We have a number of pre-1990 surveys in mulu for which all the centreline data has been lost but we still have the drawn up surveys. We have regenerated centreline data by inventing survey stations and legs and working out XYZ coordinates for them (very laboriously)2, but now I am putting these surveys into Therion. At the moment I am using the XYZ base survey data and putting in stations for these on the plan. However the positions of these stations on the drawing are very aproximate (I don't actually have the original plot showing exactly where the made-up stations are - one can just just guess from the look of the survey). But this introduces errors of approx the width of the passage, which can be several meters, and is adding distortion to the drawings I am putting in. I am wondering if there is a better way - so I put the drawings in and say 'these are all we have - you work it out', rather than using ill-defined stations which were only made up by taking off this very same drawing in the first place but now they have been through numerous innacuracies and distortions. Can therion do this? And how would I work out the scaling? Can I use a scale bar or something and just copy the scale into each scrap? Wookey -- Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK Tel +44 (0) 1223 811679 work: http://www.aleph1.co.uk/ play: http://www.chaos.org.uk/~wookey/
I am wondering if there is a better way - so I put the drawings in and say 'these are all we have - you work it out', rather than using ill-defined stations which were only made up by taking off this very same drawing in the first place but now they have been through numerous innacuracies and distortions.
There two possibilities how to solve this: 1. To position each scrap - in principle you need only two stations for each scrap. You do not need to enter a lot of stations. Then in output you can hide those virtual stations using symbol-hide group centerline. 2. You can use projection none for that - but then you will need to calibrate each scrap at real coordinates (if you have some grid on your paper, this should not be a big problem). 3. We can change the behaviour of therion and allow calibration of scraps using only -scale option also in other projections (currently it is allowed only in the "none" projection) But in any case, you will need to specify -scale for each scrap.
Can therion do this? And how would I work out the scaling? Can I use a scale bar or something and just copy the scale into each scrap?
In the scrap -scale option, you can specify coordinates of two real and corresponding two image points. There is no scalebar or something like that. Copying can be done in text editor (very carefully :) Regards, S.
+++ Stacho Mudrak [04-12-29 09:44 +0100]:
I am wondering if there is a better way - so I put the drawings in and say 'these are all we have - you work it out', rather than using ill-defined stations which were only made up by taking off this very same drawing in the first place but now they have been through numerous innacuracies and distortions.
There two possibilities how to solve this:
1. To position each scrap - in principle you need only two stations for each scrap. You do not need to enter a lot of stations. Then in output you can hide those virtual stations using symbol-hide group centerline.
Yes I've been doing this - and since discovering debug mode I've been correcting the stations which were badly placed. Nevertheless there seem to be some errors in our original 're-generation' of the fake data which means there is some stretching of of the image. It would be better to use the image itself as the reference, with the only stations involved being ones where real surveys connect. I tried adding one passage (for which we didn't even have 'fake' data) as 'projection none', but this just produced an error (something like 'incompatible projection'). Is it possible to mix projection none and plan in a map (and have joins between them)? If not how would you add a load of old 'lost data' survey to some new data (which is what I need to do).
2. You can use projection none for that - but then you will need to calibrate each scrap at real coordinates (if you have some grid on your paper, this should not be a big problem).
I don't have a grid on the paper. There is nothing but a scale bar in one corner. What exactly am I trying to do here? There are 8 number to be given - 4 picture scale points and 4 real scale points. Let us say I have a scrap on a sheet of paper and I know nothing but the distance that corresponds to 50m. How might I go about making up some numbers for the scrap? And would I then just copy those numbers to all the other scraps? I suspect we need a little how-to for importing such 'lost data' surveys to therion, and how to best combine them with new survey data.
3. We can change the behaviour of therion and allow calibration of scraps using only -scale option also in other projections (currently it is allowed only in the "none" projection) But in any case, you will need to specify -scale for each scrap.
Would this solve the 'incompatible projection' problem when connecting to plan projection scraps?
