MOSPREP

The data to mosaic.

Input images can vary in projection, spatial resolution, or both. When the projection varies, the default output projection will be that which occurs most commonly in the input imagery.

You can specify the value of the parameter by using any of the following options:

• A specific data-file name or folder path and file name; for example, irvine.pix or C:\data\irvine.pix.
• A specific file-name pattern or all of the files in a folder; for example, C:\data\*.pix or c:\data.
• A text file that contains references to multiple files with, optionally, different parameter settings per file; for example, C:\list.txt or C:\list.mfile.

• The file, containing multiple files and different parameter values, must have a .txt or .mfile file name extension.
• List each set of parameters on its own line in the given order.
• Specify string parameters in quotation marks.
• Do not specify numeric parameters in quotation marks.
• A quotation mark preceded by a backslash is used as part of the string; for example, 'ab\"c' is used as: ab"c.
• Use an exclamation mark (!) before writing a comment in the text file.
• Use semicolons to delimit the parameters.
• When you do not want to specify a value, leave the position between the semicolons empty (";;"). Alternatively, if there are no remaining parameters to be specified at the end of a line, you can leave it blank.
• Parameters in a text file may or may not be specified in the algorithm:
  • When a parameter is specified in the algorithm and in the text file, the value provided in the text file is used.
  • When a parameter is not specified in the text file, you can do so in the algorithm.
  • Each line in the text file must follow this format:

    <FILE>; [<NODATVAL>]

    For example, C:\Data\image.pix; 0 indicates that a zero will be used as the background for the image.pix file.

    You can also specify more than one NoData value. For example, C:\Data\image.pix; -32768, 0 indicates that -32768 will be used as the background for the first channel of the image.pix file, and zero (0) will be used for the second channel. You can, if necessary, skip a channel when specifying NoData so that all pixels in the channel are valid. For example, if you want to mosaic three images, each with different rules for NoData, specify your entry in the MFILE text file as follows:

    "c:\data\image1.pix"; -9999      ! all channels use -9999 as their NoData
    "c:\data\image2.pix"; 0, -9999   ! channel 1 uses 0 for its NoData and the remaining channels use -9999
    "c:\data\image3.pix"; ,0,-9999   ! channel 1 does not have a NoData (notice the comma before the zero), channel 2 uses 0 for its NoData and the remaining channels use -9999
                                            

The name of the mosaic project file (.mos) to create.

An accompanying folder, with the same base name, is also created. The folder contains auxiliary information that makes up the mosaic project created by the mosaic-preparation process. Unique image IDs are created automatically based on the names of the source files and listed in the project file.



You can open the output mosaic project in CATALYST Professional Mosaic Tool to enhance the mosaic and perform quality-control procedures.

The background value of the input images.

A pixel you designate as NoData is excluded from normalization, color balancing, and cutline generation. Images often have NoData defined in their metadata so this parameter can be defaulted.

When you specify a text file as input, you can specify the NoData value of an image in the file. The NoData value specified in the input text file takes precedence over all other sources. When a value is not specified in the text file, precedence is applied to the NO_DATA_VALUE metadata tag of the input image. If NO_DATA_VALUE metadata exists for an input channel, its value is used; otherwise, the NO_DATA_VALUE at the file level is used. Finally, if a NoData value is available in neither the text file nor the metadata, you can specify the value for this parameter.

1. Text file
2. Channel metadata
3. File metadata
4. NoData parameter value

• NONE: Source images are processed in the order in the text file specified for MFILE. The first image in the file is the first processed. When the value of MFILE includes a wildcard or is a folder, the images are processed in ascending alphabetical order.
• NEARESTCENTER, [<starting_image>]: When you specify a starting_image, it is processed first; otherwise, the image most central of the input scenes is processed first. Subsequent images are processed based on which image has its center closest to the center of the first image. The center point is determined based on the rectangular bounding box of each image.
• MAXINTERSECT, [<starting_image>]: When you specify a starting_image, it is processed first; otherwise, the image most central of the input scenes is processed first. Subsequent images are processed in turn, based on which image most intersects (overlaps or covers) the images processed already. Each calculation, such as determining overlap amounts, is based on the rectangular bounding box of each image.

• NONE: No adjustment of the pixel values of the source images occurs.
• HOTSPOT: Removes hot spots from the input images.

  A hot spot is a common distortion in aerial photographs and optical-satellite images. With aerial photographs, distortion is caused typically by solar reflections that appear circular (also known as lens flare). With optical-satellite images, distortion appears typically as a striped pattern.

  HOTSPOT normalizes the brightness over the image by fitting a Gaussian surface to the brightness values. Smaller spot reflections, such as from lakes, cars, or buildings are not removed.

• ADAPTIVE, [<image_percent>]: Balances the brightness and contrast by using a moving window. The method is well suited for use on images with a large, irregular bright-and-dark pattern that cannot be modeled to a Gaussian surface. Patterns that model to a Gaussian surface are better processed with HOTSPOT.

