To appear Previous: 4.4 Summary

5 Lossy Compression Experiments

Since elevation data can have many errors[2], and in any case is represented only to the nearest meter, it is a waste to store the data at a higher accuracy than is justified. Therefore lossy compression methods might be appropriate sometimes. Indeed, sometimes we gather data faster than we can easily store or transmit it. Altho the our first suggestion here might be simply to lower the spatial resolution or to truncate the number of bits used to store each point, a more sophisticated approach, such as one of the algorithms described below, might allow us to keep more information in fewer bpp.

  
Table 5: Histogram of Errors Caused by Various JPEG Parameter Settings on the Scaled Adir512

• 5.1 Scaling the Elevations
• 5.2 Lossy Compression of the Scaled Data
• 5.3 Said and Pearlman