Differential Correction (GPS and GIS) Part 4

PROJECT 4F—The OK Campus

More GPS Files and DOQs

In this Project we verify again that differential correction has great value—here we superimpose a GPS trace, on a digital orthophoto, made by walking.

{__ } Make a Pathfinder Office project named UK_Campus_yis.

Specify its directory to be_:\GPS2GIS\UK_Campus_yis.

{_} Use the operating system to copy the contents of_: \GPS2GIS\UK_Campus to __:\GPS2GIS\UK_Campus_yis by doing the following:

Minimize the entire Pathfinder Office window. Open two windows showing the contents of both folders UK_Campus and UK_Campus_yis. The latter should be empty except for folders labeled "backup," "base," and "export." Click in the title bar of (UK_Campus. Hold Ctrl and press the "A" key to select all folders and files. Hold down the CTRL key and press the "C" key to copy the highlighted files and directories onto the operating system clipboard. Click the title bar of (UK_Campus_yis, hold down Ctrl, and press the "V" key to paste the contents of the clipboard into that folder. Say Yes to any complaints that folders already exist.

The directory __:\GPS2GIS\UK_Campus_yis should now contain at least the following:

• the rover file r020919b.ssf

• the doqqqq image file J41se8w.tif and its world file J41se8w.tfw


• a directory "base" that contains U0020919.exe and L0020919.ssf

• two empty directories, "export"17 and "backup"

The file r020919b.ssf was made by a student carrying a GeoExplorer on the University of Kentucky campus, from a point almost on South Lime-stone Street, along a sidewalk that traverses the front lawn (which contains the McVey monument), past the Administration building, across the southwest border of the plaza in front of the main faculty office building, along a sidewalk between two major library buildings (which blocked the signal in that area), and past the president’s home. He turned right on Rose Street, traveling southwest until coming to a main UK entrance where he turned right again into a parking area. Another right turn led him to intersect his earlier path. (If you are interested in an Internet map of the place, the coordinates are those of the McVey monument: 38*02’20.57581"N and 84*30’18. 67673"W.)

The files j41se8w.tif & j41se8w.tfw constitute a digital orthophoto of a portion of the UK campus.

The files in the directory "base" (U0020919.exe and L0020919.ssf) are files downloaded from two Trimble Community Base Stations in Lexington. Either of them could be used to adjust the points in r020919b.ssf to make r020919b.cor.

{_} Fix up the parameters: In Pathfinder Office, under Options, make the Distance units U.S. Survey Feet. For Coordinate System, pick U.S. State Plane 1983. The Zone should be Kentucky North 1601. Coordinate Units should be U.S. Survey Feet; the Altitude Units should be Survey Feet as well. Make the Time Zone Eastern Std USA.

{_} Open and display the waypoint file McVey_yis.wpt file. (You may have to browse for it. It’s in_:\GPS2GIS\McVey_NGS. You may also have to zoom to the full extent of the open feature and background files. Close any other than the waypoint.)

{_} Display the file j41se8w.tif in the background (after removing any other files in the Load Background Files window).

{__ } Bring up the Map and make it full screen. Open the Position Properties window. The Map Window may shrink. Resize the Map window and move the Position Properties window off to the side so that both windows are visible and not overlapping. Zoom to Extents. Open and display the rover file r020919b.ssf.

{__ } Make the Not In Feature layer be displayed with heavy red dots; the dots should not be joined.

{_} Zoom up on the beginning of GPS track. It is pretty hard to see what the path was. Using the Position Properties window, look at the location of the First fix. Click the DOPs tab. Move several positions along the track using the > button. Note that when the path takes a sudden jump a different set of four "Satellites" is used to compute the position. This shows that the error produced by one set of satellites may be quite different from the error produced by another set.

{__ } Note the position of the McVey monument relative to the GPS track.

Correct the Rover File

{__ } Click on the button on the Start bar that will bring up the window of _:\GPS2GIS\UK_Campus_yis. (Alternatively you could click on the project folder on the Project Toolbar.) Pounce on the "base" folder to bring up the list of files in __:\GPS2GIS \UK_Campus_yis\base. Pounce on U0020919.exe to have the computer execute it. You should shortly see the appearance of U0020919.ssf. The self-extracting (.exe) base file just created an .ssf file. Bring back Pathfinder Office if you minimized it.