  ADAPTIVE adjusts the brightness and contrast over local areas, improving image detail, while reducing the bright-and-dark pattern over the entire image. The method applies an adaptive enhancement by using a moving window to calculate the adjustment for each pixel value.

  The size of the window affects the results of the filter. You specify the window size by entering the percentage value of the filter size; for example, to represent a 35-percent window, enter 35. If you do not specify a value for image_percent, a default value of 20, or 20 percent, is applied.

  ADAPTIVE calculates the mean and standard deviation of the gray levels in the window and adjusts the values to match the overall mean and standard deviation of the image. Mean adjusts the brightness and standard deviation adjusts the contrast.

• BUNDLE, [AUTOSCALE], [DODGE=MANY, [SAMPLE=<size>], [DIST=<distance>]]: Produces best-looking color-balancing results by using two complimentary phases.

  First, a global adjustment of the mean and sigma of each image is computed by using a "block-bundle" method between it and each of its overlapping images. The method makes large adjustments between images to make them more balanced.

  Second, a series of dodging points is determined automatically to make smaller local adjustments between pairs of images. The order of the input images is ignored, because all of the overlapping images are used to adjust one image.

  You can also specify the keyword AUTOSCALE to have all input images stretched linearly and have a common scale. This can improve the results when the dynamic ranges of various input images vary significantly. You can also specify the keyword DODGE=MANY to allow the creation of multiple dodging points for each edges across the whole project, rather than just one dodging point per edge. When DODGE=MANY is specified, you can also use the SAMPLE=<size> and DIST=<distance> keywords to further control the dodging points. To adjust the number of full-resolution pixels that are used to compute each dodging point specify a value for the sample <size>. To adjust the spacing between the dodging points specify a <distance> in the units of the project. When unspecified, a default <size> of 64x64 is used, while a <distance> is computed automatically. For example, BALSPEC="BUNDLE, DODGE=MANY, SAMPLE=128, DIST=1000" would result in dodging points being created every 1000 meters, with each point using the surrounding 128x128 pixels.
• OVERLAP: Bases color balancing on matching overlapped areas of the input images.

  The rasters must have a sufficient overlap (greater than 10 percent) between each pair of adjacent input rasters and have good-quality georeferencing. You can also specify a local (LOCLMASK) or global mask (GLOBMASK). The mask must have some pixels within the overlap area to affect the output.

• LUT, <LUT_segment_number>: Color balancing is determined by lookup table (LUT) segments stored previously from the source images.

  When you specify a LUT value, MOSPREP looks for one LUT segment for each image channel; for example, an RGB-color image being added to the mosaic must have three LUT segments per input file.

  LUT-segment numbers must be provided explicitly using LUT_segment_numbers. The first LUT is used for the first image channel, the second LUT for the second channel, and so on. All input images must have the specified LUT segments; otherwise, the process will stop and an error message is displayed.

• HISTOGRAM, [<match_area_multiple>], [<trim_percentage>]: Color balancing is based on matching the histogram of the input image to that of the target mosaic image, sequentially.

  The match_area_multiple value determines the amount of data in the mosaic file to use to compute the reference histogram.

  • When you specify a value of 1, the reference histogram is computed based on an area equal to the overlap between the incoming image and the output mosaic.
  • When you specify a value of 3, the reference histogram is computed based on an area equal to three times the overlap between the incoming image and the output mosaic.
• REFERENCE, <fili_ref>: Color balancing is based on matching source-image histograms with those of the specified reference image.

  The reference-image file is specified by fili_ref. The reference image need not be the same resolution or projection as the input images, provided it can be reprojected into the projection of the input image. That is, neither precise georeferencing nor overlap are required.

• NEIGHBORHOOD, [<meanb_adjustment>], [<outlier_correction>], [<outlier_threshold>]: Determines a set of model coefficients that change each image pixel based on the pixel values of the intersecting (neighboring) pixels, regardless of image order.
  The overall pixel value of each individual image is adjusted iteratively, such that the value is similar to those of its neighboring images.

  • meanb_adjustment, set to 3 by default, adjusts the overall tonal value of the image pixels in each overlapping region to minimize the outliers.
  • outlier_correction excludes extreme pixel values–outliers–from the computation of the coefficients. The default is YES.
  • outlier_threshold is for use only when the value of outlier_correction is YES. The default value is 1, which excludes from the computation of the coefficients pixel values beyond the mean difference value, plus one standard deviation.

  When you specify NEIGHBORHOOD, MOSPREP uses the default values of meanb_adjustment, outlier_correction, and outlier_threshold.

