The Filter tool is used to filter a continuous or classified raster using a specified filter kernel. Grids generally comprise of equally spaced data values located across rows and columns. When filtering a grid, each grid node of the output grid is calculated as a function of the corresponding node and its neighboring nodes. The size of the neighborhood of cells used in the filtering process is defined by the size of the filter or kernel.
A filter or kernel is defined as an array of grid nodes, with each node being assigned a set of filter weights. Since the filter neighborhood needs to be the central node of the grid, filters generally have odd number of rows and columns.
The filter neighborhood is shifted across the grid and a new value for each grid node is calculated in the output image according to the filter method selected.
A typical kernel calculates the average (mean) value for each m x n neighborhood of cells. Each cell value in a grid is replaced with the weighted value of the filter neighborhood, including the central node value. A typical 3x3 kernel will be represented as:
1 1 1 1 1 1 1 1 1
Available Grid Filters
A convolution filter is supplied in MapInfo Pro Advanced. Convolution is a neighborhood operation where each output cell is the weighted sum of all neighboring cells. The weight matrix in this case is known as a convolution kernel. The types of convolution filters that are present in MapInfo Pro Advanced are:
Supported raster field types for Filter.
| Filter Type | Continuous | Classified | Imagery | Image Palette |
|---|---|---|---|---|
| Enhancement | Yes | No | No | No |
| Smoothing Filter | Yes | No | No | No |
| Custom | Yes | No | No | No |
| Focal | Yes | Yes | No | No |
| Classified | No | Yes | No | No |