No rojak today (haven’t found anything interesting enough this past week), so I thought I’d post a few thoughts about my current research. The warhol-esque image represents a major part of my work, the digital image analysis and my attempts to re-present the rock art into meaningful bits of information.
While I do have physical finds to examine, the bulk of my data consists of the visual imagery that’s been recorded on forms and by the digital camera. For each rock art element that’s been recorded in the field, digital image analysis on the computer helps to confirm – or more often, expand on – what’s been observed by the naked eye. This is particularly important since I’m building an inventory of what’s on the walls of the site. The top left image is the ‘raw’ image rom the site using a DSLR. The subject of the photo is the anthromorph, but you can see some other rock paintings to the right and one rectangular shape between the left leg and foot, as well as some exfoliation on the bottom left corner.
On the top right is the same image run through a decorrelation stretch, a nifty image processing filter that basically translates the original colour matrix (RGB) into a different one (in this case, LDS) which results in a false-colour image that makes certain colours more distinctive than others. The type of colours enhanced depends on the type of stretch matrix used – and you can see here that the anthromorph turns up purple, and some red dots that were nearly invisible to the naked eye also show up quite nicely in this image.
The combination of the original and false-colour image are then used to derive two more images, the highlight photo (bottom left) and the shape photo (bottom right). A highlight photo is basically a black and white version of the original with the rock art subject in colour. It makes a good illustrative device to draw attention to a single element, particularly in more crowded images with complex compositions. The shape photo is a negative version of the highlight photo – it only shows the isolated black-and-white element on a white background.
Interestingly enough, this rock painting was first described as a ‘man with a club’ – after the digital image analysis, it looks like the ‘club’ appears to be entirely separate – in fact, what may be the left ‘hand’ has been weathered away by exfoliation, so there’s no way to be sure – which means that I’ll have to record the ‘anthomorph’ and the ‘hand’ as two seperate elements.
All this tweaking on photoshop is quite time-consuming. A small investment in a computer tablet (where you use a stylus instead of a mouse) goes quite a long way in making accurate digital tracings – not to mention the frustrations with using a mouse for such precise work!