DIMPLE 3.0

Dimple 3.0 is the leader in remote sensing image analysis - available for Mac OS and Windows!

This page lists in detail the features of DIMPLE 3.0. Screen shots are provided for many features.

Full 32 bit data processing
DIMPLE is the first image analysis system to make working with extended data as easy as working with 8 bit. DIMPLE supports 8, 16, and 32 bit signed and unsigned integer, 32 bit real, 16 and 24 bit RGB, and vector imagery. You can mix data types in a single multispectral image, and all operations allow you to choose the data type and stretch for output.
You can use images in memory for maximum performance, or from disk when working with large imagery.

File format support
As well as a range of built-in file formats, DIMPLE allows you to import data from virtually any file containing Band Interleaved by Line (BIL) or Band Sequential (BSQ) data, by describing the layout of the data in the file. Built-in support is provided for DISIMP, DXF, ER Mapper, LGSOWG/CEOS, microBRIAN, ERDAS, PICT (Mac OS only), SPOT CCT, and TIFF.

Multispectral processing
DIMPLE allows you to work with images with up to 32,000 channels. The multiband window makes managing multispectral and hyperspectral data easy. You can give channels meaningful names, choose which channels are to be used in multispectral operations, move channels to disk or into memory for optimal performance, view individual channels, and more.
As you move over an image, the pixel graph shows the spectral signature of the pixel under the pointer. Use the pixel graph to identify features quickly, or to observe changes in pixel values.

Registration & Rectification
Registration of imagery to a map grid, image-to-image registration, and geometric correction all take advantage of DIMPLE's interactive user interface. For image-to-grid registration, create a Ground Control Point (GCP) model by selecting GCPs on the image. Specify their grid coordinates. Create a polynomial model. DIMPLE allows you to specify a linear, quadratic, or cubic model, and to detect outliers by excluding any GCP from the model. A GCP report shows the GCPs and mapping polynomials in detail.
For image-to-image registration, you follow the same process, except that instead of specifying grid coordinates, you select Relative Control Points (RCPs) on a "master" image. This master image can be either a raster or a vector image. If the master image is a raster image, you can automatically locate RCPs using mean-normalised similarity or correlation algorithms.
When resampling an image, DIMPLE allows you to use nearest neighbour, bilinear interpolation, or cubic convolution methods.
Mathematically modelled rectification includes corrections for aspect ratio, skew, earth rotation skew, rotation, reflection, and affine transformations.
DIMPLE supports over 900 projected coordinate systems and over 160 geographic coordinate systems, as well as user-defined systems.

Mosaicing
Use the Join Images command to mosaic images. You can control how overlapping areas are mosaiced, and specify the output image type and bounds in registered coordinates.

Classification
For supervised classification, DIMPLE provides training tools, a comprehensive range of classifiers, and post-classification analysis tools. Training sets can be created from rectangular, irregular, or polygonal areas. The training set editor allows you to select theme colours, change class names, examine statistics, and merge training sets. A range of supervised classifiers is available: Bayesian maximum likelihood; minimum distance (Euclidean, city block, or chess board); parallelepiped; and Gaussian probability. For unsupervised classification, k-means and minimum distance classifiers are available. A dynamic report showing the results of each iteration assists in analysing the progress of an iterative classification.
DIMPLE provides sophisticated tools for both training and post-classification analysis. Statistical (minimum-maximum; mean & standard deviation) and canonical variate cross plots provide an intuitive and visual way of assessing class separation. Post-classification reports allow in-depth statistical analysis of classes.
For reclassification, the class editor allows you to examine class statistics and compare them with training area statistics; change class colours; merge similar classes; and delete classes. For annotating a classified image, you can add a class key.

Filters
In addition to a library of built-in filters for sharpening, smoothing, noise removal, edge detection, and edge enhancement, DIMPLE provides the ability to define your own adaptive filters. Convolution kernels up to 9x9 in size can be specified, as well as conditions under which the filter will operate. Custom filters saved in the "Filters" folder appear in the Enhance menu, and can be instantly accessed in the same way as built-in filters. In addition, log intensity, median, and rank-order filters are available.

Histogram
The image histogram allows analysis of intensity levels in an image, as well as interactive contrast enhancement. A range of stretches is available: linear stretch, with saturation by pixel range, percentage range, or mean ± n std dev; piecewise linear stretch; histogram shift; histogram equalisation; gaussian stretch; and exponential stretch. Commonly used stretches can be saved with specific parameters for fast access. The colour components histogram allows histogram modifications on Red-Green-Blue, Hue-Saturation-Intensity, or Cyan-Magenta-Yellow channels of an image.

