The Introduction for this project can be found below under the post Data Collection Proposal for the Hadleyville Cemetery, which was posted Wednesday, September 7th, 2016.
Study Area
Where is the study area located?
-The location of the cemetery is in the western part of Eau Claire County, in the state of Wisconsin. A locator map of where in Wisconsin this is can be found below in figure 1.2. The landscape around it is hilly with scattered tree clusters and farmland.
What time of the year was the data collected?
This data was collected in mid-September, which is the fall season for Wisconsin.
Methods
What combination of geospatial tools did the class to use in order to conduct the survey? Why?
- The class at first started out using a survey grade GPS unit to geocode the exact locations of each headstone, create an attribute table to record all legible information of the tombstones and finally to take pictures of each headstones. Others tools that were used included the Inspire, Rededge, and Phantom UAV drones. These UAV drone were able to take areal photos of the cemetery to be used as a base map. Lastly each group of students had a pen and note book that they used to create hard copies of all of the information on the headstones.
How did accuracy balance with the time involved to gather the data?
-Due to the large amount time it took to gather data by way of the GPS unit, the classes ended up not using it. In its place the areal photo taken by the Phantom had sufficient detail to make digitizing the each individual tombstone possible using ArcMaps. While this method saved hours of work the accuracy of each tombstone location decreased. The accuracy of the UAV photo is accurate up to centimeters but when it is switched to the raster format in ArcMaps the accuracy depends on how zoomed in on each tombstone the software was when the digitizing took place.
How was data recorded? List the different methods and state why a pure digital approach is not always best. What media types are being used for data collection? Formats?
- The Inspire drone used a camera to take pictures of the cemetery that would later be compiled into one image by use of the software Pix4DMapper. The Survey Grade GPS unit has the ability to compile a list of attributes for each point that was taken. The attributes included were name of the deceased, year born, year of death, and the legibility of the tombstone. This data would later be transformed over to ESRI's ArcMaps where it could be overlaid on the Inspire UAV imagery.
This method of data retrieval at times can be the best and quickest way to gather the data that was missing, however there is one very large drawback to this method, signal interference. The class ran into this problem after an hour of data collection. In the southwest corner of the cemetery there are a few very large trees that are located directly over a a handful of tombstones. This prohibited the Survey Grade GPS from getting a strong enough signal to record the locations of those particular tombstones. This foliage also created problems for the UAV drone. When the data from the drone was being compiled into one image the program was unable to create a proper image of the data. The foliage in that southwest corner created large shadows, due to the collection occurring at 4:00PM in the afternoon, which put the sun directly behind these trees. These shadows lead to the software program being unable to compile the data correctly. The imagery from the Inspire was very distorted in the that same corner.
Another issue that occurred in regards to the GPS unit was the time it took for each tombstone to be properly collect and have all of the attributes added. Each tombstone took three to four minutes to collect all of the data. Also before each group could get started the previous group had to show go through all of the steps to properly operate the GPS unit. Due to this set back the GPS unit was not used, in its place the areal photo taken by the phantom had sufficient detail to make digitizing the each individual tombstone possible using ArcMaps. When the digitizing took place a
What equipment failures occurred if any? What was done to remedy the situation?
- The equipment failure was covered in the previous paragraph. The steps the class took to remedy the stated problems was to first create a hard copy diagram of the cemetery and give it a standard coding system that the class would use. The system was simple, the row furthest to the west was row one, the row second closest to the west was row to and so forth. To find the exact location of a particular tombstone in each row a letter system was added to each row. So the tombstone in the southern most point in each row was labeled as "A" and the further north down the row a tombstone was the further down the alphabet it was labeled. So Row 1 tombstone 5 was labeled as "1E". The class used this method to label all of the tombstones in the cemetery along with the attributes of each tombstone. With these location of each tombstone the class will be able to digitize the each of the tombstones and add the attribute data to each location.
How was the hard copy data transferred towards use in the GIS?
- Each group of students collected a portion of the information on the tombstone. This data was compiled into a online google doc. This google doc contained a field with the ID values that were given to each tombstone. At the same time during the digitization process two fields were created one that will contain raster data. The other was a PointID field that contained the same ID values as the attribute information on the Google Doc. Once the digitization was done the table could be joined to the points by the PointID fields. The raster field allows for images to be stored for each tombstone, figure 1.1 shows an image of the attributes table.
A couple issues did occur during the normalization of the data. This biggest was whether or not each group counted a family memorial stone as a tombstone or not. Those that did had extra tombstones on there sheets and those that did not were lacking those memorial stones. It took a combined effort of every member going through their notes to create an accurate count of tombstones. In retrospect it would have been a good idea to get a total count of all tombstones before we began the project.
The last major issue the class ran into was getting pictures for each tombstone. A website was found that had the name and an image of all tombstones on it. The website is linked below in this post.
How did you combine the UAV data with your survey data?
- The UAV data was the base map that and allowed for the digitization of each tombstone.
Results
(Figure 1.1: Show an image of the attribute table for the tombstones for the Hadleyville Cemetery. In the middle of the table is an image of the tombstone. An image of each tombstone is attached to each feature in this feature class.)
(Figure 1.2: Is the final map of the Hadleyville Cemetery. In the bottom left hand corner of this map is the locator map that shows where in Wisconsin Eau Claire County is as well as the location of Hadleyville Cemetery.)
How did data collection methods transfer into time spent creating the GIS? What was done to remedy the situation?
- The most time consuming parts of this project were the collecting of Tombstones with the GPS unit, which was remedied by digitizing. The second most time consuming issue was compiling every persons individual data into one collective dataset. This was accomplished by using Google Docs which allows many people to work on a single document at the same time. Errors for this project for the most part lie in not having perfectly legible tombstones and not being perfectly accurate with digitizing. There is little that can be done to rectify the legibility of a tombstone but the digitizing can be as good as possible by taking a little extra time to zoom in on each individual tombstone.
What might have been done to facilitate data collection in terms of equipment and refining the method?
- The only way to make this better would have been to spend more time as a class to figure out how to collect this data as well as what data should be collected. For example had the class broken up the cemetery into sections and assigned every group a section to collect there would be less chance of missing data or collecting the same data. Another thing that could have made this easier would have been to clarify earlier on that this was a class project not just a group project.
Conclusion
How did the methods transfer to the overall objectives of the project?
- The methods did a very good job overall of meeting the objectives of this project. Every tombstone was collected and a database with all of the information of each tombstone is available and can be easily updated.
How did the mixed formats of data collection relate to the accuracy and expediency of the survey?
- There was very little of mixed formats. The UAV was collected by using the UTM Zone 15 and all of the digitization was done using the same coordinate system. The limitations in accuracy are on the skill of the technician to digitize the tombstones correctly and the accuracy of the UAV, which again is accurate down to the centimeter.
Are the potential sources of error negligible and does the final product overall provide something better than the original situation?
Yes, the errors are negligible and the county now has an excellent source of information on the graves of everyone buried in Hadleyville.
Describe the overall success of the survey, and speculate on how this GIS will be of use for continued record keeping.
This project was a success. The coordination could have been better earlier on but the final project is something to be proud of. This will be very easy to maintain and update in the future. Every time a new grave is added to this cemetery, simply add a new location for it by digitization and fill in the attributes for the new grave. The only draw back to this system is that if a grave is added in the middle of a row the PointIDs for that entire row will have to shift to accommodate the new grave. This is a minor problem that only takes minutes to correct.
No comments:
Post a Comment