基于SVM方法的航空高光谱赤潮信息提取与分析

Airborne hyperspectral red tide information extraction and analysis based on SVM

从仪器到外业数据采集再到内业数据处理,较全面的阐述了航空高光谱赤潮监测的工作流程,并着重讨论了航空高光谱赤潮信息提取这一关键技术.采用专门针对小样本高维度的支持向量机模式识别方法,以2013年7月4日浙江舟山东福山赤潮高光谱数据为例,通过采集7个点位的赤潮海水和正常海水波谱曲线,研究分析赤潮海水的差异性特征,开展赤潮信息提取技术的研究,并对相同样本不同的提取方法进行了对比,通过分析,进一步证明了支持向量机方法的准确性和优越性.航空高光谱赤潮信息提取技术的研究可为航空赤潮业务化监测提供必要的技术基础.

This article comprehensively expounds the red tide monitoring process by airborne hyperspectral from instrument to outdoor data acquisition and inside data processing. Red tide information extraction was discussed in detail. Taking the data of 4 July, 2013 as an example, SVM method was proved the most accuracy among other information extraction methods by the same samples. It was also proved by 7 spectral curves of red tides and normal sea water. Study of red tide information extraction by airborne hyperspeetral can provide technical support for operational monitoring of marine red tide.