• NONE: Images are added to the mosaic as is and the pixel values are not adjusted

LOCLMASK = (NONE | BIT | VEC | <n>), (NONE | BIT | VEC | <n>)

• NONE: Do not exclude any pixels from the computation.
• BIT: Use when at least some of the source images have bitmap segments that define exclusion regions. When one of the source images contains multiple bitmaps, the last one is used. All image pixels covered by the bitmap are excluded from calculation.
• VEC: Use when at least some of the source images have vector segments that define exclusion regions. When one of the source images contains multiple vector segments, the last one is used. This vector segment must be a polygon, and all image pixels enclosed by the polygon are excluded from the calculation.
• <n>: Use the bitmap or vector segment of the specified number, <n>, as the mask.
  Note: If two segments share the same number, MOSPREP uses a bitmap segment over a vector segment to define the mask area.

• BIT, BIT: Uses the last-created bitmap for both color balancing and cutline generation.
• NONE, VEC: Uses all data for color balancing, but excludes data under polygons in the last-created vector segment from cutline generation.
• VEC, 3: Uses the last-created vector segment for color balancing, and uses segment 3 as a local cutline constraint.
• 4: Uses segment 4 to limit color balancing, but does not constrain cutlines.

GLOBFILE = (<cb_file> | NONE), (<cut_file>)

• <cb_file>: Path and file name of the mask file to use for global color balancing
• <cut_file>: Path and file name of the mask file to use for global cutline-avoidance areas
• NONE: Keyword used to specify no color balancing, but you can specify a cutline file; for example, NONE, <cut_file>

GLOBMASK = (-1 | <n>), (m)

• <n>: Defines the segment number that corresponds to the color-balancing mask (cb_file) in the file specified for GLOBMASK. This value must be specified as -1 if–and only if–no color balancing is defined (GLOBFILE = NONE).
• <m>: Defines the segment number that corresponds to the cutline-avoidance mask (cut_file) in the file specified by GLOBMASK.

• Specify an existing cutline-constraint file
• Automatically generate a cutline-constraint file using the classic method which is based on the center of each image
• Automatically generate a cutline-constraint file using a "strips" method which is based on the overlap between images

• Name of the vector (vector_file)
• Name of the attribute field that contains the image source (field_name)
• Number of the vector segment that contains the constraint polygon (segment_number)

1. the value of its file level metadata tag: SourceID, or if that tag does not exist, then
2. the base name, without the extension, of the input source image's file name.

• MINSQDIFF, [(<vector_file>, [<field_name>], [<segment_number>]) | (AUTOCONSTRAIN[=CLASSIC[,<thiessen_factor>]] | [=STRIPS[,<overlap_factor>]]), [VECTHIN=<thin_tolerance>]]: The minimum squared difference method. Well suited to to most mosaicking projects, this method, in most cases, produces the cleanest cutlines. The algorithm determines a cutline in each overlapped area between two adjacent images, with minimum squared differences of gray values at the same locations of the region in all image channels.
• MINDIFF, [(<vector_file>, [<field_name>], [<segment_number>]) | (AUTOCONSTRAIN[=CLASSIC[,<thiessen_factor>]] | [=STRIPS[,<overlap_factor>]]), [VECTHIN=<thin_tolerance>]]: The minimum difference method. With this method, the algorithm determines a cutline in each overlapped area between two adjacent images, with minimum differences of gray values at the same locations of the region.
• MINRELDIFF, [(<vector_file>, [<field_name>], [<segment_number>]) | (AUTOCONSTRAIN[=CLASSIC[,<thiessen_factor>]] | [=STRIPS[,<overlap_factor>]]), [VECTHIN=<thin_tolerance>]]: The minimum relative difference method. Similar to MINDIFF, this method provides better output when similar sections of data appear dissimilar in various images.
• EDGE, [(<vector_file>, [<field_name>], [<segment_number>]) | (AUTOCONSTRAIN[=CLASSIC[,<thiessen_factor>]] | [=STRIPS[,<overlap_factor>]]), [VECTHIN=<thin_tolerance>]]: The edge-feature method. This method provides better output in urban-area mosaicking or in images containing many linear features. The objective of the EDGE method is to avoid placing cutlines across linear features.
• MAXDATA, [(<vector_file>, [<field_name>], [<segment_number>]) | (AUTOCONSTRAIN[=CLASSIC[,<thiessen_factor>]] | [=STRIPS[,<overlap_factor>]]), [VECTHIN=<thin_tolerance>]]: The cutlines to be on the boundary of the real image pixels, meaning that NoData pixels are ignored when the image boundary is determined.
• IMPORT, <vector_file>, [<field_name>], [<segment_number>]: Uses the specified polygons as cutlines, exactly as they appear in the vector file.
• FILEEXT: The cutline to be the extents of the input image or images and will not exclude NoData pixels from cutline generation.

• ERODEEXT: shrinks each of the data-extents polygons by an amount determined automatically using the resolution of the mosaic preview. The polygons are buffered by the determined amount and retain only the area of the original polygon that is inside of the buffer boundary.
• ERODEEXT=<erosion_distance>: is similar to ERODEEXT, except uses the buffer distance specified via <erosion_distance>. The applicable units are those of the projection of your project. For example, to shrink the data-extents polygons of UTM data by 5 meters, enter CUTMTHD="MINSQDIFF, ERODEEXT=5".