基于气象卫星云图的红外吸收带火山特征分析

飞行目标在2.7和4.3μm谱段附近具有较强的红外辐射,因此这两个波段是探测飞行目标的最佳波段,但是由于这两个波段并非大气窗口,不被大多数遥感器包含。对这两个谱段的典型地物特性开展研究具有重要的价值,但由于缺乏必要的数据获取能力,经常面临数据缺乏的问题。世界范围内频频有各种程度的火山爆发,火山爆发时温度较高的火山口,是否对于天基红外探测系统典型目标探测有影响,其影响程度如何一直缺乏相关的分析和研究。基于大气辐射传输理论,利用多元统计分析得到波段转换模型,使用气象卫星已有波段获得红外吸收谱段数据。将火点像元视为明火和背景的混合像元,采用目标与背景分离的方式描述高温目标像元的热辐射。对高温目标辐射量,在气溶胶模式固定的情况下,选取观测天顶角、大气可降水量、大气廓线为自变量影响因子。对于背景辐射亮度关系,选取观测天顶角、大气可降水量、大气廓线、背景温度为自变量影响因子,利用多元统计,建立相关模型。利用对地面热状态非常敏感的风云三号可见光红外扫描辐射计第3通道数据的统计特征探测火山口,获取高温目标在特定波段的表观多维特征并定量分析。火山的多维特征分析,主要从时间和空间两个维度展开。时间维度是对同一火山在不同时间的数据进行分析,空间维度上,主要统计火山口的辐射亮度和亮度温度的空间分布特征。一般气象卫星分辨率较低,单纯利用像元个数表示火山面积,明显夸大了火山的实际面积,所以基于亚像元特性对火山进行分析,将混合像元火点视为明火和背景的组合,运用线性光谱混合模型,通过混合像元的辐射率精确计算火山高温点的面积和温度,提高定量分析精度。分析结果表明:通过仿真手段结合多元统计分析方法建立高温目标的波段转换模型是一种可行的预研手段。在2.7～2.95μm谱段,火山口在弱背景环境下可能会对高温目标造成干扰,而在4.2～4.45μm谱段,火山口能量远高于一般地表类型,是不可忽视的干扰。

Analysis of Infrared Absorption Band for Volcano Based on Meteorological Satellite Cloud Image

The targets have strong infrared radiation near the 2.7 and 4.3 μm bands, so these two bands are also the best bands to detect the flying target, but since these two bands are not atmospheric windows, they are not included in most remote sensors. It is of great value to study the typical features of infrared absorption spectrum. However, due to the lack of necessary data acquisition ability, it often faces the problem of lack of data. There are frequent volcanic eruptions in various parts of the world. Whether volcanic eruptions have an impact on typical target detection or not is lacking in relevant analysis and research. In this paper, we get the data which transformed from the meteorological satellite data, through the wave band conversion model, based on the theory of atmospheric radiation transmission and the multivariate statistical analysis. The High temperature pixel is regarded as a mixed pixel of flame and background, and the target radiation is separated from the background to describe the thermal radiation of high temperature target pixel. When the aerosol mode is fixed, the observation zenith angle, atmospheric precipitation and the atmospheric profile are the influencing factors of the independent variable. For the background radiation brightness, the observation zenith angle, atmospheric precipitation in the atmosphere, atmospheric profile and background temperature are the influential factors of independent variables, and multivariate statistics are used to establish the relevant models. The volcano was detected by using the statistical characteristics of the third channel data of the FY-3 VIRR to obtain the apparent multidimensional features and quantitative analysis in time dimension, and the data of the same volcano are analyzed at different times. In spatial dimension, the spatial distribution characteristics of radiation brightness and luminance temperature of the crater are statistically calculated. The resolution of normal meteorological satellite is quite low, if we use pixel resolution to represent the volcanic area, the actual area of the volcano will be significantly exaggerated. So in this paper we propose a sub-pixel characteristics analysis method to increase the quantitative analysis accuracy. A combination of a mixture of pixels is considered as a combination of flame and background, and a linear spectral mixture model is used to accurately calculate the area and temperature of a volcano’s high temperature point by the emissivity of the mixed pixel. The results show that the 2.7~2.95 μm crater may interfere with the high temperature target in the weak background environment. In 4.2~4.45 μm band, the crater has been proved to be a potential disturbance that can not be ignored. Its energy is much higher than the general surface type.