Sunday, February 10, 2013

Assignment 2: Refining your terrain survey




Introduction:

This assignment was about uploading our x,y,z data, analyzing it and seeing where improvements could be made. We were to look at how our data came out and then go back out and recollect new data that would better represent our terrain. This exercise was meant to show us that there is always room for improvement and that it is important to always revaluate your data.

Methods:

We first loaded our x,y,z coordinates into ArcMap. From there we made five different interpolation models form the 3D Analyst toolbox of our terrain. These models are IDW, Natural Neighbors, Kriging, Spline, and TIN. Figures 1-5 show the results of these models using ArcScene. With the TIN we converted it to a raster before importing it to ArcScene. After looking at all the different models that we had, our group decided that Spline was the best model for displaying our data.
Figure 1: This is our IDW model of the original terrain model
Figure 2: This is our Nearest Neighbor model for the original terrain data
Figure 3: This is our Kriging model for the original terrain data
Figure 4: This is our Spline model for the original terrain data
Figure 5: This is our TIN model for the original terrain data


Our first terrain model we made out of snow and came back the next day to take our points. However we ran into the problem of the snow melting between the time of creation and the time of collecting points. To get around that, this time we created the terrain then collected the points. Due to the Wisconsin weather, our original model not only melted but was also covered up by about 2in of new snow. This forced us to build a new terrain. We kept our new terrain close to the old one so we could do a comparison.   

Our dimensions for the planter box are 100cm x 230cm. The origin of our planter box is in the lower left hand corner. The grid size is 5cm x 5cm. We tacked on string along the y-axis every 5cm. For the x-axis we put down 2 measuring tapes and had a stick that straddled the box so we could get the points (Figure 6). Me and Joel were down on the ground collecting the data points with Kent upstairs entering the points onto an Excel sheet. I would measure call out the coordinate and depth to Joel, who would tell data to Kent over the phone(Figure 8).
After we collected all the points from our new terrain we imported them into ArcMap and ran the Spline interpolation model on it. As a group we decided that the Spline model worked best for our data. We imported the Spline into ArcScene to show it in 3D model (Figure 9).
Figure 6: Joel waiting to take more points after a brief run inside to warm up
Figure 7: Our method for taking the data points.

Figure 8: Kent entering our data points into Excel
Figure 9: Our final model with the Spline interpolation

Discussion:




Figure 11: Our first terrain.
We ran into some issues with our first model with the cell size that we used. Our first terrain had a cell size of 10cm x 10cm. This turned out to be too large to accurately represent our data. There was a river on the left side of the box running up and down. However it did not show up very well in our original model. There was also the problem that most of our terrain melted the first time around. To avoid that problem the second time, we built the model then immediately took the data points. (Figures 10-11)
Figure 10: Our second terrain.

Figure 12: Our data points from the first terrain
Figure 13: Our data points from the second terrain
Another change we made was the first time Kent wrote down the points on a sheet of paper then transferred them to Excel. This took about half hour; this was with only nearly 300 points (Figure 12). We ran into the problem that as the data was transferred from paper to computer errors would be introduced. Several points were entered wrong and we had to go back and correct them. This time we were going to do a grid with 5cm x 5cm cells, so we decided that it would be easier and faster to just enter the points directly into the computer instead of transferring them between paper and computer. This turned out to be a good idea because in the end we ended up with nearly 1000 points (Figure 13).


Of the interpolation methods we used TIN (Figure 5) was by far the worst. It was blocky and a poor representation. Kriging (Figure 3) was another model that didn’t correctly model our data. The way it showed the difference in elevation was not very smooth. IDW (Figure 1) and Nearest Neighbor (Figure 2) were smoother but you could still tell were the cells were. Spline (Figure 4) was the smoothest of the models.

Results

This was a good exercise. It showed us how to do a survey from start to finish. We had to collect the data points then bring them into ArcMap and run some sort of interpolation on them. This exercise showed a good range of the different interpolation models that one could use. The only down side was the weather. But that is just the way field work is, one has to learn to adjust to the unpredictability of weather. Our group worked well with each other. We all had a job to do and accomplished our task.

No comments:

Post a Comment