[投稿] 麻煩大家幫我看看>"<
先說好 我的英文真的很破 不過我盡力了...
不要批太兇>..<
---------------------------------------------------------------------------
Cloud and haze hides the satellite images are the most main noise of remote
sensing. In the past, there are many research attempts to use different
method to carry one the image recover, and also obtains a lot of commendable
result. Nevertheless, actually less of researches regarding how to compare
the achievement in the view point of the general application. Remote sensing
tech has already become the most essential information of environmental
monitor, resource investigate and other research material. Among them, optics
satellite images are also the major data sources. The shortcoming is easy to
influence by atmosphere environment and produce comparatively much noise,
even if may provide multi-spectral and high resolution image.
Middle these disturbances, cloud cover or hide in Taiwan or Asia might be the
most common noise, the cloud have the serious influence to the image, and
uncomplicated to overcome. This phenomenon causes in the inherent
information of image to lose and to reduce massively, and have serious impact
in the monitor application. In this paper we review the recent cloud removal
research developed in fast Fourier transform filter, computer vision, and
image process, and describe the motivation of the concepts within each of
these disciplines.
We can issue the surface reflected the sunlight completely when received, and
the images from sensor will have the reaction of cloud distributions (Liu et
al, 2006). Thus it may be known that the information losing of land surface
reflection cause from the massive solar radial has been reflected by cloud
and haze in the images, and bring the higher gray level value.
Many research utilization different algorithms and the technology carry on
processing to reduce the influence of cloud and haze in the image. Then we
introduce the methods which the past research institute proposed. (1) the
Dark-target approach method(e.g. Teillet, 1995; Chaze, 1988; Chaze, 1989)
(2)Histogram-match method(e.g. Richter, 1996; Artamonov, 1999) (3)the Liang
algorithm method(e.g. Liang, 2001; Liang, 2002) (4)the HOT method(e.g. Zhang
et al., 2002) (5)the Imagery filter method(e.g. Liu, 1984; Zhao et al., 1996)
and (6)Multi-data fusion method(e.g. Du, 2002; Wang, 1999), and each method
have their own advantage and shortcomings. And the image filter and
multi-data method is the mainly method used in our research to process the
cloud and haze of the image.
Above-mentioned methods, at present the cloud-cover defused methods mostly
aim to the multi-spectral image, utilize the different absorption and
transmissivity characteristic of individual wavelength to process the
demotion image. There are two ways to reduce cloud cover influence. One way
is mosaic method. This method mosaic the different images form different
times serious to reduce the atmosphere influence (McClain et al. 1985).
Another way is operation reducing. This way use image filter operation or
image transform operation to reduce the influence (Lai et al 2004). Although
each way has it advantage and disadvantage, for obtaining the information
from the image, the mosaic method in this case is inadvisable. However, when
using operation method to reducing the cloud cover influence, it is using the
whole image to calculate and operate to reduce. It using the whole image to
operation and reduce cloud covering. Therefore, the image information of
uncovered area will be distortion during the process of calculate and
operate. As a result, the information on the image will be erroneous and
missing.
The research in the past using the visualization of arbitrary perspective,
image value statistics and the variation from entropy analysis (e.g., Dong et
al., 2005; Li et al., 2006; Liu et al., 2008; Chen et al., 2007) to evaluated
the image restoration result. But there are few research carry on the benefit
or the diversity in the reality application which between the pre-processing
and after processing. Therefore we will using the image information addendum
and image filter technique to restore the cloud-covered image, and capitalize
the supervise classification to estimated the reinstate achievement. That is
mean we could use this method to understand the effectivity in environmental
observation and management and another research or application of remote
sensing.
以上為第一小段 先麻煩大家看看了 後續還有...
感恩,修改的話回文或是站內信都可以
另外問題太多想要直接跟我說的話
我的MSN是 vessel0904@hotmail.com
可以問我的電話 我再視情況撥給您(窮學生繳不起太多電話費)
--
【in America】 ║【in Taiwan】 ωaterζantasy
trick or treats!!!! ＊ ║ 不給糖就導彈!!!!
╯╰ = = ⊙ / \ ∞ _ _ ║ \ / ∞ _ _|| = =
﹏ \ 3 / 口 ∞∞_ ▽ ║ 口 ♂♂♂♂♂♂ ∞∞ △ -
╯ / =3 ￣￣ ╰╰ ║ <共> ║║║║║║ ￣￣ ︶台 ((█
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.134.110.252
※ 編輯: vessel0904 來自: 140.134.110.252 (04/26 22:05)