結(jié)合多尺度空間濾波和層級(jí)網(wǎng)絡(luò)的基于結(jié)構(gòu)保持的高光譜特征選擇
發(fā)布時(shí)間:2018-01-16 11:20
本文關(guān)鍵詞:結(jié)合多尺度空間濾波和層級(jí)網(wǎng)絡(luò)的基于結(jié)構(gòu)保持的高光譜特征選擇 出處:《光子學(xué)報(bào)》2017年05期 論文類型:期刊論文
更多相關(guān)文章: 高光譜圖像 特征選擇 雙邊濾波 空間近鄰 流形學(xué)習(xí) 層級(jí)網(wǎng)絡(luò)
【摘要】:為了充分利用高光譜圖像蘊(yùn)含的豐富的光譜信息和空間信息,提出了結(jié)合多尺度空間濾波和層級(jí)網(wǎng)絡(luò)的基于結(jié)構(gòu)保持的高光譜特征選擇算法.算法利用基于l2,1范數(shù)的數(shù)學(xué)模型,選出同時(shí)保存全局相似性結(jié)構(gòu)和局部流形結(jié)構(gòu)的特征子集;在多個(gè)尺度的窗口中使用雙邊濾波,自適應(yīng)計(jì)算濾波核,自動(dòng)在光譜數(shù)據(jù)中融入空間信息,增強(qiáng)了類內(nèi)相似性和類間相異性,避免了參量選擇;引入層級(jí)結(jié)構(gòu)實(shí)現(xiàn)空間信息和光譜信息的深入融合,提高了分類準(zhǔn)確度;討論了層級(jí)數(shù)目和窗口尺度個(gè)數(shù)對(duì)分類準(zhǔn)確度的影響.在Indian Pines和PaviaU兩個(gè)數(shù)據(jù)集的實(shí)驗(yàn)表明,該算法在大部分地物種類上的分類準(zhǔn)確度都有較大幅度的提升,總體分類準(zhǔn)確度分別達(dá)到90.98%和94.20%,相比其他方法明顯提高了地物分類準(zhǔn)確度.
[Abstract]:In order to make full use of the rich spectral information and spatial information contained in hyperspectral images. A hyperspectral feature selection algorithm combining multi-scale spatial filtering and hierarchical network is proposed. The algorithm uses a mathematical model based on l2m-1 norm. The feature subsets of both global similarity structure and local manifold structure are selected. Using bilateral filtering in multi-scale windows, adaptive computing filter kernel, automatically incorporating spatial information into spectral data, enhancing intra-class similarity and inter-class differences, avoiding parameter selection; The hierarchical structure is introduced to realize the deep fusion of spatial information and spectral information, and the classification accuracy is improved. The effects of the number of levels and the number of window scales on the classification accuracy are discussed, and the experimental results of Indian Pines and PaviaU are given. The classification accuracy of the algorithm has been greatly improved in most kinds of ground objects, and the overall classification accuracy has reached 90.98% and 94.20%, respectively. Compared with other methods, the classification accuracy of ground objects is improved obviously.
【作者單位】: 火箭軍工程大學(xué)信息工程系;中國(guó)科學(xué)院西安光學(xué)精密機(jī)械研究所;西安交通大學(xué)電信學(xué)院;
【基金】:國(guó)家自然科學(xué)基金(No.61401471) 中國(guó)博士后基金(No.2014M562636)資助~~
【分類號(hào)】:TP751
【正文快照】: 1-3000150conventional methods.0引言近年來(lái)高光譜遙感技術(shù)發(fā)展迅速,更高的光譜和空間分辨率能捕獲地物精細(xì)的光譜響應(yīng)及空間細(xì)節(jié)特征,帶來(lái)了更精細(xì)的光譜波段和更豐富的地物信息,對(duì)分類任務(wù)提出了更高要求.高光譜圖像數(shù)據(jù)集(圖像立方體)具有圖譜合一的特點(diǎn)[1-5].雖然豐富的
【相似文獻(xiàn)】
相關(guān)期刊論文 前1條
1 王圓圓;李京;;基于決策樹(shù)的高光譜數(shù)據(jù)特征選擇及其對(duì)分類結(jié)果的影響分析[J];遙感學(xué)報(bào);2007年01期
,本文編號(hào):1432820
本文鏈接:http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/1432820.html
最近更新
教材專著