Practical Bud Observation, Measurement and Analysis
The aim is to secure “good” pictures of “good” buds at least once a day, as far as possible at the same time of day and under identical conditions each day.
A “good” bud is one to which the path geometry can be applied without too many problems: a bud with a loose “segment”, for example, or with obvious distortion of one sort or another will clearly not be selected for observation, unless the imperfections can be dealt with in some way. The Australian Oak, upper right, can be rescued by “image surgery.” In general a good bud, like the Scottish Beech on the left, will have a relatively smooth profile – but what “relatively smooth” actually means in practice is a matter of judgement and experience: some buds, such as the Maple from California, lower right, by nature have “stepped” or otherwise somewhat sculpted profiles and some of these can fall within our geometric “grasp”, so to say. There are a few tricks we can apply to them at measurement time.
A “good picture” of a bud is, essentially, one in which the bud’s profiles stand in clear contrast to the background—and to the bud’s own, interior parts, for that matter—so that they can be accurately located and measured. The picture of the Beech, left, is a good one in this sense. The picture of the Oak is less good, but useable.
Shadows, both on the bud itself (as on the Oak, left, which has “lost” its lower left profile to a shadow) and cast by the bud on the background such that they butt against the bud image, present difficulties, and are to be avoided.
A sheet of glass behind the bud, but some distance in front of the background, is useful in this respect: transparent things cannot support shadows. The glass can also help to station the bud at a standard distance from the camera lens, for focus and so on, and can bear both the scaling marks needed for calibration, and a note of a bud’s identity, either in felt-pen on the glass, or on a slip of paper clipped to the sheet alongside (not behind!) the bud. A bit of graph paper with the ID written on it neatly meets both of these needs.
The glass also helps abate another nuisance: namely, back-scatter. This is specular light scattered from the background on to the profile of the bud. It can over-ride the profile’s native colour, spoiling the contrast and making location of the profile uncertain. The further the bud is from the background, the smaller the back-scatter.
As the glass becomes soiled in time, the sheet should be readily removable for cleaning—and in this connection glass is to be preferred over plastic materials, as it resists scratching.
The camera’s flash is convenient for standard illumination, but if the source is built into the camera’s body, it is not often in the best place to do the job we want it to do, and can’t be moved. Many cameras, if not most, are not designed for close-up flash-photography. And the flash, even if it is well placed, is generally far too bright. One can in such a case either limit the flash aperture with black electricians’ tape, or reduce the light from a full aperture by overlaying it with layers of white paper or finely-woven white cloth. Sometimes one can “bounce” the light from a misplaced flash to where it is needed using judiciously placed masks and reflective surfaces. The rig sketched above worked quite well. Strips of Perspex can sometimes be used to conduct the light, fibre-optic-fashion, from the flash source to the bud. The strips can be (gently) heated and bent into shape by hand. See one result above right.
In many cases, it will probably be easiest to provide an alternative means of illumination. Two torch (flashlight) bulbs flanking the lens of the camera, but away from it to right and left far enough to ensure that reflections of them in the glass are not seen by the camera, will be found to work very nicely, eliminating those shadows mentioned above. Of course, the bulbs need a power supply. Some engineering is always needed! Banks of LEDs can also be used.
Though silhouettes are possible, and work very well when only the profiles of the buds are wanted, they do hide structural details of the buds that may be wanted for study some day, so it is probably wisest to take the pictures with the buds frontally lit, “full moon” style. Some buds so photographed do present measurement problems, but these can usually be coped with fairly easily, so, in the interests of posterity, “full moon” photos probably represent the best option. After all, the photos can only each be taken once. It will gall if we one day find that we want to review structural information from archived buds, and then find that we have prevented ourselves from obtaining it!
The background used in rigs to date has so far generally been either matt black or matt white. If there are shadowed edges on the bud, it can be difficult to distinguish these edges from a black background. If the background is white, the edge-shadows cease to be a problem, but instead, the afore-mentioned problem of back-scatter can become quite significant. Recent work has suggested that neither black nor white is necessarily the best choice. Buds have a dominant colour. It is probable that the best background is one with a saturated version of the complement of the buds’ dominant colour, or, at any rate, with a very different colour from the bud. Then a filter (either on the camera lens, or in computer software) favouring the dominant colour should help to “lift the bud clear” of its background. First attempts along these lines have shown promise.
