Mosaic from Orthos module

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 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: The source images are not resorted; rather, they are processed in the order provided.
• Nearest to Center: If an image is specified as a value for Starting Image, that image is processed first. If no starting image is specified, the image most central of the input scenes is processed first. Subsequent images are processed in turn, based on which has its center closest to the center of the first image. All calculations (determining the center point) are based on the rectangular bounding box of each image.
• Maximum Intersection: If an image is specified as a value for Starting Image, that image is processed first. If no starting image is specified, the image most central of all the input scenes is processed first. Subsequent images are processed in turn, based on which image most intersects (overlaps or covers) the images already processed. All calculations, such as determining overlap amounts, are based simply on the rectangular extents bounding box of each image.

• Hot Spot: Removes any hot spots from the input images. A hot spot is a common distortion in aerial photographs and optical satellite images. This distortion is typically the result of solar reflections that appear circular in photographs, and tends to appear as a striped pattern in optical satellite images.

  When Hot Spot is selected as the normalization method, hot spots are removed from the images by normalizing the brightness over the image by fitting a Gaussian surface to the brightness values. It does not remove the smaller spot reflections from lakes, cars, or buildings.

• Adaptive: Balances the brightness and contrast using a moving window. The adaptive filter is recommended for 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 handled by Hot Spot.

  The adaptive filter adjusts the brightness and contrast over local areas, thereby improving image detail, while reducing the bright-and-dark pattern over the entire image. It applies an adaptive enhancement using a moving window to calculate the adjustment for each pixel value. The filter calculates the mean and standard deviation of the gray levels within the window and adjusts the values to match the overall mean and standard deviation of the image. The mean is used to adjust the brightness and the standard deviation is used to adjust the contrast.

  Note: Because of the intensive background processing required, Adaptive normalization can be slow to produce results.
• None: Applies no normalization to the input images.

• Bundle: Typically produces the best-looking color-balancing results using two complimentary phases. First, a global adjustment of the mean and sigma of each image is computed using a "block-bundle" method between it and each of its overlapping images. This global adjustment produces large adjustments between images to make them more balanced with each other. Second, a series of dodging points is determined automatically to make smaller local adjustments between pairs of images. This technique ignores the order of the input images, because all of the overlapping images are used to adjust one image. Bundle is recommended for most images.
• Overlap: Performs a least-squares analysis in the overlap areas between images in the mosaic to determine the optimal radiometry for the final mosaic. When the mosaic is created, the computed transformation is then applied to each image as it is added to the mosaic.
• LUT: Uses previously stored lookup-table segments from the source images to control the color balancing. The module seeks out a LUT segment for each image channel. When this option is selected, the LUT-segment numbers must be specified for Color Balancing Specification Extra Options.
• Histogram: Performs color balancing based on matching the input image histograms to the target mosaic image, sequentially. This method does not require very precise georeferencing and overlap.
• Reference: Color balancing is based on matching source-image histograms with those of the specified reference images. When this option is selected, the reference image must be specified for Color Balancing Specification Extra Options.
• Neighborhood: Determines a set of color-balancing coefficients that change each image pixel based on the pixels values of the intersecting (neighboring) images, regardless of z-order. The overall pixel value of each individual image is adjusted iteratively, so that the value is similar to those of its neighbor images. Overlap regions between the image and its neighboring images are used to solve the equations of a multifactor model, which represents the pixel values of the image.
• None: Images are added to the mosaic exactly as they are and the pixel values are not be further adjusted.

• Bundle: <Autoscale>

  Autoscale can be specified for all input images to be stretched linearly and have a common scale. This can improve the results when different input images have significantly different dynamic ranges.

• LUT: <LUT_segment_number>

  LUT_segment_numbers explicitly specify the LUT-segment numbers to use for color balancing. 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 processing of the job stops and a message detailing the event appears.

