BioImages - Notes on Scanning

The following relates to my use of the BlackWidow 9630SP flatbed scanner, in the early years, from 1997 until 2002, when it was replaced by a Hewlett Packard Scanjet 5470c. Shortly after this I stopped using scanners on biological specimens and all my more recent images are from digital cameras. Some people have continued to use scanners to photograph leaf-miners, using the transparency adaptors to good effect, but I find I get better results photographing them on a small cheap light box.

Most of the early images were simply scanned on a flatbed scanner. Laying a specimen on the glass platen is surprisingly effective. Plants can be arranged to give "herbarium specimen" images - as though pressed but still retaining colour. Leaves, plant diseases, leaf mines and other flat objects are a natural for scanning, but 3-D object come out well too. Toadstools lie on the platen and usually give good results, as do vertebrate skulls. Dead insects if well relaxed can be positioned in life-like poses. Medium to large flies, especially those with large wings, will lay on their backs and you can usually get the wings to stay open and stick to the glass; breathing on the platen sometimes helps if it hasn't warmed up yet. Beetles are easy. Living moths will sometimes stay put long enough, especially if the platen has not yet warmed up, but tend to shed a lot of scales and I haven't had good results. Caterpillars on the other hand will often lie there without moving. I've even scanned pond life in a little well made of a loop of blue-tak stuck on the platen. (Don't try this at home!)

Specimens often won't lie in the desired orientation. Toadstools will lie on their sides and the flat capped ones will lie on their caps, but there are some positions you can't get iwthout intrusive props. Hymenoptera and some cylindrical beetles often refuse to lie on their backs, the weight of the legs pulling them over. Skulls will also only lie in a few orientations. You get three-quarter shots rather than pure side views.

Cover the specimen with a piece of paper to provide a background. I use bits of polystyrene of varying thickness to provide support. Experiment with different shades of grey; your local art shop will sell a range of colours, but neutral greys seem to be increasingly unfashionable. A black background is good to show hairiness of stems etc. Because scanners have a relatively low dynamic range (ie black and white are quite close) it is useful to choose the tone or the distance to the background paper to avoid sharp contrasts with the subject. A white subject on a black background is OK, but a black subject on a white background will give trouble unless the background is much further away.

Scanners are calibrated for pictures on white paper, so exposure often needs to be changed for natural objects. Pale toadstool flesh needs about 80%, most leaves and insects about 150% and black beetles 200% (the maximum on the scanner). Exposure acts by changing the rate of motion of the scanning bar.

Depth of focus depends on resolution. An object which scans out of focus at 600 DPI may look fine at 200 DPI. Most things will only lie in orientations which need less depth of focus than a sphere. Even for a sphere, whatever its size it's generally possible to get a reasonably sharp result at 600 pixels square and this is enough for a web image.

The scanner bar is shorter than the width of the platen and the cells are arranged so that they point progressively outwards towards the ends. This has meant occasional rescans because my polystyrene supports were visible when they shouldn't be. Also the light is brighter in the centre. This shows in the backgrounds of larger objects so it is best to always place large thick subjects in the middle of the platen; for smaller or thin subjects it doesn't matter.

Colours are usually understated: it's worth increasing the colour saturation or contrast a bit. You may also want to reduce the blue slightly to remove the blue tinge of the screen. Yellows and oranges often scan very badly and need to be manually increased which I find very difficult.

I use a Black Widow 9630SP scanner with optical resolution of 600 x 1200 DPI, 24 bit colour. 600 DPI is equivalent to 6-8 times magnification when displayed on screen. Black Widow claimed their scanner was specially engineered to increase depth of focus for scanning 3-D objects, but I don't know if this was just advertising talk, or even if it applied to this model.

Scanners are available with greater colour depth than 24 bit (eg 36 bit).

Some current scanners have optical resolution of 1200 x 1200 or higher, equivalent to 12-15 times magnification on screen, depending on screen resolution. Such scanners are ideal for leaf infections (rusts, smuts etc) and leaf mines (especially if used in conjunction with a transparency adapter - but photographing on a cheap slide viewer is even better).

Finally, keep the platen scrupulously clean, especially when using black backgrounds, and remove grit very, very carefully!

The above text relates to my own personal experience and is offered in good faith. It does not constitute a recommendation.