【摘要】：云量不僅是影響地氣系統(tǒng)輻射收支平衡的重要參數(shù),同時(shí)也是研究大氣環(huán)流及氣候變化的重要指標(biāo)。云量計(jì)算又與云檢測息息相關(guān),衛(wèi)星云圖分類方法的檢測精度直接影響著云量計(jì)算的準(zhǔn)確率。在衛(wèi)星云圖云檢測處理的實(shí)際應(yīng)用中,擴(kuò)大訓(xùn)練集是提升分類精度途經(jīng)之一。然而,大量的已標(biāo)記數(shù)據(jù)集需要耗費(fèi)大量的人力和物力成本。在遙感領(lǐng)域,現(xiàn)代高分辨率傳感器技術(shù)的飛速發(fā)展,使得收集未標(biāo)記數(shù)據(jù)變得更加容易和經(jīng)濟(jì)。因此,通過少量已標(biāo)記訓(xùn)練樣本和大量未標(biāo)記樣本提高算法的檢測性能就顯得很有意義。本文基于機(jī)器學(xué)習(xí)理論,將主動(dòng)學(xué)習(xí)與極限學(xué)習(xí)機(jī)相結(jié)合,充分挖掘衛(wèi)星云圖分類中大量樣本的有用信息,用以少量已標(biāo)記樣本,快速提高分類器的性能,提高檢測的精度,減少人工標(biāo)記成本。論文完成的主要工作如下:(1)研究極限學(xué)習(xí)機(jī)的樣本不確定性評估策略,用于主動(dòng)在線極限學(xué)習(xí)機(jī),并通過與極限學(xué)習(xí)機(jī)、主動(dòng)支持向量機(jī)以及主動(dòng)極限學(xué)習(xí)機(jī)在4種不同公共數(shù)據(jù)下的性能表現(xiàn),證明了所提出的主動(dòng)在線極限學(xué)習(xí)機(jī)的有效性。(2)運(yùn)用主動(dòng)在線極限學(xué)習(xí)機(jī)進(jìn)行云檢測,對原始衛(wèi)星云圖進(jìn)行樣本提取、預(yù)處理后,已極限學(xué)習(xí)機(jī)作為基本分類器,采用非確定性抽樣提取信息豐富的樣本,進(jìn)行主動(dòng)在線學(xué)習(xí),實(shí)現(xiàn)薄云、厚云、晴空以及薄云和厚云交界的檢測。在不降低分類器性能的前提下,減少樣本人工標(biāo)注成本,縮減分類器訓(xùn)練時(shí)間。通過與閾值法、主動(dòng)支持向量機(jī)、ELM實(shí)驗(yàn)比較,驗(yàn)證本文提出的方法在處理衛(wèi)星云圖數(shù)據(jù)時(shí)的有效性。(3)將檢測后的衛(wèi)星云圖,利用“空間相關(guān)法”在云檢測的基礎(chǔ)上進(jìn)行云量計(jì)算,并與4種不同的算法進(jìn)行對比實(shí)驗(yàn),最后通過與專家標(biāo)定的標(biāo)準(zhǔn)數(shù)據(jù)庫進(jìn)行對比分析,改進(jìn)并完善衛(wèi)星云圖云量計(jì)算模型。
[Abstract]:Cloud cover is not only an important parameter to influence the balance of radiation budget in terrestrial atmosphere system, but also an important index to study atmospheric circulation and climate change. Cloud calculation is closely related to cloud detection. The accuracy of satellite cloud image classification directly affects the accuracy of cloud calculation. In the practical application of cloud detection and processing of satellite cloud image, expanding the training set is one of the ways to improve the classification accuracy. However, a large number of marked data sets require a lot of human and material costs. In the field of remote sensing, the rapid development of modern high-resolution sensor technology makes it easier and more economical to collect unlabeled data. Therefore, it is significant to improve the detection performance of the algorithm through a small number of labeled training samples and a large number of unlabeled samples. Based on the theory of machine learning, this paper combines active learning with extreme learning machine to fully mine the useful information of a large number of samples in satellite cloud image classification, so as to quickly improve the performance of classifier and improve the accuracy of detection by using a small number of labeled samples. Reduce the cost of manual marking. The main work of this paper is as follows: (1) the sample uncertainty evaluation strategy of LLM is studied, which is used for active online LLM, and through LLM, The performance of active support vector machine and active extreme learning machine under four kinds of common data proves the effectiveness of the proposed active online extreme learning machine. (2) Cloud detection is carried out by using active online extreme learning machine. After preprocessing, the extreme learning machine is used as the basic classifier, the samples with abundant information are extracted by non-deterministic sampling, and active online learning is carried out to realize the thin cloud and thick cloud. Clear skies and the detection of the boundary between thin clouds and thick clouds. Without reducing the performance of classifier, the cost of manual labeling of samples is reduced, and the training time of classifier is reduced. Comparing with the threshold method and the ELM experiment of active support vector machine, the validity of the proposed method in the processing of satellite cloud image data is verified. (3) the detected satellite cloud image will be obtained. The spatial correlation method is used to calculate cloud amount on the basis of cloud detection, and compared with four different algorithms. Finally, it is compared and analyzed with the standard database calibrated by experts. The cloud volume calculation model of satellite cloud image is improved and improved.
【學(xué)位授予單位】：南京信息工程大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：P412.27;TP18

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 黃廣斌;;超限學(xué)習(xí)機(jī)筑夢智能物聯(lián)網(wǎng)時(shí)代[J];軟件和集成電路;2017年01期

2 靳澤群;張玲;劉神聰;易凡;;基于BP神經(jīng)網(wǎng)絡(luò)的云檢測和云相態(tài)識別[J];光學(xué)與光電技術(shù);2016年05期

3 段煉;郭庚常;;風(fēng)云衛(wèi)星云圖自動(dòng)識別追蹤MCS的方法[J];電子科技;2016年04期

4 楊富圣;高鐵杠;;一種基于差分DCT系數(shù)對直方圖的圖像取證方法[J];電子學(xué)報(bào);2016年01期

5 王保義;趙碩;張少敏;;基于云計(jì)算和極限學(xué)習(xí)機(jī)的分布式電力負(fù)荷預(yù)測算法[J];電網(wǎng)技術(shù);2014年02期

6 付倩;韓飛;葉松林;;一種改進(jìn)的集成在線順序極限學(xué)習(xí)機(jī)[J];無線通信技術(shù);2013年03期

7 胡義明;梁忠民;王軍;楊好周;;考慮抽樣不確定性的水文設(shè)計(jì)值估計(jì)[J];水科學(xué)進(jìn)展;2013年05期

8 王珍鈺;王熙照;;基于近鄰熵的主動(dòng)學(xué)習(xí)算法[J];模式識別與人工智能;2011年01期

9 賀衛(wèi)中;周維娜;束平;;基于改進(jìn)支持向量機(jī)方法的影像分類和目標(biāo)識別研究[J];遙感信息;2010年06期

10 王衛(wèi)星;王李平;員志超;;一種基于最大類間后驗(yàn)概率的Canny邊緣檢測算法[J];計(jì)算機(jī)應(yīng)用;2009年04期

相關(guān)博士學(xué)位論文 前1條

1 姚明臣;機(jī)器學(xué)習(xí)和神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)中的若干問題研究[D];大連理工大學(xué);2016年

相關(guān)碩士學(xué)位論文 前1條

1 關(guān)文淵;風(fēng)電功率預(yù)測不確定性分析及應(yīng)用研究[D];華北電力大學(xué)(北京);2016年

，

本文編號：2242292