In: Wagner W., Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
386 
disagreement may be caused by the nature of SSIM measure, 
which uses comparison of luminance and contrast of the 
images. For this example, PC ZNCC assessment is not skewed 
and coincides with results of HPCC and Correlation. 
The second assessment setup is expected to illustrate increase of 
spectral consistency with simultaneous decrease of spatial 
consistency on the fusion results (GIF-2 method, change of 
parameter set). Dependency graphs of the assessment measures 
are presented in Figure 2 (spectral consistency: SSIM mean, 
ERGAS mean, SAM) and in Figure 3 (spatial consistency: 
SSIM PAN (mean), ERGAS PAN, CORR PAN (mean), HPCC 
(mean), PC ZNCC (mean)). Since the ideal values for SSIM, 
ERGAS, and SAM are (respectively) 1, 0, and 0, the spectral 
consistency measures are increasing (Figure 2). This 
corresponds to our assumption and expectation. For the spatial 
consistency assessment, the ideal values for SSIM PAN, 
ERGAS PAN, CORR PAN, HPCC, PC ZNCC are 
(respectively) 1,0, 1, 1, and 1, the spatial consistency measures 
are decreasing (Figure 3). This also corresponds to our 
assumption and expectation for spatial consistency. Figure 3 
clearly illustrates similar trend of the PC ZNCC with all the 
other spatial consistency measures. 
Table 2 illustrates common trend on spatial consistency 
between the results obtained by known measures and the PC- 
based metric. Spatial consistency assessment using PC 
illustrates expected decrease of spatial consistency. Also, the 
PC ZNCC measure is more sensitive to change of spatial 
consistency, so it is easier to assess and compare the quality of 
the image. 
Visual assessment shows that the best spatial consistency have 
the IHS, PCA, GIF-1, and GIF-2 methods while ATWT resulted 
in slightly blurred edges (Figure 4), and coincides with the 
results of numerical assessment using PC. Figure 5 presents 
corresponding fragments of panchromatic and fused image (IHS 
fusion), edge maps (Sobel operator), and maximum moment of 
PC covariance (indicator of edge strength). It should be noted 
that PC feature map should not be confused with edge map. 
Figure 5 illustrates difference of image intensity and contrast 
(subfigures a), b)). Different edge maps are produced by edge 
detection operators (subfigures c, d). It is also demonstrated that 
the PC is more stable to intensity and contrast change 
(subfigures e, f). 
5. CONCLUSIONS 
Not many papers report on spatial consistency assessment of 
pan-sharpened data. Therefore, a need for robust and sufficient 
measures still exists. Application of phase congruency for 
spatial consistency assessment is proposed. Multiscale nature of 
phase congruency as well as invariance to intensity and contrast 
change allows more thorough analysis of fused data, comparing 
to single-scale edge detection methods. Identical trend with 
different assessment measures and with visual assessment 
showed that phase congruency is relevant for spatial 
consistency assessment, and the decision on the consistency can 
be made with higher confidence. Also it was found that ERGAS 
and SSIM provided less stability for spatial consistency 
assessment than correlation and edge-based measures. It should 
be noted that sometimes use of single assessment measure is not 
sufficient and may give skewed results (not all the 
characteristics of the fused data are revealed). Therefore use of 
several assessment measures increases confidence over 
calculated results. 
REFERENCES 
Aiazzi, B., Alparone, L., Baronti, S., and Garzelli, A., 2002 
Context-driven fusion of high spatial and spectral resolution 
images based on oversampled multiresolution analysis, IEEE 
TGRS, 40(10), pp. 2300-2312. 
Alparone L., Baronti S., Garzelli, A., and Nencini, F., 2003. A 
global quality measurement of pan-sharpened multispectral 
imagery. IEEE GRSL, 1(4), pp. 313-317. 
Ehlers, M., 2004. Spectral characteristics preserving 
imagefusion based on Fourier domain filtering. In: Remote 
Sensingfor Environmental Monitoring, GIS Applications, and 
Geology IV, Bellingham, USA, pp. 1-13. 
Gonzalez-Audicana, M., Otazu, X., Fors, O., and Seco, A., 
2005. Comparison between Mallat's and the ‘a trous’ discrete 
wavelet transform based algorithms for the fusion of 
multispectral and panchromatic images. IJRS, 26(3), pp. 595- 
614. 
Kovesi, P., 1999. Image features from phase congruency. 
Videre: A Journal of Computer Vision Research, 1(3), pp. 2-26. 
Lillo-Saavedra, M., Gonzalo, C., Arquero, A., and Martinez, E., 
2005. Fusion of multispectral and panchromatic satellite sensor 
imagery based on tailored filtering in the Fourier domain. IJRS, 
26(6), pp. 1263-1268. 
Liu, Z., Forsyth, D., and Laganiere, R., 2008. A feature-based 
metric for the quantitative evaluation of pixel-level image 
fusion, Computer Vision and Image Understanding, 109(1), pp. 
56-68. 
Pradhan, P., King, R., Younan, N., and Holcomb, D., 2006. 
Estimation of the number of decomposition levels for a 
wavelet-based multiresolution multisensor image fusion. IEEE 
TGRS, 44(12), pp. 3674-3686. 
Shi, W., Zhu, C., Zhu, C., and Yang, X., 2003. Multi-band 
wavelet for fusing SPOT panchromatic and multispectral 
images. PE & RS, 69(5), pp. 513-520. 
Starovoitov, V., Makarau, A., Zakharov, I., and Dovnar, D., 
2007. Fusion of reconstructed multispectral images. In IEEE 
International Geoscience and Remote Sensing Symposium, 
Barcelona, Spain, pp. 5146-5149. 
Wald, L., Ranchin, T., and Mangolini, M., 1997. Fusion of 
satellite images of different spatial resolutions: assessing the 
quality of resulting images. PE & RS, 63(6), pp. 691-699. 
Wang, Z., Bovik, A., Sheikh, H., and Simoncelli, E., 2004. 
Image quality assessment: From error visibility to structural 
similarity, IEEE Transactions on Image Processing, 13(4), pp. 
600-612. 
Wang, Z., Ziou, D., Armenakis, C., Li, D., and Li, Q., 2005. A 
Comparative Analysis of Image Fusion Methods, IEEE TGRS, 
43(6), pp. 1391-1402. 
Wang, Z., and Bovik, A., 2009. Mean squared error: love it or 
leave it? - A new look at signal fidelity measures. IEEE Signal 
Processing Magazine, 26(1), pp. 98-117. 
Welch, R., Ehlers, W., 1987. Merging multiresolution SPOT 
HRV and Landsat TM data. PE & RS, 53(3), pp. 301-303.