Welcome to the Harris Geospatial documentation center. Here you will find reference guides and help documents.


  >  Docs Center  >  Crop Science  >  Find Developing Hotspots

Find Developing Hotspots

Find Developing Hotspots

Use Find Developing Hotspots to look for anomalies between two rasters of the same spatial extent, from different dates. The result is a classification image that shows statistically significant differences during the intervening time period. Areas that change at an average rate are suppressed, while areas that change at a higher rate (positive or negative) are highlighted. For agricultural studies, this can be useful for identifying areas that are growing faster or slower, relative to the entire image. Or, it can reveal areas that were damaged by events such as hail storms and drought.

Two tools are available: The Find Developing Hotspots tool accepts two single-band rasters from different dates (Time 1 and Time 2). The Find Developing Hotspots with Spectral Index tool accepts two multi-band images from different dates, and it requires you to choose a spectral index to compute for each image.

These tools are designed for analyzing individual fields. You should spatially subset the Time 1 and Time 2 images to only include your area of interest, or mask out the rest of the images outside of the field extent. Then ensure that they are co-registered. See Example of Subsetting by ROI in the Find Hotspots topic for instructions.

This tool is part of ENVI Crop Science, which requires a separate license and installation. You can also write a script to find developing hotspots using ENVIAgFindDevelopingHotspotsTask or ENVIAgFindDevelopingHotspotsWithSpectralIndexTask.

See the following sections:

Background


The Find Developing Hotspot tools compute Z-scores using Getis-Ord Gi* statistics (Getis and Ord, 1992) on two co-registered images from different dates. See the Find Hotspots topic for more information about Getis-Ord Gi* statistics and hotspot analysis.

The second Getis-Ord Gi* image (Time 2) is subtracted from the first Getis-Ord Gi* image (Time 1). The result is a classification image that shows changes that are much bigger or lower than the average change between the two images.

For example, suppose that the Time 1 and Time 2 images represent a variable such as Normalized Difference Vegetation Index (NDVI). The following plots provide a graphical example. The difference in the global mean image value is plotted between the two time periods. For the developing hotspot image, each pixel represents how much the local mean changes, relative to the global mean. Their absolute position on the y-axis does not matter; the rate of change is of interest.

In the following plot, if the local mean of the variable for a given pixel changes at the same rate as the global mean, that pixel is considered neutral and is colored light grey in the developing hotspot image. This plot shows two grey lines, one for each pixel.

In the next plot, three different pixels show increasingly negative changes ("---" the most negative to "-" the least negative), relative to the global mean difference of the variable.

In the final plot, three different pixels show increasingly positive changes ("+" the least positive to "+++" the most positive), relative to the global mean difference of the variable.

Pixels that show highly negative or positive changes indicate areas in a field that may be growing or dying at a faster rate, relative to the entire field.

Run the Tools


Choose one of the following options:

Find Developing Hotspots

Use this option to find developing hotspots using a single-band raster, without specifying a spectral index image.

  1. From the Toolbox, select Crop Science > Find Developing Hotspots.
  2. Select an Input Time 1 Raster and Input Time 2 Raster. These are single-band images from different dates.
  3. Enter a Distance value, which is the radius of the area around each pixel used to calculate the local mean. If the input raster has a valid spatial reference, then the radius is measured in meters. Otherwise, it is measured in pixels. The radius determines the spatial resolution over which the clustering occurs. A good starting point is the pixel size of the input image.
  4. Enter a Threshold value. This is the first of three standard deviations from the mean difference used to create a color slice classification of Z-score differences. The default value is 3 standard deviations. Lower values reveal more hotspots, while higher values reveal fewer hotspots.
  5. Click the Browse button next to Output Raster, and select an output filename and directory.
  6. Enable the Preview check box to see a preview of the settings before you click OK to process the data. The preview is calculated only on the area in the Image window and uses the resolution level at which you are viewing the image.
  7. Select the Display result option to display the output image when processing is complete.
  8. Click OK. ENVI adds the resulting output to the Data Manager and Layer Manager, and it displays the output in the Image window.