• Histogram: [<match_area_multiple>], [<trim_percentage>]

  • match_area_multiple: Used to determine the amount of data in the mosaic file used to compute the reference histogram. If 1 is specified, the reference histogram is computed based on an area equal to the overlap between the incoming image and the output mosaic. If 3 is specified, the reference histogram is computed based on an area equal to three times the overlap between the incoming image and the output mosaic
  • trim_percentage: A number representing the percentage removed from the upper and lower parts of the histogram range. Trimming is to ignore the specified trim percentage of the total histogram area from the ends of both tails in a gray-value histogram when matching with another target histogram. Trimming is applied on both the input image and the target image. The default value is 2 percent.

• Reference: <fili_ref>

  fili_ref specifies the reference image file

• Neighborhood: [<meanbias_adjustment>], [<outlier_correction>], [<outlier_threshold>]

  • meanbias_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: Used to determine if extreme pixel values considered as outliers should be excluded from the computation of the coefficients. The default is set to YES, meaning that an outlier correction is applied.
  • outlier_threshold: Used only if outlier_correction is set to YES; it is set to 1 by default, meaning that pixel values that fall beyond the mean-difference value + 1 standard deviation are excluded from the computation of the coefficients.

• Last Vector: Sets the last-created vector segment in the source image as the local color-balance mask layer. The vector segment must contain a polygon, and all image pixels enclosed within it is excluded from the color-balancing calculation. When this option is selected, but no vector segment exists, all pixels are considered valid candidates for color balancing.
• Last Bitmap: Sets the last-created bitmap segment in the source image as the local color-balance mask layer. All image pixels covered by the bitmap is excluded from the color-balancing calculation. When this option is selected, but no bitmap segment exists, all pixels are considered valid candidates for color balancing.
• Specific Segment: Defines a specific segment in the source images as the local color-balance mask layer. All image pixels covered by the mask area is excluded from the color-balancing calculation. When this option is selected, a value must be specified for Local Color Balance Mask Segment.
• None: All pixels are considered valid candidates for color balancing.

• Last Vector: Sets the last-created vector segment in file specified for Global Color Balance Mask File as the global color-balance mask layer. The vector segment must contain a polygon, and all image pixels enclosed within it is excluded from the color balancing calculation.
• Last Bitmap: Sets the last-created bitmap segment in the file specified for Global Color Balance Mask File as the global color-balance mask layer. All image pixels covered by the bitmap is excluded from the color-balancing calculation.
• Specific Segment: Defines a specific segment in the file specified for Global Color Balance Mask File as the global color-balance mask layer. All image pixels covered by the mask area is excluded from the color-balancing calculation. When this option is selected, you must specify a value for Global Color Balance Mask Segment.

• Minimum Squared Difference: Specifies the minimum-squared-difference method. This method is suitable for most mosaicking projects, and 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.
• Minimum Difference: Specifies the minimum-difference method. This method is suitable for most mosaicking projects. 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
• Minimum Relative Difference: Specifies the minimum-relative-difference method. This method is similar to Minimum Difference, but provides better output in cases where similar sections of data appear dissimilar in different images
• Edge: Specifies 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.
• Maximum Data: Specifies that cutlines is on the boundary of the real image pixels, meaning that NoData pixels is ignored when the image boundary is determined.
• Import: Uses the specified polygons as cutlines exactly as they appear in the vector file.
• File Extents: Specifies that the cutline is the extents of the input images and does not exclude NoData pixels from the cutline generation process.

[<vector_file>], [<field_name>], [<segment_number>]

• vector_file is the name of an existing vector layer to use as a constraint polygon. Using a constraining vector layer is useful when at least some of the input images include features, such as clouds, that you want to exclude from the final mosaic.
• field_name is the attribute field that contains the image source.
• segment_number is the number of the vector segment that contains the constraint polygon.

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.

• Automatic: Determine whether constraints are required and, if so, applies one for each input image.
• Classic: Create one constraint polygon for each input image using PCI traditional algorithm that was developed for aerial images that are approximately square and contain a lot of overlap. You can adjust the option by specifying a value for the Thiessen Factor. If you do not specify a value, the default value of 50 is applied.
• Strips: Create one constraint polygon for each input image using PCI algorithm that was developed for projects that have a clear pattern of strips, such as often seen with ADS data. Images must be of similar size and align in at least one direction from, for instance, aerial flight lines or satellite swaths. Cutline-constraining polygons is generated for each image based on where it and its neighbors have actual data. This method first tries to group the images in rows or columns, and chooses the strips with the most consistent distance between images. It then removes the overlap between images in each strip using, at most, two neighboring images. Finally overlap is remove between the different strips leaving a constraining polygon. The extents of the polygons can be manipulated by specifying a value for the overlap Factor. If you do not specify a value, the default value of 20 is applied.
• No: Does not apply constraint polygons.

