Figure 4 displays the differences between the 
GPS/DR positions and GPS positions — in other 
words, the combined GPS/DR “modeled” trajectory 
of the van compared to GPS alone — and indicates 
an average error due to modeling of about 30cm. 
20.7 
® | 
815 + 
F [B 
aor À 
o | E 
TE | | 
  
Figure 4. Difference between the GPS and 
integrated GPS/DR trajectories 
Figure 5 displays the differences between the 
GPSVan™-determined coordinates of QA/QC 
points along the railroad and the statically measured 
QC points. By comparing Figures 4 and 5 we see 
that the photogrammetric feature extraction and 
post-processing add very little to the overall error 
budget. The dominant error is from the DR system. 
While we cannot isolate each error component, 
indications are that ©, (from the DR system) and 
e 
error budget. Having said that, the 50cm average 
planimetric results are enormously impressive 
(Bossler and Toth, 1995). 
Opeator are by far the largest components of the 
-— T N 
C © 
pf 
Qf rn LAO 
   
No. of surveys 
  
Figure 5. Distribution of differences at QC points 
4. CONCLUSION 
GPSVan™ technology is being used to collect 
current, accurate, and complete spatially referenced 
digital data for railroad right-of-ways. The two- 
dimensional (horizontal) positional accuracy 
obtained in the BNR project for well-defined 
features is approximately 50cm (4jo K. +0, ) 
without any data editing. Clearly, even more 
accuracy can be squeezed from the system and that 
142 
will likely occur. However, we are now at the point 
where questions such as "Where on the top of the 
rail is the coordinate?" are critical in the context of 
such accuracy. Now our attention should be directe 
toward presenting these data in a more enhanced 
fashion and toward integrating these data with other 
datasets. It is also clear that real-time data will be 
valuable for checking on errors (QA/QC) and for 
changing surveying strategies in the field. We are 
very close to being able to do just that, and we 
expect that this system will be available 
commercially in 2-3 years. 
Mobile Mapping Systems, primarily because of 
GPS, have revolutionized the mapping sciences. The 
next steps in multimedia presentation, real-time 
processing, and integration with other data will also 
be enormously exciting. 
S. REFERENCES 
Blaho, G., and Toth, C., 1995. Field Experiences 
with a Fully Digital Mobile Stereo Image 
Acquisition System, Proc. Mobile Mapping 
Symposium, OSU, pp. 97-104. 
Bossler, J.D., and Toth, Ch, 1995. Accuracies 
Obtained by the GPSVan™, Proc. GIS/LIS Annual 
Conference, Vol. 1, pp. 70-77. 
He, G.P., Cunningham, D., and Bossler, J.D., 1994a. 
Spatial Data Collection with the GPSVan Mobile 
Mapping System, Symp. of ISPRS Comm. IV., Vol. 
30, Part 4, pp. 106-113. 
He, G.P., Dedes, G., Orvets, G., and Bossler, J.D, 
1994b. (Generation of transportation GIS by 
Integrating GPS, INS and Computer Vision 
Technology, Proceedings of the 3rd Int'l Coll of 
LIESMARS, WTUSM, pp. 91-99. 
He, G.P., Novak, K., and Tang, W., 1994c. The 
Accuracy of Features  Positioned with the 
GPSVan™, Symp. of ISPRS Comm. IL, Vol. 30, 
Part 2, pp. 480-486. 
Novak, K., and Bossler, J.D., 1995. Development 
and Application of the Highway Mapping System of 
Ohio State University, Photogrammetric Record, 
15(85), pp. 123-134. 
Toth, C., 1995. Experiences with a Fully Digital 
Image Acquisition System, Proc. ASPRS-ACSM 
Annual Convention, Vol. 2, pp. 18-24. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996