International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Times of edge protection smoothing Naught | One | Two | Three
Entropy (bits) 6.17 6.11 6.08 6.06
Information amount (bits) 3.44 2.99 2.44 2.08
Average code length 6.20 6.13 6.11 6.11
Hofman Coding Compression rate 1.29 1.30 1.31 1:31
Uni-directional Average code length 4.50 4.13 3.63 3.32
differential coding Compression rate 1.78 1.94 2.20 241
Bidirectional Average code length 3.66 4.05 3.56 322
differential coding Compression rate 2.19 1.98 2.42 2.49
Table 4. Comparisons of Huffman coding and differential coding
corresponding to different times of edge protection smoothing
Urbanareas | Farmland | Mountain areas | water
Naught Average code length 6.12 4.89 4.13 3-27
Compression rate 1.61 2.21 2.52 332
One time Average code length 5.81 4.78 4.01 3.10
smoothing Compression rate 1.74 2:51 2.09 4.36
Two Times Average code length 5.63 4.70 3.07 3.06
smoothing Compression rate 1.99 3.02 3.81 8:17
Three times Average code length 5.51 4.65 3.95 3.07
smoothing Compression rate 3.23 3.45 4.29 7.25
Table 5. Unidirection differential coding of different
land cover types after different times of edge protection smoothing
transmission channels will be effectively utilized. The edge-
s protection smoothing is a bit of time-consuming, but enough
T time is available for on-board smoothing when a satellite is far
away from the ground receiving stations. This strategy proves
g- to possess much higher compression rates that would improve
| Ä the utilization rates of the expensive satellite Earth
$ - A P My communication infrastructures. It is of especial importance for a
Aud Js M, IL W 1 country, exemplified by China, which has not yet constitute a
2 A | A d j worldwide ground reception station system, and is capable of
A = meeting the requirements of cartographic satellites.
S S 4 ir
NS vl Mi LUN : REFERENCE
e | E Deng B., Lin Z. J., 2010. Onboard data lossless compression of
2 + S remote sensing image based on "District Forecast" differential
di coding. Science of Surveying and Mapping, 35(1), pp. 10-12
T T Hu Z. M., 1979. Orthogonal Transformations in Digital Signal
0 100 200 300 400 500
Processing. Posts and Telecom Press, Beijing.
Lin Z. J., 1988. Multi-information and Multi-criterion Image
Matching. Wuhan Technical University of Surveying and
Mapping, Wuhan
Figure 7. A-A sectional drawing
4. CONCLUSION Lin Z. J., Zhang Y. H., 2006. Measurement of information and
uncertainty of remote sensing and GIS data, Geomatics and
Obviously, the proposed differential encoding for lossless Information Science of Wuhan University, 37), pp. 569-572.
compression is feasible. Besides, this method is simple and easy
to implement which makes it worthy of application in data Tao C. K., Tao. C. K., 2004. Optical Information Theory.
transmission to economize in resources and improve Science Press, Beijing, pp. 36-55.
transmission efficiency. In addition, the satellite orbits may be
partitioned into districts according to historical aerial images,
remote sensing images, GIS data and so on, together with the
establishment of different compression modules, which would ACKNOWLEDGEMENT
serve to compress different land cover images and maximize the
compression rates of lossless compressions. Thus, the valuable The research is partially supported by “973 Program” grants
storage space on-board and the limited band width of (No. 41071286) and (No. 41171346 ).