{__} Differentially correct r020919b.ssf using U0020919.ssf: Since you have two base station files, you can ask Pathfinder Office to choose one of them based on the Preferred Base File Prefix. Choose U. Then when you get the screen showing the Reference Position of the base station, write down the Latitude, Longitude, and Altitude here:

Lat:_Lon:_Alt:_

Complete the process, making sure the resulting COR file goes into __:\GPS2GIS\UK_Campus_yis.

{_} Display r020919b.cor. It should look a lot more reasonable.

Compare the path with the description of the GPS track given above. Use zoom and pan (including the Auto-pan to Selection feature if you want to move along the path using the Position Properties window) to look at how closely the track follows the sidewalks.

(This is a good place to notice a characteristic of orthophotographs. Early in the walk, as the GPS track crosses a plaza, you will notice a tall building to the northeast of the track. Zoom up on that building. You can see the southwest side of the building clearly. You may assume that, in reality, the top of the building is precisely over the base, though on the photo the top of this 18-story building appears to be displaced by some 80 feet horizontally. (Use Measure.) The lesson here is that digital orthophotos depict locations accurately only at ground level.)

{__ } Use the operating system to copy (Ctrl-C) U0020919.SSF that is in __:\GPS2GIS\UK_Campus_yis\base and paste (Ctrl-V) it into __:\GPS2GIS\UK_Campus_yis.

{__ } Open U0020919.SSF (in the Base folder of__:\GPS2GIS \UK_ Campus_yis) as a Background file. Even though it is a base station file it may be displayed (and differentially corrected as you will see shortly). Zoom to the full extent of all open and background files. Under View ~ Layers ~ Background display Not in Feature files with a thinness pink line. The building over which the fixes cluster is the Forestry Building, where the OK Community Base Station is located. Zoom in so that you can see this building and those surrounding it. Measure the span of the fixes in this file: ______ This will give you a good idea of the best accuracy that could be expected when selective availability was active.

Correct a Base Station File

{__ } Start the process of differentially correcting U0020919.ssf withL0020919.ssf (both in the "base" folder) to produce U0020919.corin the __:\GPS2GIS\UK_Campus_yis folder: In the Differential Correction window choose Settings to make certain that only corrected positions are placed in the output file. (Make sure that you specify the preferred base station file letter to be "L," not "U"; correcting a base station file with itself would be sort of cheating.) When finished with the differential correction process, display this file. Note that the points all fall within a circle of diameter of less than a foot, except for one outlier that pushes the span to about 1.3 feet. You will have to zoom out considerably to see the context of the fixes.

This exercise gives you an idea of how well differential correction can work. Basically, for using GPS for GIS, where accuracy is all-important, differential correction is vital. It is therefore relatively unimportant that selective availability was removed. It is hard to imagine a situation in which you would not differentially correct GPS files that are going to be used to build a spatial database.

{_} Use the grouping utility to make the file U0020919_pt.cor.

{__} In the __:\GPS2GIS\UK_Campus_yis folder you will find a waypoint file named UK_Community_base_Station.wpt. It consists of a waypoint named Calvin. Open that file and display the waypoint.

{__ } Repeat the process of correcting a base station file, but this time correct a file that was taken after selective availability was eliminated. Correct U0050219.SSF with L0050219.SSF.

{_} Change settings: Lat-Lon, WGS 1984, ddmmssss, meters all around, and HAE. By the way, changing the projection will force Pathfinder Office to unload the DOQ, since image files cannot be projected and maintain their integrity.

{__ } Use the record editor to derive statistics for the file U0020919. Compare the results with the location of the base station that you wrote down earlier.

Surveyed: Lat:_Lon:_Alt:_

Calculated: Lat:_Lon:_Alt:_

Differences:___

{__ } Move to the parent directory: __:\GPS2GIS \UK_Campus_yis. Delete the .tif file.

