先说好 我的英文真的很破 不过我尽力了... 不要批太凶>..< --------------------------------------------------------------------------- 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. 以上为第一小段 先麻烦大家看看了 後续还有... 感恩,修改的话回文或是站内信都可以 另外问题太多想要直接跟我说的话 我的MSN是 [email protected] 可以问我的电话 我再视情况拨给您(穷学生缴不起太多电话费) -- 【in America】 ║【in Taiwan】 ωaterζantasy trick or treats!!!! ＊ ║ 不给糖就导弹!!!! ╯╰ = = ⊙ / \ ∞ _ _ ║　 \ / ∞ _ _|| = = ﹏ \ 3 / 口 ∞∞_ ▽ ║ 口 ♂♂♂♂♂♂ ∞∞ △ - ╯ / =3 ￣￣ ╰╰ ║ <共> ║║║║║║ ￣￣ ︶台 ((█ --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 140.134.110.252 ※ 编辑: vessel0904 来自: 140.134.110.252 (04/26 22:05)