: Lycaeninae) species, where patterns previously not visible to humans appear (Fig. 6). But even in more inconspicuous species, it is often worthwhile to record the UV information. To my knowledge, digital colour information in digitization projects is recorded practically 100% in the RGB colour space. The reasons for this seem obvious: RGB colour spaces are perfectly red, green to human needs, all devices – from digital cameras to monitors – work with the three colour channels red, green and blue. The downside, however, is that this view is very anthropocentric. The RGB colour space is certainly appropriate for almost all types of cultural assets, but it is questionable for biological subjects and the digitization of biological collections. A comparison: Would it make sense to document the vocalizations of grasshoppers, bats and elephants only in the sound range that humans can hear? The question seems absurd, because it is of course clear that ultrasound and infrasound are relevant for these organisms, and that biological work naturally records these frequencies that are inaudible to the human ear. The same question applies to vision: Why is the UV range largely ignored when digitizing collections? Humans cannot see in the UV range – but the vast majority of species can, in particular most arthropods, but also most groups of vertebrates (with the exception of most mammal species including humans), in particular many bird species are able to perceive UV (Ödeen and Håstad 2010). Fig. 9d shows an example of the spectral range covered by birds (example: Eurasian Blue Tit, Cyanistes caeruleus (L.)) and insects (example: Honeybee, Apis mellifera L.). The figure shows that a modified camera has spectral sensitivities that are close to those of an insect eye (trichromats: UV, blue and green) and those of a bird (tetrachromats: UV, blue, green, red). With the help of false-colour images, these can be recreated, for example, in the respective ‘viewpoints’ (Fig. 9d).

Figure 9. 

a. UV photography provides an additional fourth colour channel (UV), in addition to the usual channels B = Blue, G = Green, R = Red; b. Normal RGB image; c. False colour image with colour information GBU, shown in the usual RGB colour space; d. Measured relative sensitivity of a camera sensor in four colour channels (modified Sony Alpha 7) compared to the relative absorption of the cones in a bird’s eye (Eurasian Blue tit: Hart and Vorobyev 2005) and in an insect (Honeybee: Briscoe and Chittka 2001). Eurasian Blue tit: UV, S, M, L refer to the ultraviolet- (UVS), short-wavelength- (SWS), medium-wavelength- (MWS) and long-wavelength-sensitive (LWS) single cones, respectively.

One reason that the recording of brightness information in the UV range has been neglected is certainly that the same systems that are used for photographing ‘normal’ cultural assets simply cannot be used. The recording of UV data is currently (still) a ‘tinkering task’, and efficient concepts and tools for the evaluation of data seem to be largely lacking. The necessary cameras that are comparable in comfort and price to normal RGB cameras cannot yet be bought off the shelf. However, as the example above also shows, it is not at all impossible to produce high-quality UV images with reasonable financial means (Prutchi 2017). In certain cases it also seems worthwhile to record other patterns, such as UV-induced visual fluorescence (Brecko et al. 2016).