| | Exercise 3 - Change detection analysis using subtract function
In this section you will use the Image Arithmetic function in LeoWorks in order to undertake a change detection of our study area.
Open band 5 of the 1984 image, selecting the following crop parameters (Step 2, section 3.1 of the LEOWorks tutorial): Left=260, Right=660, Top=410, Bottom=920. This subset is centred on the Sacalin Island. Repeat the process for band 5 of the image from 2000 using the same crop limits.
Use the Arithmetic function of LEOWorks (section 3.6 of the LEOWorks 3 tutorial) to subtract band 5 of the 1984 image from band 5 of the 2000 image. The result should look like this:
Band 5 of Landsat image acquired in 1984 subtracted from band 5 of Landsat image acquired in 2000
The image created as a result of the subtract operation is a grayscale with values from 0 to 255. As it is difficult to make a visual difference analysis using a grayscale image, we will convert it to a colour image. We will do that using the LEOWorks LUT editor. Following the steps outlined in section 4.5 of the LEOWorks 3 tutorial, edit the LUT range using the following values: 0-15 (blue), 16-135 (green) and 136-255 (red).
In order to reduce the maximum number of different colours to 256, some image formats (e.g. 8-bit TIFF) use an 8-bit Look Up Table (LUT). Using this method, it is only necessary to store an 8-bit index into the colourmap for each pixel, rather than the full 24-bit colour value. Thus 8-bit colour image formats consist of two parts: a Look Up Table describing what colours are present in the image, and the table of index values for each pixel in the image.
Band 5 of Landsat image acquired in 1984 subtracted from band 5 of Landsat image acquired in 2000
The resulting image should look like the above. In this image, blue represents water, green describes the new position of the island in 2000, and red shows the areas where the island has been eroded. Changes at the waterline can now be identified. In the mainland part of the delta, changes are a result of different vegetation types or of water level fluctuation. This image subtraction method is useful primarily at the boundary between land and water, describing very well erosion or accretion areas.
| |
