Scholarlink

<p>Mangrove ecosystems are critically important for coastal biodiversity, carbon storage, and the protection of the shoreline. Despite this, their mapping and monitoring have potential challenges due to dense canopy structures and limited field accessibility. The past studies rely largely on field-based training or ground spectral data, limiting their applicability in such environments. In contrast, the present study demonstrates the potential of AVIRIS-NG hyperspectral imagery to extract spectral signatures directly from the image itself, thereby reducing dependency on extensive field campaigns. Airborne hyperspectral sensors, such as AVIRIS-NG, having high spectral resolution, enable capturing detailed spectral signatures comparable to spectroradiometers, providing an effective and reliable alternative for spectral signature identification. The study was conducted using AVIRIS-NG hyperspectral data combined with statistical separability measures to identify mangrove spectral signatures in the Marine National Park (MNP), Jamnagar, Gulf of Kutch, Gujarat. Spectral separability analysis was achieved using one-way ANOVA, stepwise linear discriminant analysis (SWLDA), and Jeffrey–Matusita (JM) distance to differentiate between mangrove spectral signatures. Reference spectra of <i>Avicennia marina (AM)</i>, a dominant mangrove species, were employed to match and validate spectra identified by spectral separability analysis. The proposed workflow successfully extracted spectral signatures for three mangrove species classes, with JM distance values exceeding 1.8, indicating strong separability between extracted mangrove groups. ANOVA results identified significant spectral separability (<i>p</i> &lt; 0.05) between species groups, and the outcome of SWLDA showed a potential set of wavelengths for discrimination within the range of visible, shortwave, and infrared regions. Furthermore, one of the extracted spectral signatures is accurately matched with reference spectra of <i>Avicennia marina (AM).</i> Overall, the study demonstrates the potential of hyperspectral remote sensing to effectively capture species-level spectral information in inaccessible coastal areas. The methodology demonstrates a potentially scalable and reliable framework applicable to other mangrove regions and a wide variety of species, supporting conservation planning, ecosystem mapping and monitoring, and future models for remote sensing-based habitat assessment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mangrove spectral signature identification and separability analysis using a hyperspectral remote sensing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月7日

查看全文
  • Khushbu Maurya,
  • Seema Mahajan

摘要

Mangrove ecosystems are critically important for coastal biodiversity, carbon storage, and the protection of the shoreline. Despite this, their mapping and monitoring have potential challenges due to dense canopy structures and limited field accessibility. The past studies rely largely on field-based training or ground spectral data, limiting their applicability in such environments. In contrast, the present study demonstrates the potential of AVIRIS-NG hyperspectral imagery to extract spectral signatures directly from the image itself, thereby reducing dependency on extensive field campaigns. Airborne hyperspectral sensors, such as AVIRIS-NG, having high spectral resolution, enable capturing detailed spectral signatures comparable to spectroradiometers, providing an effective and reliable alternative for spectral signature identification. The study was conducted using AVIRIS-NG hyperspectral data combined with statistical separability measures to identify mangrove spectral signatures in the Marine National Park (MNP), Jamnagar, Gulf of Kutch, Gujarat. Spectral separability analysis was achieved using one-way ANOVA, stepwise linear discriminant analysis (SWLDA), and Jeffrey–Matusita (JM) distance to differentiate between mangrove spectral signatures. Reference spectra of Avicennia marina (AM), a dominant mangrove species, were employed to match and validate spectra identified by spectral separability analysis. The proposed workflow successfully extracted spectral signatures for three mangrove species classes, with JM distance values exceeding 1.8, indicating strong separability between extracted mangrove groups. ANOVA results identified significant spectral separability (p < 0.05) between species groups, and the outcome of SWLDA showed a potential set of wavelengths for discrimination within the range of visible, shortwave, and infrared regions. Furthermore, one of the extracted spectral signatures is accurately matched with reference spectra of Avicennia marina (AM). Overall, the study demonstrates the potential of hyperspectral remote sensing to effectively capture species-level spectral information in inaccessible coastal areas. The methodology demonstrates a potentially scalable and reliable framework applicable to other mangrove regions and a wide variety of species, supporting conservation planning, ecosystem mapping and monitoring, and future models for remote sensing-based habitat assessment.