Going out to do a practice survey. Basically just getting used to the equipment. I picked a spot I figured would be easy with no worries of trespassing or worrying local prop owners. I'll be working inside a city park that borders a community forest. There is one building, a community lodge, and several trails. The park is bordered by several streets on two sides and the forest on the other two. The entrance to the park is from an unmarked road that curves off from a marked intersection. The City and County have a play/prop searchable database online which only provided an assessors map. I checked the surrounding plots with the same results. Due to Covid access to the City Recorder is very limited. The assessment map gives an outline of the surrounding plats, the streets and the park boundaries with a N arrow and 1" to 200' conversion note. No BM's, coordinates or any other location marks. The GIS database provided several survey BM's in the City. They are all a few miles away from the park, mainly around the freeway, and are blocked from view at the park by buildings. To get an accurate tie in will I have to start at the closest BM and due a horizontal track back to the park or is there another source I'm overlooking. GPS does not line up well with the park boundaries and I don't have access to an accurate GPS device anyways. Any help would be appreciated.
Joki, How are you planning to compute your coordinates and carry elevations? Be a good exercise to compute your coordinates by hand using lattitude and departure and trig your elevations. See how close you can close the x,y,z in a closed loop. think you'll quickly learn the limitations of the equipment you're using. Let us know how well you close.?ÿ
I'd recommend just practicing a closed traverse as a beginning exercise.?ÿ No need to tie to anybody else's coordinates for now.
In a park, I'd pick the most inconspicuous traverse points possible.?ÿ If there is sidewalk or paved parking lot, put a nail in a joint.?ÿ If you have to put one in a grassy area, use a small piece of wood set flush to the ground.?ÿ
You can still expect to get hassled by other park users or police.?ÿ Surveying scares people because they don't understand it and will assume you are going to do something to the detriment of their enjoyment.?ÿ I was once pacing distances in a wooded public park (actually looking for a geocache where GPS reception was terrible) and was confronted by a nearby homeowner who was sure I was going to pave a road through his favorite natural area.?ÿ It was quite tense before I convinced him otherwise.?ÿ When I was practicing triangulation using the lights on radio towers around town I got kicked out of a mall's nearly unused parking ramp because the security guy saw something he wasn't accustomed to and told me he "couldn't give me permission." I asked the mall director for permission and he said "against company policy" which I translate as "I don't know what the h*** you are doing so don't want to get myself in trouble if it's bad."
Good luck.
I'm not official, but even if you can't get a record map or description yet, I'd start by taking a look for any monuments at park corners and roam out beyond looking for any that might be corners of neighboring properties.?ÿ Start your traverse blind like others are suggesting.?ÿ Make as big a run around the park as you'd like shooting all your sideshots, including the external corners - might want to shoot those as control points if possible.?ÿ At least 2 or more of those would be good.?ÿ Carry on and close your traverse.
I don't have access to TBC, StarNet or anything cool like that to adjust my survey post processing so when I do have to do it I do it by 'hand'.?ÿ Good way to learn (and to also learn that you want to be using these other automated methods).?ÿ Adjust it and keep trying to find a deed, subdivision map or plats that would located any of the monuments you may have found and shot.?ÿ If so, check them for accuracy against your survey and begin the puzzle of why they're not nuts on - and they won't be.
As far as adjusting by hand, one thing that's helped me is point numbering. My station points start at 1 (ex hub, first traverse station, etc) and don't usually go beyond 20.?ÿ My sideshots start at 100, but if I make it to 179 and move forward to the next station, I start my new setup sideshots at 200.?ÿ Sometimes my 3rd station will start at 450 or 500, so it's not correlated directly to the station, it just makes it clear on import what set of sideshots were associated with what setup.?ÿ And of course good point naming is critical for making CAD even possible but can dramatically increase speed there as well.
I'm open to correction in any of this as well.?ÿ Not licensed, but just now ready to take my FS while the only testing location within 2200 miles has closed down for FauxVID, I mean, Corona Virus safety protocol.
Keeping your point numbers in different ranges is definitely a good idea. Lots of ways to do it. I have done comparisons to deed distances on a couple of lots.?ÿ I numbered the deed points 1, 2, ... and my measurements of them 501, 502, ... in exact correspondence, and any extra points needed for sighting 801, 802, ...
I did that to make it easy to compare in a least-squares fit which I did with a program of my own creation.
I did consider the closed loop scenario with picking an arbitrary point, parking lot or corner of the lodge building. Since I don't have a reliable way to verify GPS and no benchmarks to work from how would I verify my measurements? I guess once I plot it out I could then draft the measurements and elevattions and see how well the lines match up. Still just learning the methods here. I'm using an old Craftsman transit, wood rod in 8ths of an inch and a 300' fiberglass tape with no help, so not expecting any kind of accuracy on this one. I was planning on picking a spot and shoot to the nearest permanent object at the edge of the park. Continue this around the perimeter with BS and FS additions and subtractions until I returned the the starting point. Once I draft these figures it should all line up, right?
