TIN Model and Single Beam Data
There are many obstacles to create a useful TIN from a single beam data set. Although the data set might have a high level of accuracy the corresponding TIN can be almost unusable.
The main problem is to prevent the creation of undesirable triangles along the edges and wide triangles inside of TIN body. These triangles contribute to volume calculation error and contour accuracy.
To prevent such situations, we use the original planned line file that guided the vessel as the data was collected. The line file will prevent the TIN generator from connecting any three points that belong to the same line.
|
The Initial Data dialog has an additional check box that allows you to align the TIN and the corresponding line file. The line file name should be entered in the "Section File" field. Set the TIN Max Side as usual and click OK to create your model. There are a few requirements regarding the planned line file:
Failure to meet any of the requirements will result in a badly created TIN that will lead to incorrect volume calculations and very strange contours. An attempt to TIN multibeam data with the method will produce the same effect. The following pictures show the standard and the new methods applied to a real data set. Looking at the first pair of pictures, you will see that we have achieved both goals: |
 |
- Eliminate wild triangles along the edges.
- Eliminate wide triangles inside the TIN body.
The second pair shows the contours. You can see that the new method eliminates all disruption along the edges or survey lines caused by poor triangle formation
 |
The next set of pictures is an artificial example. Our first picture shows an imaginary channel 200m wide, 50m deep with 1:1 slope on both ends.
Four survey lines cross the channel at an angle of less than 45 degrees. The following pair of pictures are Solid TIN models with contours. You can see that the standard method creates bumps along the channel edge making a nonexistant volume buildup that affects the volume calculation. The new method calculates the volume (below 0.0 level) correctly 1 125 000.00 volume units. An error of about 1.3%.
