Lab 3 - Model Builder & Python

2.1 The Model

The portion of the image of interest (this was arbitrary for the current analysis) was framed in the display. Under the "Environments" tab in Model Builder, "Processing Extent" was set to "Same as Display", which allowed us to select only the pixels contained in our display for analysis. The "Times" tool, under the "Spatial Analyst -> Math" toolbox was used to simulate treatment of the image. 1 was selected as the constant with which to multiply the input TIF. This created a subset with the same values as the original TIF, covering an extent equivalent to the display extent we specified. The model is presented in Figure 2, its output in Figure 3.

2.2 Exporting Python Script

Once the model was executable, the code underlying the model tools was exported, using Model -> Export -> To Python Script on the Model Builder main menu. The script was opened in the Notepad++ text editor. Figure 4 shows the script as it was exported from Model Builder. For a more detailed image of the script, see Figure 15. The general logic for approaching this problem is as follows:

• Ensure that the Spatial Analyst License is activated:

• Set extents (in python, called Geo-processing environments):

• Allow the program to overwrite existing files using:

• Establish local variables with clear names:

• Apply the Spatial Analyst Times tool:

2.2.1 Programming in Python

The next portion of the lab focused on basic programming skills. This included an introduction to:

• variables
• lists
• loops
• branches in logic
• opening and writing to a file

These will not be described in detail here, but their application to the problem addressed in this exercise will become evident as it is further explored.

3.0 Classification Using Python

The skills applied to extract the portion of the image above are now applied to "slice" the image into 5 classification levels, using the "Natural Breaks" algorithm. First, the model was prepared using Model Builder. Figure 5 shows the model, and Figure 6 shows the output when the Slice Model is run on the B10 layer of the LANDSAT image file.

3.1 Code Logic

The Python script developed for the single file was exported from Model Builder, and then adjusted to accommodate all seven files. First, a list is created containing the locations of all seven files. A file is opened to store Raster Data, specifically Standard Deviation (STD) and Mean values. A For Loop is developed to iterate through the files and perform the specified action, which was, in this case, "slicing" the provided rasters into 5 segments using the Natural Breaks algorithm. Inside the For Loop, new files are created to store the output of the GIS process, the "Slice" is executed, the STD and Mean Raster data are extracted, and written to the RasterData.txt file. A sample of the data is displayed below, in Figure 7. At this point, we exit the for loop and close the file. The raster output produced when running the code a single time is shown in Figures 8-14. More specifics on the approach to the problem can be found in the comment section of the code, seen in Figure 16.

3.2 Post Lab Thoughts

The code in its current iteration is quite inflexible, as the file names have to be manually input by the user. A more robust and flexible code would access the folder where the files are held and iterate through the files automatically.

Ideas from others:

• set workspace to folder: arcpy.env.workspace = "filepath to folder with image files"
• using "L5018024_*" allows for any number of characters after the 1st, added flexibility
• use ListRasters - arcpy.ListRasters("L5018024_*", "TIF") -> makes a list of everything in the folder (rasters only), specifically .TIF files in this case