Digital cameras are the best choice for bud photography: after the initial outlay, they cost next to nothing to run. There is essentially no processing delay, and the pictures are in a form immediately suited to the computers on which most of the measuring and other processing is likely to be done. Images can readily be transmitted by wire to others. They can be compressed and conveniently archived. The resolution of digital cameras, good as it now is, is still not as good as film cameras, but this does not matter. Digital resolution is already better than it needs to be for the bud work. 640 by 480 is more than sufficient. Even a web cam will do at a pinch, and nowadays, very small but good TV cameras are widely and cheaply available, along with interfaces to computers, which will allow bud observation to be automated. Automatic observation is as yet little developed, and holds much promise.
Many relatively inexpensive and otherwise entirely suitable digital cameras suffer from a major deficiency from our point of view: they cannot, unassisted, get sufficiently close to the buds. Only top-of-the-range—and most expensive—cameras seem to be able to do this. This generally means that an extra lens needs to be attached, somehow, over the original lens of such cameras. Some come with rings for the purpose. Most do not. This is a moment for the amateur engineer!
Lenses have been attached successfully with dental wax, of all things. Frames have been fashioned in various ways and in various materials and secured to their cameras by various means. However the extra lens is attached, it must be perfectly stable in use. And if the new lens is not a fully-corrected camera lens in its own right (one might try a simple hand magnifying lens, for example), there will be trying problems with optical distortion, chromatic aberration, and so on. Optical distortion is liable to compromise bud measurement, especially if the bud cannot be held in one place in the field of view from picture to picture. An attempt is being made to address the problem of optical distortion in the software, and it may eventually succeed, but, all things considered, it is best to get hold of a properly manufactured camera lens, if one can.
The attitude (tilt) and whereabouts of the bud in the image frame are not of too much concern (given a good lens), if they fall within reasonable limits, say about thirty degrees tilt either way, and the inner three quarters of the frame area – unless it is wished directly to compare one image with another, in which case attempts should be made to standardise both. This is a fairly tricky affair, to which we will return in a moment, in connection with another, more important, matter.
By far and away the most essential thing to keep constant from photo to photo of a bud is its aspect. In one sense, it does not much matter which aspect is selected at the outset, but it is vital that that aspect, once chosen, is rigorously preserved. As far as is practically possible, the bud must always have the same orientation around its long axis with respect to the camera. Experience suggests that for usefully consistent and valid results the aspect should be held to within at most plus or minus two degrees.
Of the various tricks and devices that can be deployed to help preserve aspect, probably the most effective is simple practise. As will be seen fairly immediately when sessions are attempted, a great many circumstances intrude themselves on practical photography! One deals with them, one by one, each on its merits, as best one can, and over time, one establishes working, and workable, habits. In due course, the camera will come to be presented to each bud in the same way each time almost automatically. It is actually rather remarkable how precisely the procedures come to be repeated, without conscious effort. It says much for our unconscious faculties!
It is possible to attach a “cross peg” permanently to the stem of a bud (not to the bud itself), that engages in some suitable way with an attachment to the camera lens- barrel or body, so as to keep the orientation of camera to bud precisely constant. This is tricky, and is not as permanent as one would like, but, when it works, it works very well. The cross peg (a match stick serves well!) takes care of the other, less important, orientations at the same time. The cross peg can be wired in place if the stem is a woody one, or it can be “super glued”. But the wire eventually works loose—and the super glue may murder the bud!
Strapping a bud with thin wire to a slide of Perspex has been tried, with some success. The slide is dropped into a receptacle on the camera for the photograph. There was some anxiety about confining the bud in such a gross fashion, but, in the event, the buds seemed not to mind too greatly! However, the wire bonds do obscure the profiles where they cross them, which can interfere with measurement, and the bud can and does squirm in its bonds.
As may have been gathered, this constant orientation business presents us with a difficult, technical problem, to which no wholly satisfactory solution has yet been found, other than human adaptability in the field. There is room for ingenuity here. If anyone finds a good solution, please let us all know what it is! Lawrence Edwards has tried yet a different tack: he took five or six pictures per bud per session, and either averaged the measurements from them all, and found the l of this “composite” bud, or averaged the individual l values. He said this yielded good, reliable and consistent results, “the best he’d managed.”
We end this section on photography and observation with a remark about daylight. Why do we prefer not to use it? The main reason is its variability. Direct sunlight is simply too bright; daylight generally creates unpredictable shadows; it is not available at night – and so on. Our bud measurements are best served by controlled, constant and favourable lighting conditions, and, as part of achieving them, we actually need to exclude ambient light by enclosing the bud in a nearly light-tight shell for the moment of photography.