高级检索

基于多维超体积分割的露天煤矿生态系统变化监测

Monitoring of ecosystem changes in open-pit coal mines based on the segmentation of multi-dimensional hypervolume

  • 摘要: 露天煤矿开采和生态修复共同驱动矿山生态系统发生显著变化,新时期矿山土地复垦与生态修复监测评价将生态系统退化和恢复状况作为必选指标,但缺少有效监测方法。利用哨兵卫星数据反演矿山生态系统植被和土壤参数,并基于生态位原理开发露天煤矿生态系统多维超体积分割方法,将其应用到准能集团露天煤矿2018至2024年生态系统状态变化监测中。研究结果表明:① 综合利用哨兵卫星的极化、光谱和纹理特征变量,可以准确有效估测关键植被和土壤参数;② 随着生态系统参数数量的增加,不同生态系统状态空间的Jaccard相似系数降低,在超体积维数为3时,各生态系统状态空间的重叠度小,可分性高,植被覆盖度、群落结构指数、地面粗糙度、土壤湿度和土壤有机质是超体积分割的关键参数;③ 基于多维超体积分割,可以有效识别露天煤矿生态系统状态变化区域及幅度,生态系统状态跃迁幅度越大,监测精度越高,总体精度超过80%;④ 研究区2018至2024年剥离区和开采区生态系统状态以负向跃迁为主,排土区和复垦区生态系统状态以正向跃迁为主,表明准能集团露天煤矿有效实施了边采边复对策。总体来看,基于哨兵卫星数据和多维超体积分割方法可以实现对露天煤矿开采前、中、后生态系统退化和恢复状况的全过程监测,今后应加强多源多时相遥感数据的融合与协同分析,实现露天煤矿生态系统状态、结构和功能的一体化监测。

     

    Abstract: Open-pit coal mining and ecological restoration jointly drive significant changes in the mine ecosystem. In the new era, monitoring and evaluation of mine land reclamation and ecological restoration consider ecosystem degradation and recovery as essential indicators, but effective monitoring methods are lacking. Sentinel satellite data is used to invert vegetation and soil parameters. Based on the ecological niche theory, a segmentation method of multi-dimensional hypervolume of ecosystems in open-pit coal mine is developed and applied to monitor ecosystem changes in the open-pit coal mine of Zhunneng Group from 2018 to 2024. The research results show that: ① Comprehensive utilization of Sentinel satellite polarization, spectral, and texture feature variables can accurately and effectively estimate vegetation and soil parameters; ② As the number of ecosystem parameters increases, the Jaccard similarity coefficient of different ecosystem state spaces decrease. When the hypervolume dimension is 3, the overlap degree of each ecosystem state space is small and the separability is high. Vegetation coverage, community structure index, ground roughness, soil moisture, and soil organic matter are key ecosystem parameters for hypervolume segmentation; ③ Based on multi-dimensional hypervolume segmentation, the state change areas and amplitudes of the open-pit coal mine ecosystem can be effectively identified. The larger the amplitude of ecosystem state transitions, the higher the monitoring accuracy, with an overall accuracy exceeding 80%; ④ From 2018 to 2024, the ecosystem state in the stripping and mining areas of the study area mainly experienced negative transitions, while the ecosystem state in the dumping and reclamation areas mainly experienced positive transitions, indicating that open-pit coal mine of Zhunneng Group has effectively implemented the strategy of concurrent mining and reclamation. Overall, based on Sentinel satellite data and multi-dimensional hypervolume segmentation method, the whole process monitoring of ecosystem degradation and recovery status in open-pit coal mining before, during, and after mining can be achieved. In the future, the integration and collaborative analysis of multi-source and multi-temporal remote sensing data should be strengthened to achieve integrated monitoring of the state, structure, and function of the ecosystem in open-pit coal mine.

     

/

返回文章
返回