高光谱特征辨别潴育型麻沙泥田和潮沙泥田水稻土

Discrimination between hydromorphic alluvial sandy mud paddy and tide sandy mud paddy based on hyperspectral characteristics

为了快速识别和划分不同类型水稻土,该研究以江西省兴国县潴育型麻沙泥田和潴育型潮沙泥田为研究对象,以样点方差分析为基础,采用光谱重采样处理、包络线去除处理及光谱一阶微分变换3种光谱分析方法,提取有效的光谱吸收特征参数(spectral absorption feature parameter,SAFP),分析得出两土种有效光谱特征。结果表明:识别潴育型水稻土最有效的是波谷波长位置的共性参数,而区分两土种最有效的是波峰宽度、波谷宽度及波谷倾斜角特性参数。重采样光谱特征参数分析中最有效的是波峰宽度:麻沙泥田的波峰宽度为58 nm,而潮沙泥田为27 nm,麻沙泥田波形起点高,下滑速度比潮沙泥田更快,吸收宽度大;去包络线算法提取光谱吸收特征参数发现,辨别麻沙泥田与潮沙泥田最有效的是波谷宽度,前者在波长482 nm处波谷宽度为26 nm,后者在波长484 nm处波谷宽度为178 nm,潮沙泥田波谷宽度大了将近6倍,而且麻沙泥田波谷倾斜角比潮沙泥田更大,麻沙泥田趋于突变型,而潮沙泥田趋于平缓型;光谱一阶微分方法提取特征参数中区分两土种最有效的是波谷倾斜角:麻沙泥田与潮沙泥田在波段471 nm处分别为-0.00573°和0.00573°,前者相邻波谷走向为负向坡,而后者为正向坡。研究可为高光谱技术辨别不同水稻土类型提供参考。

The application of hyperspectral analysis techniques can extract hyperspectral characteristic parameters and identify soil types. In contrast to the traditional methods, this application is more convenient and time-saving. This paper took two types of Typic Hapli-Stagnic Anthrosols （the hydromorphic alluvial sandy mud paddy and the hydromorphic tide sandy mud paddy） in Xingguo County as the research objects to analyze the hyperspectal characteristics and to extract the effective hyperspectral absorption characteristic parameters （SAFP） for the two soils. In the research, we used three hyperspectral methods including the hyperspectral resample process, the continuum removal, and the first-order differential transform of hyperspectral reflectivity. The results showed that the wavelength position of peaks and valleys can be used to effectively identify the two different Anthrosols. The wave peak width, the wave valley width and the wave valley tilt angle were the parameters to distinguish the two types of Anthrosols. In addition, the analysis of the hyperspectral characteristics parameters after the hyperspectral resample process showed that the difference between the two types of soils was the wave peak reflectivity, and the wave peak reflectivity was the most effective parameter. The hydromorphic alluvial sandy mud paddy belonged to the high reflectivity mud type. The declining rate of the hydromorphic alluvial sandy mud paddy hyperspectral reflectivity was larger and the hyperspectral curve absorption width was significantly larger than that of the hydromorphic tide sandy mud paddy. Moreover, the analysis of the hyperspectral characteristics parameters after the continuum removal showed that the wave valley width can be used to distinguish the two types of Typic Hapli-Stagnic Anthrosols. The wave valley widths of the two soils were 26 and 178 nm, respectively in the vicinity of the wavelength 480 nm. Finally, the analysis of the hyperspectral characteristics parameters after the first-order differential transform of hyperspectral reflectivity showed that the wave valley tilt angle was the other effective parameter to distinguish the two types of Anthrosols. The wave valley tilt angle of the hydromorphic alluvial sandy mud paddy and the hydromorphic tide sandy mud paddy were-0.00573＇ and 0.00573＇, respectively at the wavelength 471 nm. This showed that the hyperspectral reflectivity increase trend of the hydromorphic alluvial sandy mud paddy was negative in the vicinity of the wavelength 471 nm, while the hyperspectral reflectivity increase trend of the hydromorphic tide mud paddy was positive in the vicinity of the wavelength 471 nm. This paper can provide a reference for the application of hyperspectral analysis techniques to identify different Stagnic Anthrosols soil.