As far as locals are concerned I figured I'd have to be prepared to explain myself no matter where I practice. Hoping a park is a bit better than the City streets. Of course there are some transient campers in the woods adjacent that I'm going to have to watch out for.
I think you should start by trying to "close the horizon":
Set up on any point, pick out "natural sights (building edges, poles, towers, etc., but not cows) to break the circle into at least four random parts; measure each part separately and see if the angles you measure add up to 360?ø.?ÿ Repeat a few times with different set-ups.
Then establish a triangle, measure the three angles and see if they add up to 180?ø. Be sure to pick vertices that you can set-up on and leave a backsite on.
Then, work on a closed traverse.
In my 2+ acre yard, I have 4 60d nails set to where 3 are traverse points and the 4th is a sideshot.
Can see all 4 points from each other.
My helpers and I traverse thru the 3 points and sideshot the 4th point from each traverse point and also test any new TS and those returning from cleaning and adjustment here also.
When a TS is acting weird I take it back to these test points and compare it to the previous results.
Everything is computed and checked against every other test run.
I also use this test traverse to teach and see if my helpers are able to use the TS and data collector and evaluate their performance to see if they can move up and be an Instrument Man.
0.02
If you're just test driving some equipment (you didn't say what kind), why not just do a random traverse in the park you're in??ÿ set some temporary nails or whatever and turn and burn through them.?ÿ?ÿ
...I did that to make it easy to compare in a least-squares fit which I did with a program of my own creation.
Do you have anything posted on this or write up about it? I was curious if that was possible the other day seeing the price tags associated with some of those big box programs haha. It seems that some are saying it is not possible with MS excel. At this point I??m just really intrigued about any home grown stuff more for the purpose of learning about LSA?ÿ
For exploring simple examples of Least Squares problems, see if StarNet still offers a free demo mode. I have used an old one that lets you have 10 points.
My program is a pale imitation of the 2-D plane mode of StarNet, with a little different syntax, but handles a few hundred points and many observations.
It was a formidable programming task done in the now obsolescence Visual C++ environment under 32-bit Windows.
It only allows coordinates to be fixed. Bearings, distances, and angles must have standard errors. I have to provide the initial guess. I never got to the point of working graphics, and the attempts cause some crashing problems.
It does provide statistics, with some options for the measures. I found out that commercial programs may differ in how they define their results.
The key to making LS work for big problems is finding a robust algorithm for Singular Value Decomposition of matrices. I found some code on a university site and worked out appropriate ways to deal with the ill-conditioned roots.
I suspect there are some tricks still needed to improve convergence from an initial guess.
But despite the limitations the program does give me results that check against StarNet, and lets me work bigger problems.
If anyone is interested, I'll share my writeup that discusses linearizing angle and length measurements, outlines the matrix setup, and adds some ideas for novel measures of dependence and fit.
I'm not sure anyone would want to put up with the nuisance issues of my program.
@bill93 When you have a chance, no rush, I would be interested to see your writeup. I use the open-source GNU Gama least-squares adjustment package. A friend of mine is one of the contributors, and I've been passing him ideas for improvements and additions.
If you are using a transit you need to turn deflection angles right and left.
This makes it easier to calculate the next bearing.
@bill93 When you have a chance, no rush, I would be interested to see your writeup. I use the open-source GNU Gama least-squares adjustment package. A friend of mine is one of the contributors, and I've been passing him ideas for improvements and additions.
https://www.dropbox.com/s/esgmuxk2s0mqio4/LSQ_notes.doc?dl=0
If you are using a transit you need to turn deflection angles right and left.
This makes it easier to calculate the next bearing.
I see how that makes calculation easier if you are doing hand calculations, but "need" is a little strong because you can get the same results other ways.
?ÿ
Thanks for the suggestions. As I mentioned I will be using an older Transit level, a wood rod in 8ths and a 300' fiberglass tape with a stake. I have a long (2') nail and a roll of flagging "tape". I'll be breaking in my HP35s at the same time. The park I'm looking at is open with a building at one end. It has a mild slope over an open grass field and bordered by trees. I'm reviewing the methods of setup for a basic Horizontal run. I'll have to assign a random value for BM1 as I was unable to locate any historic BM's at the park. There is a manhole cover I can use as the start and close the loop to check my work. I can use the trees to practice trig leveling. The park is too busy to take many rod shots, so I'll stick with a basic BS-TP-FS run and plot a cross section.
