Scholarlink

Super-resolving the satellite images would be crucial in order to improve precision in tree classification which is a very important aspect as far as environmental monitoring is concerned. Super-resolution addresses the probe of increasing image resolution by recovering fine details from low-resolution inputs. Three deep learning super-resolution methods, namely SRCNN, ESPCN, and ESRGAN were applied to a set of satellite imagery that were divided into two classes: images with trees and those without them. Our research shows that ESPCN outperforms SRCNN and ESRGAN based on Peak Signal to Noise Ratio and Structural Similarity Index Measure. For super-resolved images, a Convolutional Neural Network is used to classify them into TREES or NO TREES classes. The results indicate that using ESPCN enhanced images significantly improves classification performance. Combining super-resolution with classification approaches provides a reliable framework for accurate tree detection in satellite imagery. This framework is crucial for precise assessments of tree canopy attributes, which are vital for applications like biodiversity monitoring, forest management, and environmental sustainability.

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

Super-Resolution Enhanced Tree Classification in Satellite Images Using Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Nisha Shamsudin,
  • V R Bindu

摘要

Super-resolving the satellite images would be crucial in order to improve precision in tree classification which is a very important aspect as far as environmental monitoring is concerned. Super-resolution addresses the probe of increasing image resolution by recovering fine details from low-resolution inputs. Three deep learning super-resolution methods, namely SRCNN, ESPCN, and ESRGAN were applied to a set of satellite imagery that were divided into two classes: images with trees and those without them. Our research shows that ESPCN outperforms SRCNN and ESRGAN based on Peak Signal to Noise Ratio and Structural Similarity Index Measure. For super-resolved images, a Convolutional Neural Network is used to classify them into TREES or NO TREES classes. The results indicate that using ESPCN enhanced images significantly improves classification performance. Combining super-resolution with classification approaches provides a reliable framework for accurate tree detection in satellite imagery. This framework is crucial for precise assessments of tree canopy attributes, which are vital for applications like biodiversity monitoring, forest management, and environmental sustainability.