| 摘要: |
| 【目的】基于实测光谱数据通过光谱特征分析，构建与选取适用于乌梁素海复杂水环境的挺水植被类群、沉水植被类群和漂浮藻类的光谱指标，构建决策树分类模型进行研究区长时间序列的植被类群与藻华监测。【方法】本文首先提出MNDSI光谱指数进行水生植被类群和黄苔藻华之外类型的排除，其次利用NDWI光谱水体指数提取明水区，然后提出改进的增强型植被指数（AEVI）提取挺水植被，最后利用改进的大型藻类指数（MFAI）来区分黄苔藻华与沉水植被类群。基于以上工作构建决策树分类模型实现黄河流域典型内陆湖泊水生植被类群和藻华遥感监测。将该方法应用于1986—2018年Landsat影像，研究乌梁素海水生植被类群和黄苔藻华的的时空分布特征和变化趋势。【结果】通过实测数据和GF-2 PMS影像人工视觉解译对分类模型的精度验证的总体精度均大于90%，研究表明光谱指数结合决策树分类模型是区分水生植被与黄苔藻华的有效方法，可用于分析乌梁素海植被类群和藻华时空分布、变化分析及藻华爆发事件监测。【结论】1986—2018年挺水植被类群呈现缓慢增长趋势。1986—2018年，沉水植被类群的变化特征为：1986—2013年呈现减少趋势，2013年以后具有快速增长趋势。1986—2018年，黄苔藻华无明显变化趋势，但具有爆发性。 |
| 关键词: 水生植被类群；黄苔藻华；光谱指数；决策树；乌梁素海 |
| DOI： |
| 分类号:P962 |
| 基金项目:河南省高等学校重点科研项目（基金号：20B170002）,河南省高校人文社会科学研究项目（基金号：2021-ZZJH-035） |
|
| Remote sensing monitoring of aquatic vegetation groups and algal blooms in typical inland lakes in the Yellow River Basin |
|
chenyonggui, zhuyuxiang
|
|
Henan Collage Of Surveying and Mapping
|
| Abstract: |
| 【Background】In recent decades, global lakes have suffered from ecological problems caused by water eutrophication and harmful algal blooms, and remote sensing technology is considered an effective means to monitor aquatic vegetation taxa and algal blooms. It is difficult to extract algal blooms in the complex inland lake aquatic environment of the Yellow River Basin. These areas usually grow a lot of aquatic vegetation, and a single spectral index cannot well distinguish the aquatic vegetation groups and yellow moss algal blooms in these areas, aquatic vegetation groups and yellow moss algae There is still uncertainty about the accuracy of remote sensing monitoring of aquatic vegetation groups and algal blooms.【Objective】Based on the spectral feature analysis of the measured spectral data, the classification indexes of emergent vegetation, submerged vegetation and floating algae suitable for the complex water environment of Ulansuhai Lake were selected and constructed, and a decisiontree classification model was constructed to analyze the vegetation taxa and algal blooms in the study area.【Methods】First, the MNDSI spectral index is proposed to exclude aquatic vegetation groups and types other than yellow moss algal blooms, then the NDWI spectral water index was used to extract open water areas or open water bodies, and then the improved enhanced vegetation index (AEVI) was used to extract emergent vegetation, and finally use the modified macroalgal index (MFAI) to distinguish submerged vegetation from yellow moss algal blooms. The applicability of the spectralindex was tested by the measured data and Landsat8.【Result】Spectral index combined with decision tree classification model is an effective method to distinguish aquatic vegetation from yellow moss algal blooms. The method was applied to the Landsat images during 1986—2018, and the spatial and temporal distribution characteristics and variation trends of the aquatic vegetation groups and yellow moss algal blooms in the Ulansuhai Lake Sea were studied.【Conclusion】The emergent vegetation taxa showed a slow growth trend during 1986—2018. From 1986 to 2018, the coverage area of submerged vegetation groups has the characteristics of periodic changes, specifically: a decreasing trend from 1986 to 2013, and a rapid growth trend after 2013. The yellow moss algal bloom showed no obvious trend, but it was explosive during 1986—2018. |
| Key words: aquatic vegetation group; yellow moss algal bloom; spectral index; decision tree; Ulansuhai Lake |