In the scrap -scale option, you can specify coordinates of two real and corresponding two image points. There is no scalebar or something like that. Copying can be done in text editor (very carefully :)
OK - but as above I have no idea how to give these numbers. Which way do they increase? are both sets of numbers entirely arbtrary and only the ratio matters? Are there overflow limits I should worry about? and so on... What a lot of questions! Wookey -- Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK Tel +44 (0) 1223 811679 work: http://www.aleph1.co.uk/ play: http://www.chaos.org.uk/~wookey/
At 20:45 +0000 29.12.2004, Wookey wrote: *******************************************
I don't have a grid on the paper. There is nothing but a scale bar in one corner. What exactly am I trying to do here? There are 8 number to be given - 4 picture scale points and 4 real scale points. Let us say I have a scrap on a sheet of paper and I know nothing but the distance that corresponds to 50m. How might I go about making up some numbers for the scrap?
You may make a grid in Gimp, Illustrator, ... on top of your scan and generate a "scan with grid" My 0.02 EUR M. --
At 20:45 +0000 29.12.2004, Wookey wrote: *******************************************
I tried adding one passage (for which we didn't even have 'fake' data) as 'projection none', but this just produced an error (something like 'incompatible projection').
Wookey, I hope you may use minimaly one reference point - from passage where this "fake" passage is connected to. Martin --
'incompatible projection'). Is it possible to mix projection none and plan in a map (and have joins between them)?
No it is not. In the map - all scraps have to be from same projection.
If not how would you add a load of old 'lost data' survey to some new data (which is what I need to do).
Well, If I should do this, I would probably do it following way: 1. For each paper define one virtual station, that will position this paper in real world coordinates. 2. For all scraps on this paper, I would insert into each of them only this single station (and I would give it -visibility off option, that it will not be visible in the map). 3. Scaling and rotation of scrap would be done using -scale option. If you have a horizontal 50 m scalebar on the paper, and paper is facing north, than I would use 0 0 50 0 as real scale coordinates for edges of scalebar. See attached picture. Real coordinates are not important, only their difference (vector) is. Scrap will be positioned according to stations specified in it.
- 4 picture scale points and 4 real scale points. Let us say I have a scrap on a sheet of paper and I know nothing but the distance that corresponds to 50m. How might I go about making up some numbers for the scrap?
As above. All you need to do is to specify paper and real vector. From these two vectors rotation and scaling are determined.
And would I then just copy those numbers to all the other scraps?
Yes, for all scraps lying on the same paper, these numbers would be same.
OK - but as above I have no idea how to give these numbers. Which way do they increase? are both sets of numbers entirely arbtrary and only the ratio matters? Are there overflow limits I should worry about? and so on...
Both paper and real coordinates are standard cartesian. X = east, Y = north. Yes, numbers are arbitrary, important are only relative positions. There are no overflow limits - you can probably use any numbers you like. 4. When scraps will be drawn, I would then adjust positions of those virtual paper stations, that scraps will be as close as possible. When scraps will be close to each other - then I would join them. And that's all. You will need only so as many stations, as many papers you have. Mixing projections, or calibrating scraps only using scale option I do not consider as a good idea. Once you may decide to use other coordinets (e.g. GPS), and you will need to shift all papers somewhere else. If real coordinats will be in form: fix paper1 10 20 30 fix paper2 20 30 40 it will be very simple, because it will be in one table. If scaling points would be in scrap headers, it would be very hard. Another point - this way (one station per scrap or paper) each scrap also receives some vertical information. If you would like to color scraps according to altitudes (it should be possible in the next release), it will be also possible. When scaling scraps only using -scale option, there is no vertical information. I am not sure, whether I wrote it understandable way, but I hope you understand what I wanted to say. Happy new year to everybody, S.
+++ Stacho Mudrak [04-12-31 09:19 +0100]:
'incompatible projection'). Is it possible to mix projection none and plan in a map (and have joins between them)?
No it is not. In the map - all scraps have to be from same projection.
If not how would you add a load of old 'lost data' survey to some new data (which is what I need to do).
Well, If I should do this, I would probably do it following way:
<snip description>
I am not sure, whether I wrote it understandable way, but I hope you understand what I wanted to say.
Thanx Stacho - that's exactly what I needed to know, carefully explained, and yes, I think it all made sense. I'll give it a go and get back to you if I have any problems. Another excellent answer for the FAQ :-) Wookey -- Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK Tel +44 (0) 1223 811679 work: http://www.aleph1.co.uk/ play: http://www.chaos.org.uk/~wookey/
participants (3)
-
Martin Sluka -
Stacho Mudrak -
Wookey