The Atcor Spectra Plotting window is linked to the image you select in the first step of the ATCOR - Ground Reflectance wizard.
With ATCOR spectra plotting you can apply various atmospheric settings to compare the reflectance signature produced by each variation of the settings for a given spectra. A spectrum can be composed of a single pixel or a weighted average of several pixels. Comparisons of similar features can be made between the target image and library signatures or to a reference reflectance image. By comparing the target image and reference data, you can deduce the most accurate atmospheric model and condition so that you can adjust the atmospheric settings that give the closest match between the image signature and the corresponding library signature or reference reflectance image.
Behavior of the graph
The graph that appears in the ATCOR Spectra Plotting window shows the spectra response in computed wavelength even though the sampled input image is in radiance. The nodes on the graph for target spectra are displayed hollow circles (○). The nodes on the graph for reference spectra, whether from a library file or an image, are displayed as filled diamonds (◆).
Plotting target ground reflectance
Regardless of whether you are plotting target or reference spectra, the current position of the cursor on the input ATCOR image is plotted on the graph as a black line. When you move the cursor, the graph is updated automatically with the new values. To display a fixed target, on the Target tab of the Spectra Reflectance Plot window, create a new spectrum, and then draw a point on the input (target) image.
A spectrum can consist of one or more pixels derived from one or more points, lines or polygons. The weighted average is computed for each pixel under a shape in the spectrum. Each time you add or edit a point, or change the atmospheric parameters, the plot is updated.
Reflectance values are calculated internally for each wavelength of the spectrum, based on the ATCOR settings. The values are then plotted on the graph with the middle wavelength of each band plotted on the x-axis and the reflectance measurement for that band plotted on the y-axis.
Plotting reference ground reflectance
Reference spectra can be derived from a spectral-library file, a reading from a spectrometer file, or from a surface-reflectance image atmospherically corrected and validated previously. Displaying reference spectra in the graph will depend on the source for the reference data.
| Reference image | To plot spectra samples from a reference image, first create a new spectrum. On the Editing toolbar in the Focus window, click the arrow beside New Shapes, select Point, Line, or Polygon, as applicable, and then draw features, as necessary.
The samples you collect should correspond to the same type of samples you collected for the target spectra (water, asphalt, agriculture, and so forth). |
| Reference spectral-library samples | To plot spectra samples from a spectral library, on the Reference tab, click Import Spectra. If the target image is already defined, the most appropriate library will be selected. You can also select an alternate file, if necessary. In the Open Reference Spectra window, select one or more samples from the list, and then click Add to Plot.
Each unique spectrum in the library file is added to its own spectrum record in the Reference spectra table. |
Comparing with library spectra
An important part of testing for the best atmospheric parameters is to compare the signatures computed from different pixels in the image (samples) with the library spectra of a similar feature. For example, you can compare the reflectance signature computed from a pixel of a pine tree in an image with the library signature for pine trees, and then test various atmospheric settings to determine the best match. By sampling a variety of features (pixels), you can determine whether you have the best match.
- In the Spectra Reflectance Plot window, create a new target spectra to sample.
- In the Focus window, on the Maps tab, draw a polygon around the pixel you want to sample in the target image.
For example, to compare a pine tree, draw a polygon around a point in the tree.
- In the Spectra Reflectance Plot window, on the Reference tab, click Import spectra.
- In the Open Reference Spectra window, select a feature from the list, and then click Add to Plot.
The feature you selected appears in the Reference table and the spectra is plotted on the graph.
- In the table, select either the target or the reference spectra, and then link it to the corresponding spectra (of the same feature that you want to compare).
- Viewing both the graph and the regression statistics, in the Spectra Reflectance Plot window, under ATCOR, modify the settings until you can determine the best match.
- Repeat step 2 to add more pixels of the same feature type to compare with the sample.
- Repeat steps 1 to 6 to compare other features with other library samples.
Target vs. reference band regression
You can only perform regression analysis on linked spectra. A target spectrum can only be linked to a single reference spectrum (and vice versa). When a target and reference spectrum are linked, the regression statistics (gain, bias, and so forth) are calculated.