When one does a set of scientific experiments to determine the truth about "something" one usually wants to hold fixed ("constant") most of the things that could affect the outcome of the experiments while varying one or two other things. For example, if you were trying to determine a law for how long it took an object dropped from a height to hit the ground you might want to drop the object from different heights and record the duration of the fall. But you would always want to use the same object—not a billiard ball one time and a Ping-Pong ball the next. One of the problems in determining accuracies of GPS measurements is that it is impossible to repeat experiments under exactly the same conditions. Everything changes. The satellites move. The atmosphere changes composition. And so on. So absolute statements about GPS accuracy are hard to make honestly. One approach to determining GPS accuracy is to make lots of different measurements at different times and attempt to "get an idea" of "what you might expect." Not completely satisfying, but it’s the best we can do.

What follows is a miniexperiment, easily done, that will give you an idea of the best you could expect from differential correction. We take five base station files and correct them (with another base station) and look at how much improvement was made. Base station files are good SSF files to work with since (a) we know exactly where the antennas are, (b) they collect a lot of data (every three or five seconds), and (c) their antennas are mounted so that they get really good reception.

{__} Make a project that points to the folder __:\GPS2GIS \Compare_ ssf_cor. There you will find SSF files U0070707, U0080808, U0090910,18 U0101010, and U0111111. These are files generated since selective availability went off. They are about a month apart so they used different constellations. While the file names have a certain pattern to them the files themselves probably constitute a fairly random sample of DOPs and signal strengths. In the "base" folder are base station files that will correct most of the points of the SSF files.

{_} Differentially correct the five flies. You can do all five files at once. Start the differential correction process. Select the first SSF file by clicking on its name, then select the remaining four SSF files by holding down Ctrl and clicking. Ignore complaints that the files are not rover files. Conduct a Local Search for the base files (preferred base station prefix "L") and you will see all five files set up to be corrected at once. Approve everything and wait —there are lots of points for the computer to calculate.

{_} Group each of the SSF files into five point features:

You can again process all five files at the same time. Put the results into GROUPED.SSF.

{__} Open GROUPED.SSF. Represent the waypoint Calvin with a blue triangle. Represent a Point_generic feature as a red square. You may have to Zoom to Extents to get the waypoint included on the map. Measure some distances. This gives you a good idea of the best horizontal accuracy you can expect from uncorrected data.

{__ } Open the position properties window. Click on the waypoint and note its altitude (HAE). _. Now click on the other points and note their altitudes. (You can tell which is which by the date.) You now have an idea of the best vertical accuracy you can get from uncorrected data.

{__} Group each of the COR files into GROUPED.COR. Display and measure these. Note their altitudes.

{__} Now open and display both GROUPED.SSF and GROUPED.COR together. You should now have a pretty good idea of how much differential correction can improve excellent SSF files. In general, your SSF files will not be as good—so take that into consideration.

Look at a High Resolution Color DOQ

Sometimes, as you’ve seen, DOQs are in black, white, and grayscale. In this project you will see the results of part of the walk across the University of Kentucky campus overlaid on a high-resolution color orthophoto. The grayscale orthophotos you used previously have cells a little more than three feet on a side. The color orthophoto that you will load next uses 6 inch cells.

(Unless you have a computer with a fairly sizable memory and a good bit of free disk space, you may not be able to do this part of the project. The size of the grayscale orthophotos is somewhat over two megabytes. The color orthophoto covers somewhat less territory but has resolution about six times higher (which means about 36 times as many cells in a given area, and—because it is in color—requires many more bits per cell. Therefore, it will not surprise you that the color orthophoto is larger in terms of memory size. It may surprise you that it is about 50 times larger—102 megabytes compared to about 2.2 megabytes.)

{_} Change settings: U.S. Survey Feet everywhere. U.S. State Plane 1983. Kentucky North 1601. Eastern Standard.

You may proceed in either of the two ways, depending on the speed of your CD-ROM drive and disk space available. One is to load the large color orthophotos into PathFinder Office directly from the CD-ROM; the other is to copy the data onto the hard drive and load the images from there. If you choose the latter:

{__} Make a folder in _:\GPS2GIS named Temp_Color_DOQ.

With the operating system, copy the files NorthEast.tfw and NorthEast.tif from the CD-ROM folder UK_Color_Orthos into Temp_Color_DOQ.

{__ } Examine the color orthophoto __:\GPS2GIS\NorthEast.tif together with U0020919.SSF. Zoom up on the area where the GPS track crosses the image and the office tower appears to its northeast. Notice the shadow cast by the building. Notice how crisp everything appears compared to the grayscale orthophotos. This happens partly because of color, but mostly because the area covered by each cell of the grayscale orthophoto is now represented by more than 36 cells. If you zoom out you can easily pick out individual cars and trucks. The little dots making shadows on the sidewalks are people.

