Scholarlink

This study explores the Novel Shortcut Tanish Residual Neural Network (STRNN) to thoroughly evaluate the complex spectrum properties of remote sensing data. It explores the basic ideas, principles of construction, and possible uses of STRNN in remote sensing, with a focus on how well it can categorize mangrove species. Implemented STRNN outperforms CNN and ResNet, achieving F1 score 0.94, accuracy 95.72%, and precision 0.93 on the Indian Pines dataset. Utilizing the achievements of the Indian Pines dataset, the strategy is applied to the EO1 Hyperion dataset for assessing mangrove adaptability. The methodology improves mangrove image quality by applying Speckle Noise Reduction, Balanced Histogram Equalization, and Levy Reptile Search Optimization during training. Mangrove classification utilizes Polynomial Kernelized Watershed Segmentation, with feature selection carried out by STRNN.

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

Advanced Hyperspectral Image Classification Using Shortcut Tanish Residual Neural Network and Levy Flight Search Optimization in Mangrove Classification

  • 章节

  • 发表日期：

    2025年

查看全文
  • Anindita Das Bhattacharjee,
  • Somdatta Charavortty

摘要

This study explores the Novel Shortcut Tanish Residual Neural Network (STRNN) to thoroughly evaluate the complex spectrum properties of remote sensing data. It explores the basic ideas, principles of construction, and possible uses of STRNN in remote sensing, with a focus on how well it can categorize mangrove species. Implemented STRNN outperforms CNN and ResNet, achieving F1 score 0.94, accuracy 95.72%, and precision 0.93 on the Indian Pines dataset. Utilizing the achievements of the Indian Pines dataset, the strategy is applied to the EO1 Hyperion dataset for assessing mangrove adaptability. The methodology improves mangrove image quality by applying Speckle Noise Reduction, Balanced Histogram Equalization, and Levy Reptile Search Optimization during training. Mangrove classification utilizes Polynomial Kernelized Watershed Segmentation, with feature selection carried out by STRNN.