• Last Vector: Sets the last-created vector segment in the source image as the cutline-avoidance mask layer. When this option is selected, but no vector segment exists, all pixels are considered valid candidates through which a cutline can pass.
• Last Bitmap: Sets the last-created bitmap segment in the source image as the cutline-avoidance mask layer. When this option is selected, but no bitmap segment exists, all pixels are considered valid candidates through which a cutline can pass.
• Specific Segment: Defines a specific segment in the source image as the cutline-avoidance mask layer. When this option is selected, a value must be specified for Local Cutline Avoidance Mask Segment.
• None: All pixels are valid candidates through which a cutline can pass

• Last Vector: Sets the last-created vector segment in the file specified for Global Cutline Avoidance Mask File as the global cutline-avoidance mask layer.
• Last Bitmap: Sets the last-created bitmap segment in file specified for Global Cutline Avoidance Mask File as the global cutline-avoidance mask layer.
• Specific Segment: Defines a specific segment in the file specified for Global Cutline Avoidance Mask File as the global cutline-avoidance mask layer. When you select this option, you must specify a value for Global Cutline Avoidance Mask Segment.

• Single Tile: Generates the output mosaic as a single tile.
• Vector: Use when you have an existing vector-polygon layer that contains the tile definitions you want to use for the mosaic. When you select this option, you can add options for it with Tile File, Segment Number, Field Name, and Buffer Distance, respectively.
• Dimensions: Creates tiles of a specific size. When you select this option, you can add options for it with Tile Height and Tile Width, respectively. When using this specification, the TileID values of the output tiles is generated using the convention <column_number>_<row_number>. For example, the upper-left tile always has a TileID of "1_1", the one immediately below is "1_2", and so on.
• Script File: Use when you have a text file that defines the coordinates of each tile you want to create. When you select this option, you can add options for it with Tile File.

• RASEXT: Raster extents
• GEOEXT: Geocoded extents
• LNGLATEX: Longitude and latitude extents
• RASOFFSZ: Raster offset and size
• GEOOFFSZ: Geocoded offset and size

• Geocoded Extents
• Geocoded Offset & Size
• Long/Lat Extents
• Raster Extents
• Raster Offset & Size

• CORNER: indicates that the values are relative to the upper-left corner of the first pixel in the output tile
• CENTER: indicates that the values are relative to the center of the first pixel

• Stride_X
• Stride_Y
• Ref_X
• Ref_Y

• SourceID: a text field that contains the SourceID of the source image.
• SourceFile: a text field that includes the full GDB string including path to the source input image.
• Z-order: an integer field that records the z-order of the predominate source image. The word "predominant" is used because, in the case of blending along cutlines, the source scene which had the largest contribution is the one indicated.

• Cubic: cubic convolution. This method determines the gray level from the weighted average of the 16 pixels closest to the specified input coordinates and assigns that value to the output coordinates. The resulting image is slightly sharper than one produced by bilinear interpolation, and it does not have the disjointed appearance produced by nearest-neighbor interpolation.
• Nearest Neighbor: nearest-neighbor interpolation. This method identifies the gray level of the pixel closest to the specified input coordinates and assigns that value to the output coordinates. Although this method is the most efficient in computation time, it introduces small offsets in the output image. The output image may be offset spatially by up to half a pixel, which may cause the image to have a jagged appearance.
• Bilinear: bilinear interpolation. This method determines the gray level from the weighted average of the four closest pixels to the specified input coordinates and assigns that value to the output coordinates. An image with a smoother appearance than nearest-neighbor interpolation is created, but the gray-level values are altered in the process, resulting in blurring or degraded image resolution.