{__} Delete the folder Temp_Color_DOQ.

PROJECT 4G—DOP Matters

You must not assume that Differential Correction will take out all errors, or even all significant errors. In particular, multipath errors and errors caused by high PDOP can remain after the correction procedure has been applied. The file H092702B.SSF is an 11-hour file taken by a GeoExplorer receiver with the PDOP mask set to 99. File H092720B.COR is a file made by correcting four hours of the data of file H092720B.SSF.

{__ } In the folder __:\GPS2GIS\High_PDOP open the file H092720B .SSF. Display it, joining its points with a thin black line. Note that, even though selective availability was not active, some points are very wide of the mark—as much as 200 meters.

{__ } Check on the DOP values to the most far-ranging points at the ends of spikes. Some have very high PDOP values. Others, probably the result of multipath error, have reasonable PDOPs.

{__ } Open file H092720B.COR as a background file (it’s in the same folder). Set up View ~ Layers ~ Background so that Not in Feature files display as a thin red line. Note that the spikes are still there. In a couple of cases in appears that differential correction made things worse.

The moral of the story is that differential correction cannot compensate for errors caused by high PDOP or multipath.

PROJECT 4H—Your Data II

Correct Your Own Data

In this project you will correct the data you took in Projects 2A (fixed point) and 2B (a trace made by moving the antenna). Or you might choose to correct some other data. You will learn how to find and download data from a base station.

{__} Return to the Project_:\GPS2GIS\Data_yis. Examine the data you collected in Projects 2A and 2B. What are the last seven characters of the names of the base station files? You would need to differentially correct these files?19 Recall that the format of a base station file is XYMMDDHH.SSF.

It may take a little research to determine if there is a base station within 300 miles of your location but GPS equipment vendors can usually supply the information. When you find one, fill out the form at the end of this part. If it’s a Trimble compatible station you want what the company calls Community Base Stations. One way to find one is to point your web browser at:20 www.trimble.com/trs/trslist.htm or www.trimble.com/gis/cbs/ and you can select base stations by state, province, region, and/ or city. Information on how to retrieve the base station computer file can be found here also.

{_} Bring up a browser and check out the URLs above. Explore your geographic region. It may be obvious what base station is closest to you and how to download it. In any event, let’s look at a different method before we do any downloading. Keep your computer connected to the Internet.

A more elegant way to locate a nearby base station is provided directly by the Pathfinder Office software and a Trimble Navigation web site. Once you open an SSF file that you want corrected, you can direct the software to automatically download a list of base stations to your computer, sorted in the order of distance from the positions in your SSF file. Assuming you have the correct permissions from the base station owner, the software can even download the base station file and automatically produce the COR file.

{__ } Start the differential correction utility. Select one of your data files as the Rover File. Under Base Files, click "Internet Search." In the window that comes up, click New. In the New Provider window choose "Copy the most up-to-date etc." option. What happens now may depend on how you are connected to the Internet, but what can happen is that your computer will download a list of many dozens of base stations. (Please see Figure 4—13.) The ones closest to your SSF file’s location are listed first and, if you are on the North American continent, those in South Africa, Australia, and New Zealand listed last. If you pick a base station and choose Properties you have the option of going to their web page directly or sending them E-mail.

List of Base Stations from the Internet.

Figure 4-13. List of Base Stations from the Internet.

If you OK this window you may have the option of downloading the files you need directly. Or you may not, since there may be impediments—both technical and/or commercial.21

{__ } Using the information above, plus instructions from your teacher or employer, obtain the appropriate base station files. Such files are usually kept on hand at the base station locations for at least several months after they have been collected. As I’ve described, you may well be able to download these files from the web, or through the download capability of your browser through a process called "anonymous FTP" (File Transfer Protocol). Bring the files into the Base folder of the Project you set up for the data in 2A and 2B. If the files are EXE files, or "zipped" files, convert them into SSF files.

{__ } Use the differential correction process on the files you collected. Display the SSF and COR files with Pathfinder Office software.

Next post:

Previous post: