Introduction
This project is a follow up to the previous project, creation of a digital elevation surface. The task was to import the data collected from the sand box terrain as an excel file into ArcMap and ArcScene. Then to create a surface that best represents the data collected using an interpolation method. These methods include IDW, Natural Neighbors, Kriging, Spline, and TIN which will be explained in further detail about what which is later in the blog. The next task of this assignment was to reevaluate the data and possibly revisiting the sandbox and taking more data points (if any areas seem weak) to enhance the image.
Methods
After surveying the sandbox terrain and collecting data points the next step is import the excel file into ArcMap or ArcScene. The difference between the two is that ArcScene has the capability of viewing images in 3D. This project uses different elevation points making ArcScene very useful when converting the points into an image. The first step is to add the excel file into ArcScene then converting the excel file into a point feature class. This can be done by clicking file>add data> and filling out the rest correcting to fill needs. A step by step process of how to convert data into a point feature class can be found here. No coordinate system or units were used in the process because the data would become skewed and changed if unites were used. Points will then appear on ArcScene differentiating in elevation, now ready to be converted into different interpolation techniques. Using the ArcToolbox and various techniques under 3D analyst the point feature class was converted into the five techniques I mentioned earlier.
 |
| Figure 1: Image of the ArcToolbox in ArcScene 3D Analyst Tools opened |
This section will discuss the different interpolation techniques used on ArcScene to create surfaces to view the data elevation points. More on each technique can be found here on a ArcGIS help page.
IDW
IDW stands for Inverse Distance Weighted, it estimates cell values by averaging the values of sample data points in the neighborhood of each processing cell (ArcGIS Help). The IDW was created in ArcScene to show a 3D image to better represent the collection of data. 420 points were added to the number of points box when creating this tool from the tool box. 420 points were added because our group collected 420 points.
 |
| Figure 2: 3D IDW Red- high elevation, Blue- low elevation |
Natural Neighbors
This method finds the closest subset of input samples to a query point and applies weights to them based on proportionate areas to interpolate a value. Natural neighbor works by weighting each point by how close it is to other points in the same from the cell being used (ArcGIS Help). However, this time no points were added to the map because it was not required, similar to they style of TIN. This image was created in ArcMap and is a 2D representation of the data collected.
 |
| Figure 3: 2D image of Natural Neighbor Brown representing high points and Blue low points |
Kriging
Kriging is an advanced geostatistical procedure that generates an estimated surface from a scattered set of points with z-values. Kriging determines height by looking at each value compared to other values. I cut the the number of points in half and entered 210 for the number of points in the kriging creation box. 9 classes were used to represent the data green being the deepest and dark red being the highest elevation. This is a 3D model of the data created in ArcScene.
 |
| Figure 4: 3D Kriging Red-high points, Green- low points |
Spline
Spine uses an interpolation method that estimates values using a mathematical function that minimizes overall surface curvature, resulting in a smooth surface that passes exactly through the input points (ArcGIS Help). 420 points were also entered when creating this method making it similar to natural neighbor and IDW. Spline also shows cone shaped areas that were not present in our actual creation of the data. The creation in the snow was very smooth and flat in most areas.
 |
| Figure 5: 3D Spline flipped to show the other end of the creaton |
 |
| Figure 6: 3D Spline Red - high elevation Blue - low elevation |
TIN
TIN stands for Triangulated irregular network and uses Z values and cell centers to fully cover the perimeter of the surface (Arc Help). The TIN method displayed the data in a very unique way compared to the rest of the methods. Digital triangles are placed between the nearest data points connecting the triangles.
 |
| Figure 7: 3D TIN |
Discussion
After creating all five of the interpolation methods, I came to the conclusion that the kriging method best represents the data. The kriging method, figure 4, best represents the data because of the smoothness and the easiness to interpret the data. Compared to the other 4 methods, kriging has a smoothness and clearness about the image created. Only using 210 points compared to the others where 420 were used contributed to the best image being created. If I had been more educated in ArcGIS and ArcScene the outcomes could have been a lot different, but since I am still an amateur in this area most of the creations were not clean and hard to read. A big key in choosing the kriging method was how it represented the deepest part of the model. Looking through each figure it is easy to tell the bottom right corner is very deep and hard to understand. The kriging model best represents this corner with the smoothness it represents.
Figure 2, the IDW technique, was the worst out of the five when creating a continuous surface that displayed the data. Those cone shapes you can see in the image look very bad and it does represent the data in a good way at all. I believe that the image is very skewed and cone shaped because there were 420 points collected. If there were less points collected then the image would become more clear and possibly useful.
Natural Neighbor represented the data very well but was not my favorite. This is very similar to kriging but it does not appear as smooth as kriging because of some jaggedness in the surface when changing colors.
Kriging represent the data collected by my group the best. The continuous surface created looks very smooth as each color blends into each other making the image appear clean. I used 9 classes to represent the data. The bottom right corner where the data is the deepest it looks very clean and easy to read compared to the rest of the methods.
The spline method is very similar to natural neighbor as it represents the data in an un-smooth way. The area where the data is most deep is badly represented and difficult to read in a 3D form in ArcScene, making spline not as useful as some other techniques.
Figure 7, I really liked how deep the TIN data would appear in ArcScene, really displaying the different elevations that were present in the surface. However, I would not choose this method to best represent the data collected because sometimes the triangles do not represent the smoothness of data and show them as triangular shapes instead. The data created by our group did not represent any triangles when creating the model, making the TIN method hard to represent the data.
Conclusion
This field activity along with combing the week before was very fun and challenging to complete. It definently tested critically thinking skills to come up with creating a surface and a grid system to measure the elevation. The hardest thing for me was converting the data into continuous images on ArcGIS. I have never created any of these types other than TIN, with my inexperience with these methods I found the Arc help menu very useful to create the images. I thought the team I belong to did a fantastic job in completing the task in a efficient and successful way allowing the project to be completed without many troubles.
No comments:
Post a Comment