Scholarlink

<p>Urban smart gardening for monitoring plant health contributes toward Sustainable Development Goal&#xa0;11 (Sustainable Cities and Communities). It demands real-world applicability in precision gardening, campus monitoring, and scalable smart agriculture. To use smart technologies for sustainable development, this study presents an AI-driven smart gardening system tailored for the Green Nirma Campus using UAV-based multispectral and RGB imaging. In contrast to earlier studies that concentrated on large-scale farming or disease detection, this work combines deep learning with multimodal data for intelligent urban gardening. The proposed approach used an 18-band spectrometer (410–940 nm) and RGB photos taken by a&#xa0;UAV to gather detailed physiological and image data from a&#xa0;variety of plant species. Spectral data has been processed using an Autoencoder-CNN architecture, achieving a&#xa0;classification accuracy of 98% between healthy and unhealthy plants. The RGB images integrated with spectral features through a&#xa0;dual-branch multimodal neural network are analysed using a&#xa0;frozen ResNet18 model, which achieved 99.2% accuracy. The novelty of this work lies in an integrated pipeline that involves a&#xa0;real-time data collection strategy via drones, fusion of spectral and image modalities, and its application in urban smart gardening, contributing toward scalable smart agriculture.</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

UAV-Based AI-Driven Multimodal Framework for Plant Health Monitoring for Green Campuses

  • 期刊论文

  • 发表日期：

    2026年7月2日

查看全文
  • Hridayesh Parmar,
  • Tvisha Patel,
  • Sumedha Arora,
  • Chandan Trivedi,
  • Jaiprakash Verma,
  • Swati Jain

摘要

Urban smart gardening for monitoring plant health contributes toward Sustainable Development Goal 11 (Sustainable Cities and Communities). It demands real-world applicability in precision gardening, campus monitoring, and scalable smart agriculture. To use smart technologies for sustainable development, this study presents an AI-driven smart gardening system tailored for the Green Nirma Campus using UAV-based multispectral and RGB imaging. In contrast to earlier studies that concentrated on large-scale farming or disease detection, this work combines deep learning with multimodal data for intelligent urban gardening. The proposed approach used an 18-band spectrometer (410–940 nm) and RGB photos taken by a UAV to gather detailed physiological and image data from a variety of plant species. Spectral data has been processed using an Autoencoder-CNN architecture, achieving a classification accuracy of 98% between healthy and unhealthy plants. The RGB images integrated with spectral features through a dual-branch multimodal neural network are analysed using a frozen ResNet18 model, which achieved 99.2% accuracy. The novelty of this work lies in an integrated pipeline that involves a real-time data collection strategy via drones, fusion of spectral and image modalities, and its application in urban smart gardening, contributing toward scalable smart agriculture.