光谱学与光谱分析
光譜學與光譜分析
광보학여광보분석
SPECTROSCOPY AND SPECTRAL ANALYSIS
2014年
1期
191-195
,共5页
高光谱遥感%地形起伏%地气耦合辐射%成像仿真
高光譜遙感%地形起伏%地氣耦閤輻射%成像倣真
고광보요감%지형기복%지기우합복사%성상방진
Hyperspectral remote sensing%Rugged scenes%Earth-atmosphere coupling%Simulation
邻近效应会影响高光谱遥感的定量化应用,而地气耦合辐射是邻近效应的重要组成部分。高光谱遥感数据多受地形背光和阴影的影响,不利于描述地物的光谱特性,并且地形的相互遮蔽使得辐射在大气、地表之间的耦合过程变得更加复杂。为了满足高光谱数据模拟快速性的要求,在深入分析辐射传输过程的基础上,利用邻近地物和目标地物的相对方位以及局地地形特征描述背景等效反射率,计算地气耦合辐射对传感器入瞳辐亮度的贡献,从而在起伏地形下实现了精确的地气耦合辐射建模。将此建模方法应用于高光谱遥感成像仿真,通过对实测图像的模拟,验证模型的有效性。结果表明,模拟图像与实测图像在视觉效果上具有较好的一致性;在地形起伏区域,仿真中采用改进后的地气耦合模型可提高模拟图像与实测图像的光谱相似度,同时在地形平坦区域保持了较高的仿真精度。
鄰近效應會影響高光譜遙感的定量化應用,而地氣耦閤輻射是鄰近效應的重要組成部分。高光譜遙感數據多受地形揹光和陰影的影響,不利于描述地物的光譜特性,併且地形的相互遮蔽使得輻射在大氣、地錶之間的耦閤過程變得更加複雜。為瞭滿足高光譜數據模擬快速性的要求,在深入分析輻射傳輸過程的基礎上,利用鄰近地物和目標地物的相對方位以及跼地地形特徵描述揹景等效反射率,計算地氣耦閤輻射對傳感器入瞳輻亮度的貢獻,從而在起伏地形下實現瞭精確的地氣耦閤輻射建模。將此建模方法應用于高光譜遙感成像倣真,通過對實測圖像的模擬,驗證模型的有效性。結果錶明,模擬圖像與實測圖像在視覺效果上具有較好的一緻性;在地形起伏區域,倣真中採用改進後的地氣耦閤模型可提高模擬圖像與實測圖像的光譜相似度,同時在地形平坦區域保持瞭較高的倣真精度。
린근효응회영향고광보요감적정양화응용,이지기우합복사시린근효응적중요조성부분。고광보요감수거다수지형배광화음영적영향,불리우묘술지물적광보특성,병차지형적상호차폐사득복사재대기、지표지간적우합과정변득경가복잡。위료만족고광보수거모의쾌속성적요구,재심입분석복사전수과정적기출상,이용린근지물화목표지물적상대방위이급국지지형특정묘술배경등효반사솔,계산지기우합복사대전감기입동복량도적공헌,종이재기복지형하실현료정학적지기우합복사건모。장차건모방법응용우고광보요감성상방진,통과대실측도상적모의,험증모형적유효성。결과표명,모의도상여실측도상재시각효과상구유교호적일치성;재지형기복구역,방진중채용개진후적지기우합모형가제고모의도상여실측도상적광보상사도,동시재지형평탄구역보지료교고적방진정도。
Adjacency effects may introduce errors in the quantitative applications of hyperspectral remote sensing ,of which the significant item is the earth-atmosphere coupling radiance .However ,the surrounding relief and shadow induce strong changes in hyperspectral images acquired from rugged terrain ,which is not accurate to describe the spectral characteristics .Furthermore , the radiative coupling process between the earth and the atmosphere is more complex over the rugged scenes .In order to meet the requirements of real-time processing in data simulation ,an equivalent reflectance of background was developed by taking into account the topography and the geometry between surroundings and targets based on the radiative transfer process .The contri-butions of the coupling to the signal at sensor level were then evaluated .This approach was integrated to the sensor-level radi-ance simulation model and then validated through simulating a set of actual radiance data .The results show that the visual effect of simulated images is consistent with that of observed images .It was also shown that the spectral similarity is improved over rugged scenes .In addition ,the model precision is maintained at the same level over flat scenes .