International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B1. Istanbul 2004
Optimally the image contains as much information as possible
i.e. the efficiency of its histograms should be high. At the same
time saturation of bright and dark areas should be avoided.
Based on the studies made, following recommendations were
established:
e Efficiency of R, G, B and L-histograms should be at
least 230 DNs i.e. 90%
e 99%-efficiency of R, G, B and L-histograms should
be 160-254 DNs
e 0% and 255% saturation of L-histogram should not
exceed 0.5%
e (0% and 255% saturation of R, G and B-histograms of
CIR-positive-film should not exceed 2.0%
e 0% and 255% saturation of R, G and B-histograms of
other film-types should not exceed 0.5%
The same recommendation for saturation of L-histogram is also
given by European Commission (2004). Scanner- and film-
type-detailed data as presented in Figure 4 can be used to
improve these recommendations after the 100%-histogram-QC
has been taken to full operational use.
Because the image-detailed statistics were in a flight level
quite uniform, excluding exceptional images, it can be
assumed that the calculation of histogram statistics in scanning
parameter selection phase would lead to good result in
scanning of the whole flight.
5.2 Tone model image
The created tone-model-image is one alternative that can be
used in the tone tuning for forest interpretation purposes. lt is
not optimal for all purposes, but it is a scientifically justified
and radiometrically good quality alternative for images based
on subjective criteria. The information loss mentioned in
chapter 4.3 is a consequence of high contrasts required for
forest interpretation and was expected. Because of this, MAF
decided not to use the tone-model-image (Figures 7. and 8.) in
the image enhancements for land parcel identification.
6. CONCLUSIONS
A 10096-histogram control of scanned images is highly
recommended. Calculating histograms and statistics after
scanning is an efficient process. Histogram-control should be
implemented directly to scanning software for the maximum
ease of use. For each scanned film diagrams presenting
efficiencies, 99%-efficiencies and saturation of every image
should be created and summary statistics should be calculated.
Separate alarm should be given of the images that exceed the
given tolerance values. The central data should be archived; in
a long run these statistics give valuable information of the
development of scanning process and they can be used for
optimization of scanning parameters. The histogram control
can also aid in the selection of scanning parameters.
Recommendations of tolerance values given in chapter 5.1 are
not in any way final; they will be improved when the 100%-
"histogram-QC is taken to full operational use. In addition to
100%-histogram-QC, thumbnail images of every scanned
image should be created. The complete QC-system should
contain also visual inspection of a sample of final images,
because all the radiometric errors cannot be detected by
histogram control.
Interpretations of forests and land parcels have conflicting
requirements for radiometric enhancement of orthophotos.
Specifying numeric recommendations for image enhancement
procedures appeared to be difficult. The most practical way for
defining the desired radiometry with the existing commercial
systems is a model image. One image cannot be optimal for all
purposes, thus various applications and even various user may
require different model images. In this article a tone-model-
image suitable for forestry applications was presented.
REFERENCES
Carper, W J., Lillesand, T.M. and Kiefer, R.W., 1990. The Use
of Intensity-Hue-Saturation Transformations for Merging
SPOT Panchromatic and Multispectral Image Data.
Photogrammetric Engineering & Remote Sensing, 56(4), pp.
459-467.
Conrac Division, Conrac Corporation, 1985. Raster Graphics
Handbook, second edition. Van Nostrand Reinhold Company
Inc.
European Commission, 2004. Guidelines for Best Practise and
Quality Checking of Ortho Imagery. European Commission,
Directorate General, Joint Research Centre, Institute for the
Protection and Security of the Citizen, Monitoring Agriculture
with Remote Sensing Unit. Issue 2.3.
http://marsunit.jrc.it/Mapping/Guidelines for ortho
QA/Forms/Allltems.htm (accessed 26. April, 2004)
Honkavaara, E., Kaartinen, H., Kuittinen, R., Huttunen, A.,
Jaakkola, J., 1999. Quality of FLPIS Orthophotos. Reports of
the Finnish Geodetic Institute, 99:1
Honkavaara, E., 2003. Maa- ja metsütalousministerión
J
peltolohkorekisterin ortoilmakuvatuotannon laatujürjestelmá.
(In Finnish)
Gonzalez, R. C. and Woods
processing. Addison-Wesley.
R. E. 1992. Digital image
Kosaka D., 2000. A primer an Image Histograms and Curves.
http//www.marginalsoftware.com/HowtoScan/tutorial page 1.
htm (accessed 26.April, 2004)
Koutsias, N., Karteris, M., Chuvieco, E. 2000. The use of
Intensity-Hue-Saturation — Transformation of Landsat-5
Thematic Mapper Data for Burned Land Mapping.
Photogrammetric Engineering & Remote Sensing, 66(7), pp.
829-839.
Schowegerdt R., 1997. Remote Sensing — Models and Methods
for Image Processing, second edition. Academic Press.
ACKNOWLEDGEMENTS
The investigation has been made in close co-operation between
Ministry of Agriculture and Forestry, FM-Kartta Ltd., National
Land Survey of Finland, The Forestry Development Centre
Tapio and Geoaudit Ltd. All the participants are gratefully
acknowledged.
