I need your advice on the "one step transformation" process for creating ground coordinates from grid coordinates
Our project area is a 260 km long proposed railway line corridor. The proposed railway corridor needs to be designed after doing the detailed topographical survey
The first step followed is : Established 85 pairs of control points (say GPS1 to GPS 85) along the corridor using DGPS.
Second step: Convert the control point LLH (Lat,Lon,Ht) to grid coordinates (XYZ) based on UTM projection system
Third step: Convert grid coordinates to Local coordinates using one step transformation with the method - One point localization. Here the localization base is one of the GPS stations (say GPS 83). During one point localization the ground coordinates (i.e Position & Height (P&H)) of all control points are derived w.r.t GPS 83 . Here the paramters for one point localization such as Flase easting,False Northing etc are set using UTM easting and northing of GPS83.
My question :
1. For a linear project of 260 km long, is it a right approach to create ground coordinates of all DGPS control points w.r.t to GPS 83 (which is at one end of the corridor)
2. Second question is, should multiple DGPS control points (use their UTM grid coordinates) be used for localization which may be spread out in the 260km long corridor
3. What are the advantages of using one point localization and what are its disadvantages (considering the fact that we need to re convert the ground coordinates back to true UTM coordinates).
The proposed alignment is designed on ground coordinates, and the design team is creating railway corridor centerline, right and left corridor points based on ground coordinates.
KML when created using ground coordinates (assuming they are UTM) cannot be used for visualization in google earth since they don't fall at the right location.
The final location survey : Where the corridor pillars are pegged on the ground using Total station are w.r.t to ground coordinates.
The problem I am facing is : the land mapping is done using UTM coordinates (geo-referencing of cadastral maps is done using DGPS coordinates / UTM coordinates), and we need a solution for the problem so that land acquisition proposed using the cadastral maps matches the final location survey.
This Content Originally Published by a land surveyor to Land Surveyors United Network
Very interesting question. I am not prepared to comment off the top of my head but the problems are not unlike what is involved in highway construction. Various State Highway departments have published information on this, Texas as an example. Everybody knows a rectilinear survey will fall apart (develop excessive error ) at some point if not adjusted for curvature etc.. Local construction does not normally take curvature of the earth into consideration. I never saw an architect consider it in the design of a building as an example. I would take a look at how highway engineers handle the problem if you haven't already. I am very interested in following this & reading the comments of others. Thank you for posting this very relative question.
My suggestion is that you establish one base control point using your Ground Control Coords. Then localize on that point. Set your datum as WGS84 then carry out your survey. All your observations will be stored with Ground Control Coords. Using a Cad software such as Magnet Office you can then transform to your cadastral map system or vice versa. You will most likely not get an exact fit, but it will give you a pretty good idea how the land parcels will be affected and size of the affected lots. You will get accurate sizes when the land surveyor goes out to the field survey each parcel that needs to be acquired.
Setting control points along the route and checking them with a Total Station will be good to give some level of confidence but once the GPS survey is done properly it will work out fine.
Dear Mr. Shankar,
You describe a project with several issues to be addressed.
The problem you presented is
There are many places there to trip. All three steps involve converting data to and from UTM coordinates (may only be LIKE UTM coordinates). A one-point localization runs the risk of not matching the scale and orientation of the system you wish to match. a minimum of two points with three preferred are needed to locate, scale, and rotate between systems with a minimum amount of redundancy.
If you are not sure, find someone experienced with datum transformations as protection from making a serious error.
Many thanks for the reply.
I wish to clarify a bit more on why the process I mentioned was followed.
There are two different survey teams in the project.
One team did the railway corridor alignment design - It includes establishment of DGPS control points, Topographical survey, design of the proposed railway corridor alignment and final location survey (where the corridor is pegged on the ground). The topographical survey, design, and final location survey used Ground coordinates.
A second team was given the task of cad astral mapping, which includes geo-referencing of cadastral maps, superimposing the proposed alignment on the cadastral map, and preparation of land parcel schedule. The land parcel schedule has details such as survey number, proposed area for acquisition etc. The cadastral mapping was done using UTM coordinates. The cadastral mapping team was not aware that the rail corridor design, topographical survey, and final location survey was done using ground coordinates.
I was part of the second team, and our team did all the mapping with UTM coordinates.
The Team A later reconverted the ground coordinates to grid (UTM), but they don't match the true UTM coordinates beyond 10km.
I read somewhere that the effect of localization (when on point is used as a base reference for creation of ground coordinates) is not more than beyond 10 km (6 miles).
Also, I noticed JAC replied that " A one-point localization runs the risk of not matching the scale and orientation of the system you wish to match. a minimum of two points with three preferred are needed to locate, scale, and rotate between systems with a minimum amount of redundancy".
I believe the disconnect between your work coordinated in UTM and the conversion by others.
Converting between ground coordinates and a plane projection (grid) is more involved than simply applying a scale factor. One must correct for azimuth orientation because all norths on a grid (useful grids) are parallel, which is not the case on the globe where they converge. Perhaps the most significant step that is overlooked is identifying the origin.
In most software I've seen, there is a place to enter an origin, however, it is happy if no entry is made (overlooked). This usually puts the origin from which to apply the scale is at zero and not relevant to your survey area. What is needed are already known grid coordinates of one or more points to be transformed. Then the scaling will be applied with an appropriate origin.
Localization accomplishes this when several points in the project have known coordinates in the desired projection. The scaling will be a little different than one calculated because both the "known" coordinates and the newly measured ones contain errors that will be included in the localization.
Bottom line: Find as many points observed in both surveys and use them to do a localization from one survey to the other.
I hope this helps. One "quick and dirty" check might be to try simply translating the questionable survey to the trusted survey. It may not produce final results but it may reveal what needs to be done to make it rigorous.
Dear Mr. Ravi
It is the right way for a linear project to joint all the control points together if they are in the same zone. You have also use Scale factor without changing Original UTM co-ordinate.
Multiple control points can use.
Localization of points are only helpful in the layout after final designed