Scholarlink

<p>Soil moisture (SM) is a critical component in the fields of plant growth, soil health, climate and weather forecasting, erosion management and water resource management. The continuous monitoring of SM allows enhancement of crop yield and food security. Remote sensing is a state-of-the-art technology to measure the SM globally in an efficient way without any major financial requirements. The integration of two distinct types of electromagnetic radiation i.e., optical and microwave, ensures all-weather and accurate SM monitoring. However, the classification of fused datasets remains a challenging task due to the difficulty in preserving spectral and spatial information. In this article, different image fusion methods i.e., (a) nearest neighbour diffusion (NND), (b) gram Schmidt (GS), and (c) principal component analysis (PCA) have been implemented to develop SM maps using various classifiers i.e., (a) Multilayer Perceptron (MLP) based neural network (NN), (b) support vector machine (SVM), and (c) random forest (RF). To implement these algorithms, two global-level satellite datasets i.e., MODIS along with SCATSAT-1, were utilized over a part of North Indian. The performance analysis confirmed the efficacy of NND and ANN in the development of improved SM maps with an accuracy of 94.14%. This outcome of the study allows the precision SM monitoring which directly impacts the economy, ecosystem sustainability and food protection.</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

Improved Soil Moisture Classified Maps from Fusion of Optical and Microwave Satellite Datasets

  • 期刊论文

  • 发表日期：

    2026年7月30日

查看全文
  • Ravneet Kaur,
  • Raman Maini,
  • Reet Kamal Tiwari

摘要

Soil moisture (SM) is a critical component in the fields of plant growth, soil health, climate and weather forecasting, erosion management and water resource management. The continuous monitoring of SM allows enhancement of crop yield and food security. Remote sensing is a state-of-the-art technology to measure the SM globally in an efficient way without any major financial requirements. The integration of two distinct types of electromagnetic radiation i.e., optical and microwave, ensures all-weather and accurate SM monitoring. However, the classification of fused datasets remains a challenging task due to the difficulty in preserving spectral and spatial information. In this article, different image fusion methods i.e., (a) nearest neighbour diffusion (NND), (b) gram Schmidt (GS), and (c) principal component analysis (PCA) have been implemented to develop SM maps using various classifiers i.e., (a) Multilayer Perceptron (MLP) based neural network (NN), (b) support vector machine (SVM), and (c) random forest (RF). To implement these algorithms, two global-level satellite datasets i.e., MODIS along with SCATSAT-1, were utilized over a part of North Indian. The performance analysis confirmed the efficacy of NND and ANN in the development of improved SM maps with an accuracy of 94.14%. This outcome of the study allows the precision SM monitoring which directly impacts the economy, ecosystem sustainability and food protection.