RESOLUTION MERGE OF 1:35.000 SCALE AERIAL PHOTOGRAPHS WITH LANDS AT
7 ETM IMAGERY
M. Erdogan, H.H. Maras, A. Yilmaz, Ô.T. Ôzerbil
General Command of Mapping 06100 Dikimevi, Ankara, TURKEY - (mustafa.erdogan; hakan.maras; altan.yilmaz;
tuncer.ozerbil)@ hgk.mil.tr
Commission VII, WG VII/6
KEY WORDS: Aerial Photographs, LANDSAT, Resolution Merge, Image Processing, Remote Sensing
ABSTRACT:
Merging of different data sets is often used in digital image processing to improve the visual and analytical quality of the data. The
analyst may need to merge different types of data. In this process, different data such as satellite imagery from the same sensor but
with different resolution, satellite imagery from different sensors with varying resolution, digitized aerial photography and satellite
imagery or satellite imagery with ancillary information can be merged. In this paper, the efficiency of three different merging
techniques (Principal Component, IHS, and Brovey Transform) is examined, in order to improve the spatial resolution of very high
resolution aerial photographs (1:35.000 scale panchromatic) with the Landsat 7 ETM imagery. The aim was to get best enhanced
merged aerial imagery for the visual interpretation. Because of the very big difference between the resolutions of sources, the
techniques give very different results. The general conclusion is that when the original source imagery is used, Principal Component
and Brovey Transform merging techniques should be preferred for such kind of imagery. Other methods were also tested to enhance
the merge imagery, such as, merging the multispectral Landsat 7 ETM imagery with Landsat 7 ETM panchromatic imagery at first
and merging this imagery with aerial photographs again with three different merging techniques afterwards. In another method,
multispectral Landsat 7 ETM imagery was resampled to higher resolution imageries and then panchromatic aerial imagery was
merged with this resampled image with three different merging techniques. In all approaches, Brovey Transform and Principal
Component techniques serve well the purpose of increasing resolution of the low resolution images with the high resolution images.
However, all methods should be tested in different areas by using multispectral and panchromatic images which were taken in
different time frames to define the overall performances of these methods and merging techniques.
1. INTRODUCTION
1.1 Resolution Merge
Digital images taken by airborne or spacebome sensors are very
frequently used in earth sciences and applications. The
increasing applications are due to the availability of high
quality images for a reasonable price and improved
computation power. Nowadays there is a wide range of systems
that provide images in digital format, and their interpretation
into terrestrial attributes is very dependent on their spatial and
spectral resolution. As a result of the demand for higher
classification accuracy and the need in enhanced positioning
precision there is always a need to improve the spectral and
spatial resolution of remotely sensed imagery. For most of the
systems, panchromatic images typically have higher resolution,
while multispectral images offer information in several spectral
channels. Resolution merge (also called pan-sharpening) allows
us to combine advantages of both kinds of images by merging
them into one.
A variety of resolution merging techniques is available and
described by several authors. The most common techniques are
implemented in standard image processing software packages
(IHS , PCA, Multiplicative and Brovey transforms) It is hard to
categorise the techniques into a limited number of main types
(Pohl, C.,1999, Hill et al.,1999, Bretschneider et al.,2004). In
another study conducted by Carvalho et al.,(2006), a
geostatistical merging methodology based on direct sequential
cosimulation with reference images of LandsatTM and SPOT-P
is tested. With the stochastic simulation one generates a high
spatial resolution image with the characteristics of the of the
higher spectral resolution image. It is an iterative inverse
optimization procedure that tends to reach the matching of an
objective function by preserving the spectral characteristics and
spatial pattern, as revealed by the variograms, of the higher-
spectral resolution images both in terms of descriptive statistics
and band correlation coefficients.
Digital image-merging procedures are techniques that aim at
integrating the multispectral characteristics in a high spatial
resolution image, thus producing synthetic images that combine
the advantages of both types of images. The main constraint is
to preserve the spectral information for tasks like classification
of ground cover. Ideally, the method used to merge datasets
with high-spatial resolution and high-spectral resolution should
not distort the spectral characteristics of the high spectral
resolution data. Not distorting the spectral characteristics is
important for calibrating purposes and to ensure that targets that
are spectrally separable in the original data are still separable in
the merged dataset (Chavez et al.,1991). Several methods for
spatial enhancement of low-resolution imagery combining high
and low resolution data can be used. Some widely used ones are:
Intensity-Hue Saturation, Principal Component, Multiplicative,
Color Normalized, and Brovey Transform. Before the
resolution merge, the source imagery need to be processed to
have the same reference system. The images have to be
accurately registered to the same reference system and to one
another.