Scholarlink

<p>Lack of freshwater inflow in lagoons and various anthropogenic activities has led to the shrinkage of a healthy mangrove community. This paper focuses on mapping degraded mangrove species using hyperspectral data in the Muthupet mangrove forest, Tamil Nadu, India. The Hyperion hyperspectral data, acquired on 29 September 2012, consists of 242 spectral bands. Due to the presence of water absorption bands, overlapping bands between VNIR and SWIR, and low illumination from the source to detectors, bands 1–7, 58–78, 120–132, 165–182, 185–187, 221–224, and 225–242 were effectively removed during preprocessing. As a result, 163 bands were selected for data processing. The significant steps involved in the current research include preprocessing L1Gst Hyperion data, analyzing data dimensionality for bands and pixels using Pixel Purity Index (PPI), extracting endmembers from pure pixels using the n-D Visualizer, and comparing the endmembers with a reference spectral library generated using a FieldSpec3 spectroradiometer. The optimized unique endmembers of each species were applied to the Hyperion reflectance data using the Spectral Angle Mapper (SAM). The classified image revealed that there is a significant amount of degradation in the Muthupet mangrove forest. Since the endmembers were provided in the form of pure pixels rather than object-based regions of interest in classifying the Hyperion data, reasonable accuracy was achieved. This extensive study helped identify that <i>Avicennia marina</i> is the dominant degraded species with an area of 340.7&#xa0;ha, followed by <i>Rhizophora mucronata</i> with an area of 272.4&#xa0;ha. The degradation of other species such as <i>Acanthus ilicifolius</i>, <i>Aegiceras corniculatum</i>, <i>Rhizophora apiculata</i>, and <i>Excoecaria agallocha</i> were found to have degraded occupancy areas of 167.1&#xa0;ha, 191.7&#xa0;ha, 72.5&#xa0;ha, and 54.5&#xa0;ha, respectively. The overall classification accuracy in identifying the mangrove species was found to be 85.6%, with a Kappa statistic of 0.79.</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

Enhancement of tropical degraded mangrove mapping using EO-1 satellite data: a case study of mangrove forest in Tamil Nadu, Southern India

  • 期刊论文

  • 发表日期：

    2025年5月29日

查看全文
  • Ayyakkannu Selvaraj,
  • Subbarayan Saravanan

摘要

Lack of freshwater inflow in lagoons and various anthropogenic activities has led to the shrinkage of a healthy mangrove community. This paper focuses on mapping degraded mangrove species using hyperspectral data in the Muthupet mangrove forest, Tamil Nadu, India. The Hyperion hyperspectral data, acquired on 29 September 2012, consists of 242 spectral bands. Due to the presence of water absorption bands, overlapping bands between VNIR and SWIR, and low illumination from the source to detectors, bands 1–7, 58–78, 120–132, 165–182, 185–187, 221–224, and 225–242 were effectively removed during preprocessing. As a result, 163 bands were selected for data processing. The significant steps involved in the current research include preprocessing L1Gst Hyperion data, analyzing data dimensionality for bands and pixels using Pixel Purity Index (PPI), extracting endmembers from pure pixels using the n-D Visualizer, and comparing the endmembers with a reference spectral library generated using a FieldSpec3 spectroradiometer. The optimized unique endmembers of each species were applied to the Hyperion reflectance data using the Spectral Angle Mapper (SAM). The classified image revealed that there is a significant amount of degradation in the Muthupet mangrove forest. Since the endmembers were provided in the form of pure pixels rather than object-based regions of interest in classifying the Hyperion data, reasonable accuracy was achieved. This extensive study helped identify that Avicennia marina is the dominant degraded species with an area of 340.7 ha, followed by Rhizophora mucronata with an area of 272.4 ha. The degradation of other species such as Acanthus ilicifolius, Aegiceras corniculatum, Rhizophora apiculata, and Excoecaria agallocha were found to have degraded occupancy areas of 167.1 ha, 191.7 ha, 72.5 ha, and 54.5 ha, respectively. The overall classification accuracy in identifying the mangrove species was found to be 85.6%, with a Kappa statistic of 0.79.