A tie point is a point that has ground coordinates that are not known, but is visually recognizable in the overlap area between two or more images. The corresponding image positions of tie points appearing on the overlap areas of multiple images is identified and measured. Ground coordinates for tie points are computed during block triangulation. Tie points can be measured both manually and automatically.
Tie points should be visually well-defined in all images. Ideally, they should show good contrast in two directions, like the corner of a building or a road intersection. Tie points should also be well distributed over the area of the block. Typically, nine tie points in each image are adequate for block triangulation. The figure below depicts the placement of tie points.
Point Distribution for Triangulation
In a block of images with 60% overlap and 25-30% sidelap, nine points are sufficient to tie together the block as well as individual strips, as shown in the figure below.
Tie Points in a Block
Automatic Tie Point Collection
Selecting and measuring tie points is very time-consuming and costly. Therefore, one of the major focal points of research and development in photogrammetry has concentrated on the automated triangulation where the automatic tie point collection is the main issue.
The other part of the automated triangulation is the automatic control point identification, which is still unsolved due to the complication of the scenario. There are several valuable research results available for automated triangulation (for example, Agouris and Schenk, 1996; Heipke, 1996; Krzystek, 1998; Mayr, 1995; Schenk, 1997; Tang et al, 1997; Tsingas, 1995; Wang, Y., 1998b).
After investigating the advantages and the weaknesses of the existing methods, IMAGINE Photogrammetry Project Manager was designed to incorporate an advanced method for automatic tie point collection. It is designed to work with a variety of digital images such as aerial images, satellite images, digital camera images, and close range images. It also supports the processing of multiple strips including adjacent, diagonal, and cross-strips.
Automatic tie point collection within IMAGINE Photogrammetry Project Manager successfully performs the following tasks:
- Automatic block configuration. Based on the initial input requirements, IMAGINE Photogrammetry Project Manager automatically detects the relationship of the block with respect to image adjacency.
- Automatic tie point extraction. The feature point extraction algorithms are used here to extract the candidates of tie points.
- Point transfer. Feature points appearing on multiple images are automatically matched and identified.
- Gross error detection. Erroneous points are automatically identified and removed from the solution.
- Tie point selection. The intended number of tie points defined by you is automatically selected as the final number of tie points.
Image matching strategies incorporated in IMAGINE Photogrammetry Project Manager for automatic tie point collection include the coarse-to-fine matching; feature-based matching with geometrical and topological constraints, which is simplified from the structural matching algorithm (Wang, Y., 1998b); and the least square matching for the high accuracy of tie points.