Linking target and reference spectra
In either the Target or Reference tables, select a spectra. A list of available spectra appears, in which you select the corresponding spectrum. When spectra are linked, the plot is updated, and the linked spectrum inherit the color characteristics of the defining spectrum. When you link target to reference, the target spectrum is the defining spectrum and the reference spectra inherits the qualities (color) of the target spectrum. When you link reference to target, the reference spectrum becomes the defining spectrum and the target spectrum inherits the qualities of the reference spectrum. When two spectra are linked, the regression statistics gain and bias are calculated. To calculate the regression slope, intercept, correlation coefficient (r) and coefficient of determination (r*2), you must select two or more pairs of linked spectra.
Interpreting the regression statistics
- The gain is calculated with the header gain applied.
A good match will approach a value of 1.
- The bias is calculated with the header bias applied.
A good match will approach a value of 0.
- The R2, the determination coefficient, should converge to a value of 1.
- The active gain is defined in the header for the selected band and used in the current computation.
- The active bias is defined in the header for the selected band and used in the current computation.
- The derived gain is the result of the computed regression for the selected band and is equal to 1 / active gain.
- The derived bias is the result of the computed regression for the selected band and is equal to the active bias less the regression bias.
- The regression slope for the selected band should approach a value of 1 for a good match.
- The regression intercept for the selected band should approach a value of 0 for a good match.
- The correlation coefficient (r) for the selected band should converge to a value of 1 for a good match.
Applying derived calibration coefficients
The derived calibration coefficients will update the value of Atmospheric Correction Sensor and Gain and Bias for all bands.
Goodness of "fit" of selected spectra
When linked spectra are selected in the Target or Reference table, they have an inherent quality of fit. The computed value represents the regression across all bands for all of the selected spectra. If no spectra are selected in either of the tables, all linked spectra are used in the computation. Nonlinked spectra are excluded.
Atcor settings
The graph is updated autmatically in real time with any change you make to the settings or pixel position of the cursor; that is, the curve in the graph is recalculated and replotted automatically.
Because the spectral plot assumes a constant atmosphere, output from an atmospheric correction that varies spatially will differ from the original. The same is true for the BRDF Correction and Terrain Reflectance parameters specified in the Advanced Options window.
- Aerosol type: The dominant atmospheric aerosols for a region. Select an appropriate value from the list, based on the dominant land-use type in and around the image: Rural, Desert, Urban, or Maritime.
- Atmospheric Condition: The dominant water-vapor column of the image, which is influenced mostly by the location and time of year. The Atcor Spectra Plotting wizard attempts to set the value based on the latitude and the date on which the image was captured. However, you may want to experiment by selecting other items in this list.
- Constant height: Even when you supply a digital elevation model (DEM), the calculations for reflectance are based on a flat region at the height defined by the height you specify. When working with regions of rugged terrain, enter a height that is as close as possible to the actual elevation of the pixel you are sampling, based on the DEM.
- Visibility: The horizontal visibility on the ground at time of image acquisition. This value defines the aerosol optical depth (AOD) of the atmosphere. A thick atmosphere (low visibility) means you can expect more scattering and absorption. With a higher value, you can expect less.
Comparing with library spectra
An important part of testing for the best atmospheric parameters is to compare the signatures computed from different pixels in the image (samples) with the library spectra of a similar feature. For example, you can compare the reflectance signature computed from a pixel of a pine tree in an image with the library signature for pine trees, and then test various atmospheric settings to determine the best match. You can establish confidence that you have the best match by sampling a variety of features (pixels).
- In CATALYST Professional Focus, click a pixel of the feature you want to sample.
For example, to compare a pine tree, click a pixel in the tree.
- In the Atcor Spectra Plotting window, click From Spectra File.
- In the Open Spectra window, select a feature from the list, and then click Add To Plot.
The feature you selected appears in the Displayed spectra table.
- Under Atcor, adjust the settings, as necessary, until you can determine the best match.
- Repeat steps 1 to 4 for other pixels you want to compare with the corresponding feature.