本文關(guān)鍵詞： 衛(wèi)星圖像 Mallat算法 薄云 多尺度分析 非線性增強(qiáng) 中值濾波　出處：《中國科學(xué)技術(shù)大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：高分衛(wèi)星自投入使用以來,被廣泛的應(yīng)用于災(zāi)害監(jiān)測、資源勘查以及環(huán)境保護(hù)等許多領(lǐng)域。而高分辨率衛(wèi)星數(shù)據(jù)使得在小的空間尺度上面進(jìn)行地表細(xì)節(jié)變化的觀察以及完成人為活動對環(huán)境影響的檢測等變?yōu)榭赡?具有重要的意義。由于搭載在衛(wèi)星上的高空間分辨率成像設(shè)備獲取到的遙感圖像數(shù)據(jù)會受到云的干擾,數(shù)據(jù)質(zhì)量存在不同程度的下降。當(dāng)天空中存在厚云遮擋的時候,下墊面信息會完全丟失;而在薄云覆蓋區(qū),圖像的質(zhì)量雖然會退化,但仍有可供利用的下墊面信息。為了提高圖像定量解譯的水平和圖像信息的利用率,有效地去除高分辨衛(wèi)星圖像中薄云的影響,本文針對圖像的薄云去除進(jìn)行了研究。論文的主要內(nèi)容以及結(jié)論如下:(1)首先對我們要處理圖像的類型以及表示方法進(jìn)行了介紹,然后對常規(guī)遙感圖像的退化模型以及受薄云影響的圖像的成像模型進(jìn)行了總結(jié),并從空間特征和頻率特征這兩個方面對遙感圖像云區(qū)的特征進(jìn)行了分析;(2)提出本文所用的方法,對圖像作Mallat小波分解得到高頻細(xì)節(jié)部分和低頻近似部分,依據(jù)云噪聲在分解系數(shù)中處于低頻部分而地物信息占據(jù)相對高頻部分的特點,在多尺度分析的基礎(chǔ)上,算法在最大尺度低頻圖像上按照云厚度掩模值對云區(qū)進(jìn)行線性處理;對于高頻子帶圖像根據(jù)尺度的不同運用非線性增強(qiáng)算子進(jìn)行不同程度的增強(qiáng),從而提高圖像的清晰度,最后對經(jīng)過重構(gòu)后的圖像作中值濾波以減少高頻云的影響。針對高分一號衛(wèi)星圖像進(jìn)行了試驗,試驗證明該方法能夠取得比較好的效果。(3)設(shè)計薄云去除軟件,在功能上實現(xiàn)了 TIFF圖像的讀取、顯示、保存、子圖像截取、小波分解與重構(gòu)、2種薄云去除的方法以及退出程序等操作并給出了相應(yīng)的處理結(jié)果圖。(4)以高分一號衛(wèi)星數(shù)據(jù)為例進(jìn)行試驗并將該方法與傳統(tǒng)小波變換法進(jìn)行比較分析,,該方法在除薄云的同時很好的保持了圖像細(xì)節(jié)信息,去薄云效果優(yōu)于傳統(tǒng)小波變換法,論證了本文方法的有效性。
[Abstract]:Since the high score satellite put into use, is widely used in many fields of disaster monitoring, resource exploration and environmental protection. The high resolution satellite data that were observed in the surface details of the changes above small spatial scale and influence of human activities on the environment to complete the detection of possible, is of great significance. Because the remote sensing image the data with high spatial resolution imaging equipment on the satellite access to the cloud will be subject to interference, has the low quality of the data. In the sky when there is a thick cloud cover, surface information will be completely lost; and covered in thin cloud area, although the image quality will be degraded, but there are still under the mat the information available. In order to improve utilization level of image information and image quantitative interpretation of the rate, effectively remove the influence of high resolution satellite images in the cloud, according to the image Thin cloud removal was studied. The main contents and conclusions are as follows: (1) first said methods are introduced for us to deal with the types of images and then the degradation model of conventional remote sensing images and imaging model image by thin cloud effects are summarized, and the characteristics of the remote sensing image from the cloud area these two aspects of spatial features and frequency characteristics are analyzed; (2) the method used in this paper, the Mallat wavelet decomposition to obtain high frequency details and low frequency approximation part of the image, based on the cloud noise in the low frequency part of the wavelet coefficient and the feature information occupy a relatively high frequency part characteristics, based on multiscale analysis on the algorithm in maximum scale low frequency image according to the thickness of cloud mask value the cloud region linear processing; the high-frequency subband images according to different nonlinear scale enhancement operator is not With the enhancement degree, so as to improve the clarity of the image, finally through the reconstructed image median filtering to reduce the influence of high-frequency cloud. For a high score satellite images of the test, the experiment proves that the method can achieve better results. (3) software to remove thin cloud design, realizes the function of reading TIFF, image display, storage, image capture, wavelet decomposition and reconstruction, 2 methods of thin cloud removal and exit procedures and gives the processing results of the corresponding graph. (4) to score the 1st satellite data as an example of the test and the method with the traditional wavelet transform method were compared and analyzed. This method, in addition to the cloud while preserving the image details, to better than the thin cloud effect of traditional wavelet transform method, demonstrate the effectiveness of the proposed method.

【學(xué)位授予單位】：中國科學(xué)技術(shù)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP751

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 劉澤樹;陳甫;劉建波;孫業(yè)超;;改進(jìn)HOT的高分影像自動去薄云算法[J];地理與地理信息科學(xué);2015年01期

2 文奴;楊世植;崔生成;程偉;;自適應(yīng)雙樹復(fù)小波遙感圖像復(fù)原[J];強(qiáng)激光與粒子束;2014年10期

3 曹平;章文毅;馬廣彬;;遙感衛(wèi)星成像模型研究及仿真[J];遙感信息;2014年03期

4 李海巍;楊敏華;劉益世;;基于中值濾波和小波變換的遙感影像去云[J];測繪與空間地理信息;2012年07期

5 王靜;徐妮妮;楊雪玲;;基于周期延拓的純二維小波變換[J];天津工業(yè)大學(xué)學(xué)報;2012年02期

6 楊慶怡;;圖像復(fù)原方法研究[J];計算機(jī)光盤軟件與應(yīng)用;2012年07期

7 吳素霞;宋士濤;張電學(xué);楊會民;謝新宇;常連生;;EOS/MODIS遙感影像云剔除方法[J];河北科技師范學(xué)院學(xué)報;2011年04期

8 鄭玉鳳;李海濤;顧海燕;;基于環(huán)境衛(wèi)星CCD影像的薄云去除研究[J];遙感信息;2011年03期

9 王敏;周樹道;劉志華;黃峰;梁妙元;;遙感圖像薄云薄霧的去除處理方法[J];實驗室研究與探索;2011年02期

10 占必超;吳一全;紀(jì)守新;;基于平穩(wěn)小波變換和Retinex的紅外圖像增強(qiáng)方法[J];光學(xué)學(xué)報;2010年10期

相關(guān)博士學(xué)位論文 前1條

1 周麗雅;受云霧干擾的可見光遙感影像信息補(bǔ)償技術(shù)研究[D];解放軍信息工程大學(xué);2011年

相關(guān)碩士學(xué)位論文 前9條

1 喻昕s，

本文編號：1456461