Density Slicing
The density slicing tool allows you to choose breakpoints numerically or graphically. Slices are coloured automatically, or you can manually assign colours to individual slices. For annotating a density sliced image, you can add a density slice key.

Linear Profile
The linear profile tool allows you to analyse pixel value changes across an image, for a single image or all channels of a multispectral image. The linear profile can be interactively scaled. A linear profile report shows the profile for all channels, and you can copy the pixel data to the clipboard for further analysis in other applications such as spreadsheets.

Image Ratio
Built-in image ratio functions include simple ratios (x/y), vegetation indexes (x-y)/(x+y), and log ratios ln(x/y). For more complex ratios, you can use the Image Operation Language.

Image Cross Plot
The image cross plot allows you to analyse the correlation between two channels. Cross plots can be produced in colour or grey scale to indicate the frequency of a particular pixel combination. Alternatively, a cross plot of two images can be produced using the colours of another related image; use this to relate the colours in a colour composite image to the values of images used to produce it. You can also use a classified image to colour a cross plot, thereby showing the relationship between classes in the image and pixel values.

Gridding & Interpolation
For gridding x,y,p data such as geochemical or digital terrain data, DIMPLE provides unrivalled ease of use. You import virtually any text file containing sampled data by specifying the format of the data - fixed field or variable field; single channel or multichannel. Specify gridding and interpolation parameters, and produce raster imagery. For analysing sample points, you can produce a vector image, and overlay the points onto the raster images.

3D Modelling
3D plots can be used for modelling digital terrain data or for depicting data relationships. You can produce a 3D plot of any data, and drape another image over it. Adjust the resolution, scale, and vertical exaggeration. Change the viewpoint and orientation interactively.

Image Operation Language
For custom operations, DIMPLE provides the unique Image Operation Language (IOL). Specify new operations using the built-in library of mathematical and convolution operations, from simple image ratios and vegetation indexes to adaptive filters to complex multi-step operations. DIMPLE automatically gives you a graphical interface for your operations, and you can add them to the menu for use in the same way as built-in functions.

Colour Composites
You can produce colour composite images in Red-Green-Blue (RGB), Hue-Saturation-Intensity (HSI), or Cyan-Magenta-Yellow (CMY) space. The colour components histogram allows you to manipulate the colour channels of an image, for producing an image with the optimum visual definition. You can apply built-in or user defined stretches to each channel, in the same way as operating on individual images using the image histogram. The produce grey scale command allows you to produce a grey scale representation of a colour image.

Principal components
The principal components transformation produces a set of transformed channels from a multispectral image in which correlation among channels is minimised. DIMPLE allows you to produce principal component channels and also a principal components report, showing details of covariance for the images. A related transformation is canonical components, which produces channels in which the correlation among channels for particular classes in the image is reduced.

Channel prediction
Channel prediction analyses the relationship between channels in an image, to predict the value of a channel from other channels. By basing the analysis on a selected area that excludes a given feature, channel prediction can be used to show what the response for a particular channel would be if a given feature were not present. Subtracting the predicted channel from the actual channel shows the distribution of the feature in the image.

Annotation
The annotation tools in DIMPLE allow you to produce professioally presented output. Add captions to an image, and label geographic features. Add a north arrow, scale bar, and map grid for geographically accurate output. Add a class key to a classified image, a density slice key for a density sliced image, or a colour scale indicating the values of colours in a pseudo-coloured image. Overlay vector imagery to show boundaries, roads, or other features. Print preview allows you to see what the image will look like when printed at scale, including the position of page breaks. DIMPLE can even position annotations automatically for a professional result, fast.

Cross-platform compatibility
All versions of DIMPLE share the same friendly user interface and the same file formats, while providing native performance and conforming to user interface standards on each platform. You can exchange images seamlessly with DIMPLE users on other platforms.

System requirements
DIMPLE is available for Mac OS, Windows 95, and Windows NT.

To use DIMPLE for Mac OS, you will need the following:

• A 680x0 or Power Macintosh or 100% compatible (Power Macintosh recommended)
• 16-bit or 24-bit colour, and a colour monitor
• A CD-ROM drive
• A hard drive with at least 5 MBytes of free disk space
• At least 12 MBytes of RAM (16 MBytes recommended)
• Mac OS System 7.0 or later

• A 486 or Pentium-class PC
• Windows 95, Windows NT 3.51, or Windows NT 4.0
• A video card with 16-bit or 24-bit colour, and a colour monitor
• A CD-ROM drive
• A hard drive with at least 5 MBytes of free disk space
• At least 16 MBytes of RAM