Andy Atkinson
A number of survey software packages now let you include LRUD (Left Right Up Down) passage dimensions in some form or other and will produce a plot using this information. All of these, with the exception of Toporobot, are simplistic in their approach, and if you have seen the output you will recognise the rather odd-looking characteristic shapes that occur. There are good reasons why this is so. The conventions that surveyors use when recording the data, such as 'in the direction of the survey', and 'across the passage' are not very easily defined in computing terms. Also the fact that LRUDs are typically taken at stations, and stations are often at junctions causes problems as junctions tend to be atypical, rather than typical, of the passage. There are a whole host of other things like what to do when a value is not given, a question mark is entered, and at the ends of passages. If zero is assumed at an omitted reading then a pinch-point occurs in the walls at that station - not really the desired effect.
We have been examining the details of this process with a view to implementing something in Survex, and generationg a proper 3D model of the complex Kaninchenhohle. It quickly became clear that as well as the programming difficulties there were significant problems in terms of the data that was actually collected. Obviously a set of 4 numbers conveys much less information than a sketch. Unfortunately a sketch is inherently 2-dimensional, and thus is not very helpful when trying to construct a 3D model. So the question becomes - what is the least information that needs to be recorded in order to construct a useful 3D model? Obviously the answer depends on what you want to use the finished model for, and the usual constraints on surveying manpower, time & conditions.
Andy Atkinson took a look at the specific area of improving the information contained in the LRUD data without dramatically increasing the time it took to record, or the complexity beyond the point which surveyors would stand (where relatively inexperienced Cambridge Cavers in Austria's horrible caves have a particularly low tolerance of such things). Obviously improvements of this nature are no use if surveyors think they are too much extra work. Here he presents his second iteration of the idea for comment.
The computing aspects of LRUD interpretation and
the broader issues of wall modelling need articles of their own
to explore. These will be in future issues.
The suggested format is an extension of the now standard LRUD, with a 5th column -'E' for Extension - which is used in some cases. One obvious improvement is a notation for allowing more than one value to be given in the same direction. This particularly useful in traversable rift where you really want to indicate the distance to both the actual and apparent floors, or sometimes in a wide bedding where only the centre part is person-sized. There are many possibly notations, all prone to confusion, or not completely general.
 | L & R are defined as the bisector of the legs. For the last station they are perpendicular to the last leg. ~ is used to indicate estimated distances (very useful to know which numbers may be suspect when drawing up) | ||||||
 | (220) | Where L or R in standard direction is unhelpful, irrelevant or meaningless another direction is given. The approximate bearing is given in the comments field. | |||||
Pn | P(n) |
P(n) | P(n) | (NE)(160)(260) | For pitches give NSEW instead of LRUD. Where NSEW is not appropriate (e.g. axes of elliptical shaft lie in another orientation) then use bracket notation to give bearings | ||
 | At a Junction the value that would otherwise disappear down the joining passage is given as where the wall would have been if the joining passage was not there. | ||||||
 | When a survey ends at a junction the L & R values for the surveyed passage as if it continued are given. Also given is the 'Extension'. The distance that the survey leg would need to be continued to meet the wall. | ||||||
 | At a point where a passage meets a pitch The extension is given to the far wall and the roof or floor (roof in the example) is given using the Junction notation. | ||||||
 | For a perimeter survey use C (for Chamber) in either the L (anticlockwise survey) or R (clockwise) column. Readings are on bisectors of legs | ||||||
  | C Cn | C Cn | n n | n n | n n | For a radial survey the centre point is given like this Then the other give the Extension value to the wall, as well as U & D In more complex areas L & R values can also be given if they are significant | |
 | For more complicated bits of the cave the notations given can be combined to fit the need |