Find Developing Hotspots with Spectral Index

Use this option to find developing hotspots by creating a spectral index image.

Data should be in units of reflectance prior to creating spectral indices. See the Spectral Indices topic in ENVI Help for more information.

  1. From the Toolbox, select Crop Science > Find Developing Hotspots with Spectral Index. The Query Spectral Indices for Two Time Rasters dialog appears.
  2. Select an Input Time 1 Raster and Input Time 2 Raster. These should be multispectral files with defined wavelengths.
  3. Click OK. The Find Developing Hotspots with Spectral Index dialog appears.
  4. From the Spectral Index drop-down list, select a spectral index to compute for the Time 1 and Time 2 rasters. The list includes only those indices that can be computed on both rasters.
  5. Enter a Distance value, which is the radius of the area around each pixel used to calculate the local mean. If the input raster has a valid spatial reference, then the radius is measured in meters. Otherwise, it is measured in pixels. The radius determines the spatial resolution over which the clustering occurs. A good starting point is the pixel size of the input image.
  6. Enter a Threshold value. This is the first of three standard deviations from the mean difference used to create a color slice classification of Z-score differences. The default value is 3 standard deviations. Lower values reveal more hotspots, while higher values reveal fewer hotspots.
  7. Click the Browse button next to Output Raster, and select an output filename and directory.
  8. Enable the Preview check box to see a preview of the settings before you click OK to process the data. The preview is calculated only on the area in the Image window and uses the resolution level at which you are viewing the image.
  9. Select the Display result option to display the output image when processing is complete.
  10. Click OK. ENVI adds the resulting output to the Data Manager and Layer Manager, and it displays the output in the Image window.

Example


The following example uses the Find Developing Hotspots with Spectral Index tool with two PlanetScope images, two weeks apart. The source images are four-band orthorectified tiles (4-meter resolution), courtesy of Planet, Inc. A circular soybean field was chosen for analysis.

The following steps were taken to pre-process each image:

  1. Convert the pixel values to apparent surface reflectance, using the QUick Atmospheric Correction (QUAC®) tool. This tool is available with an ENVI Atmospheric Correction Module license.
  2. Define a circular region of interest (ROI) around the soybean field.
  3. Mask the image so that pixels outside of the field have values of NoData.
  4. Spatially subset the image to the extent of the ROI.

Here is a side-by-side comparison of the two images (displayed in color infrared) after they were processed:

Next, the Find Developing Hotspot with Spectral Index tool was run, using the following values:

  • Input Time 1 Raster: July 2 image
  • Input Time 2 Raster: July 17 image
  • Spectral Index: Normalized Difference Vegetation Index
  • Search Distance: 4 meters (the spatial resolution of the PlanetScope images)
  • Threshold: 3 standard deviations

When the tool runs, ENVI creates an image showing the difference in Z-scores between the July 2 and July 17 images. It applies a color slice to the image to create color gradients for specific Z-score differences, based on the Threshold parameter. In the Layer Manager, these gradients are labeled with plus and minus symbols to represent larger positive and larger negative Z-score differences, respectively.

Neutral pixels are colored light grey. These pixels represent an average rate of growth within the field. Pixels that show a healthier than average change (based on NDVI) are colored green. Red pixels indicate areas that have become less healthy than average during the two-week period.

References


Getis, A., and J. K. Ord. "The Analysis of Spatial Association by Use of Distance Statistics." Geographic Analysis 24, no. 3 (1992): 189-206.

Mitchell, A. The ESRI Guide to GIS Analysis, Volume 2: Spatial Measurements and Statistics. Esri Press, 2005.

Ord, J. K., and A. Getis. "Local Spatial Autocorrelation Statistics: Distributional Issues and an Application." Geographical Analysis 27, no. 4 (1995): 286-306.

See Also


Find Hotspots

 



© 2019 Harris Geospatial Solutions, Inc. |  Legal
My Account    |    Store    |    Contact Us