METHODS AND APPLICATION OF QUALITY ASSESSMENT FOR REMOTE SENSING
IMAGE COMPRESSION
ZHAI Liang 3 ’ *, TANG Xinming 3 , ZHANG Guo b , ZHU Xiaoyong b
3 Chinese Academy of Surveying and Mapping, 16 Beitaiping Road, Beijing, China - (zhailiang, tang)@casm.ac.cn
b State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University,
Wuhan, China - guozhang@whu.edu.cn, zhuxytop@163.com
KEY WORDS: image compression, quality assessment for image compression, JPEG2000, SPIHT, subjective assessment,
objective assessment
ABSTRACT:
Quality assessment for remote sensing image compression is critical to compression algorithm designers and image products users.
In order to satisfy the users’ demand, image compression index such as image compression ratio should be set down before the
launch of satellite. Quality assessment for remote sensing image compression is an important objective reference for specifying
image compression ratio. Effects of compression on image quality include image interpretability quality and geometric quality.
Centering on subjective and objective assessment, image quality assessment methods and application are proposed for mapping
using remote sensing images. A prototype system for image quality assessment is proposed and applied to Resources Satellite 3
compression ratio design. Through a series of experiments, compression ratio is considered to be no more than 4:1 if JPEG2000 or
improved SPIHT algorithm were employed on board.
1. INTRODUCTION
Data compression is the process of encoding information using
fewer bits than an unencoded representation through use of
specific encoding schemes. Image compression is one of the
important aspects of data compression. In the realm of satellite
remote sensing, image compression is very important because it
helps reduce the consumption of expensive resources, such as
hard disk space or transmission bandwidth. According to the
stage of satellite remote sensing image data processing,
application of image compression in satellite remote sensing
can be divided into two aspects. One is on-board image data
compression, the other is ground image data compression.
Quality assessment for remote sensing image compression is
critical to compression algorithm designers and image products
users. In order to satisfy the users’ demand, image compression
index such as image compression ratio should be set down
before the launch of satellite. Quality assessment for remote
sensing image compression is an important objective reference
for specifying image compression ratio.
Many image compression algorithms have emerged in the last
ten years, while the quality assessment is ignored especially in
the realm of application, such as surveying and mapping. This
paper aims at proposing an integral scheme for remote sensing
image quality assessment. Subjective assessment method is
examined first. Furthermore, objective assessment including
imaging quality assessment and geometric quality assessment is
discussed. Finally, an image quality assessment software is
introduced.
2. SUBJECTIVE ASSESSMENT
Images, which are the result of imaging process mapping
physical scene properties onto a two-dimensional luminance
distribution, are carriers of visual information [T.J.W.M.
Janssen, 1999]. It is a visuo-cognitive process when an observer
views reconstructed images. Subjective assessment is regarded
as a golden rule of image quality assessment since human
observers are the ultimate receivers in most image processing
environments [Michael P. Eckert et al, 1998; Zhou Wang et al,
2002a], Usually, subjective quality assessment can be divided
into quantitative and qualitative assessment. Quantitative
assessment can be divided into absolute and relative assessment.
Absolute assessment has a group of reference images while
relative assessment not. In general, quantitative assessment
reaches a score according to some rating scale. Ahmet M.
Eskicioglu has concluded the rating scales having been used by
former researchers [Ahmet M. Eskicioglu, 2000]. Qualitative
assessment divides reconstructed image into different grades
according to grades descriptions. The American aerial imaging
community utilizes the National Imagery Interpretability Rating
Scale (NIIRS) to define and measure the quality of images and
performance of imaging systems. Through a process referred to
as "rating" an image, the NIIRS is used by imagery analysts to
assign a number which indicates the interpretability of a given
image. The NIIRS concept provides a means to directly relate
the quality of an image to the interpretation tasks for which it
may be used [K. J. Hermiston et al, 1999]. In essence, the
difference between qualitative and quantitative assessment is
the assessment result. Qualitative assessment result will be
scores, while quantitative assessment result be grades and
corresponding descriptions. Both qualitative assessment and
quantitative assessment are common methods used to describe
reconstructed image quality.
Fuzzy phenomena is everywhere in the nature, such as
meteorological phenomena, land cover classification, spatial
data quality, etc. Image compression quality is also fuzzy. One
can not tell good or bad simply. Further, image compression
quality is affected by several distortions and every type of
distortion has different effects on image quality. As to this, a
fuzzy comprehensive evaluation method is used in image
Corresponding author: Male, PhD., majors in satellite remote sensing, spatio-temporal data mining , etc.
