Scholarlink

<p>The use of space based hyperspectral data for application in crop classification or any other varietal specific classification is limited due to the low spatial resolution. With the aim to overcome such drawbacks, various multi-resolution data fusion algorithms have been developed over the years to integrate the high spectral low spatial resolution hyperspectral data with low spectral high spatial resolution multispectral data. This study presents a comparative analysis to assess the efficacy of EnMAP (Environmental Mapping and Analysis Program) hyperspectral fusion with Sentinel-2 and LISS-IV (Linear Imaging Self-Scanning Sensor-IV) for improved crop classification. Gram Schmidt spectral sharpening, a state-of-the-art fusion algorithm is utilized. The spatial and spectral qualities of the fused data are examined using ERGAS (Erreur Relative Globale Adimensionnelle de Synthèse), RMSE (Root Mean Square Error), and CC (Correlation Coefficient). All the reference and fused datasets are classified using endmembers obtained from the hyperspectral data using classification algorithms like Spectral Angle Mapper (SAM), Spectral Information Divergence (SID), Maximum Likelihood and Support Vector Machine (SVM). The accuracy assessment shows improved results in both the fused datasets in comparison to the original EnMAP hyperspectral data throughout the classification algorithms particularly the EnMAP data fused with Sentinel-2 due to the preservation of a larger number of bands.</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

Analysis of EnMAP Hyperspectral Data Fusion with Sentinel-2 and LISS-IV Multispectral Datasets for Crop Classification

  • 期刊论文

  • 发表日期：

    2026年4月18日

查看全文
  • Ovungrhoni V. Murry,
  • Jonali Goswami,
  • K. K. Sarma,
  • Prasanna Boruah

摘要

The use of space based hyperspectral data for application in crop classification or any other varietal specific classification is limited due to the low spatial resolution. With the aim to overcome such drawbacks, various multi-resolution data fusion algorithms have been developed over the years to integrate the high spectral low spatial resolution hyperspectral data with low spectral high spatial resolution multispectral data. This study presents a comparative analysis to assess the efficacy of EnMAP (Environmental Mapping and Analysis Program) hyperspectral fusion with Sentinel-2 and LISS-IV (Linear Imaging Self-Scanning Sensor-IV) for improved crop classification. Gram Schmidt spectral sharpening, a state-of-the-art fusion algorithm is utilized. The spatial and spectral qualities of the fused data are examined using ERGAS (Erreur Relative Globale Adimensionnelle de Synthèse), RMSE (Root Mean Square Error), and CC (Correlation Coefficient). All the reference and fused datasets are classified using endmembers obtained from the hyperspectral data using classification algorithms like Spectral Angle Mapper (SAM), Spectral Information Divergence (SID), Maximum Likelihood and Support Vector Machine (SVM). The accuracy assessment shows improved results in both the fused datasets in comparison to the original EnMAP hyperspectral data throughout the classification algorithms particularly the EnMAP data fused with Sentinel-2 due to the preservation of a larger number of bands.