SFAP survey planning often requires compromising about small GSD of the images and large area coverage (see Chapter 9). In order to achieve a desired degree of detail, it might be necessary to take multiple images for a given area that subsequently need to be assembled into an image mosaic. In this way, even unmanned SFAP may cover considerable areas with high resolution, provided a platform capable of travelling longer distances is used, e.g. a model airplane (see Chapter 8.5.2).
The term mosaicking is commonly used in remote sensing and geoinformation sciences when contiguous images are joined to form a single image file. However, this does not necessarily require that the images be georeferenced—rather than using GCPs for positioning the individual images in a common reference coordinate system it is also possible to use any homologous image objects for a relative spatial adjustment of the photographs (uncontrolled mosaic). This can be done either in remote sensing and GIS software or with dedicated photo-editing software, with varying degrees of automation.
Most digital camera software now includes so-called stitching functions, and there is a large choice of free and commercial panorama software for fully automatic merging of multiple photographs into larger composites. These tools can produce results that are visually highly appealing (see Fig. 11-3), but they are not geometrically correct. Note the angular skew and straight edges of the left image tiles in Figure 11-3. If the mosaic were georectified, the edges would be bent and warped following the changing relief heights of the hillslopes, similar to the bottom edge of Figure 11-1C. Remote sensing or GIS software is required for georeferenced (controlled) mosaics, which can be constructed from image tiles previously geocorrected with polynomial transformations or orthorectification.
Image mosaics mercilessly reveal exposure differences across or between photographs. Regardless of controlled or uncontrolled and no matter which software used, a good mosaic requires some sort of color matching between the stitched images in order to conceal the seams. Not all software capable of stitching adjacent images together may provide color-correction tools, and matching the image tiles for a balanced overall impression may be a difficult task especially for large mosaics with strongly varying brightness distributions (Fig. 11-4).
Was this article helpful?