ORIGINAL RESEARCH article
Front. Plant Sci.
Sec. Sustainable and Intelligent Phytoprotection
Volume 16 - 2025 | doi: 10.3389/fpls.2025.1630758
This article is part of the Research TopicAccurate Measurement and Dynamic Monitoring of Forest ParametersView all 9 articles
Reconstructed Hyperspectral Imaging for In-Situ Nutrient Prediction in Pine Needles
Provisionally accepted
Yuanhang Li
Jun DuChuangjie ZengYongshan WuJunxian Chen
Teng Long
Yongbing LongYubin Lan
Xiaoliang CheTianyi Liu*
Zhao Jing*- South China Agricultural University, Guangzhou, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
hyperspectral imaging (HSI) has become an advanced, non-destructive technology that enables rapid and accurate analysis of plant nutrients, thereby significantly improving forestry productivity and quality. However, the high cost and operational complexity of hyperspectral imaging systems have limited their practical application for in situ field detection. To address these challenges and maximize the potential of hyperspectral technology, this study proposes an innovative deep-learning-based method for in situ needle nutrient prediction by reconstructing hyperspectral images from RGB inputs. Compared to traditional methods that directly use raw hyperspectral imaging for nutrient content detection, this approach streamlines the workflow and extends the spectral range into the near-infrared band through improvements to the models. Specifically, we reconstructed hyperspectral images with a spectral range of 400-1000nm (with a resolution of 3.4nm) and a spatial resolution of 768×768 from 3channel RGB data to meet the requirements for accurate nutrient content prediction.The spectral information extracted from these hyperspectral images was combined with competitive adaptive reweighted sampling (CARS) and partial least squares regression (PLSR) to successfully predict needle nitrogen, phosphorus, and potassium content.The coefficients of determination (R²) for nitrogen, phosphorus, and potassium were 0.8523, 0.7022, and 0.8087, respectively, closely matching the results predicted from the original hyperspectral data.The proposed method significantly reduces the time and economic cost of hyperspectral technology and provides strong technical support for forestry production. It offers an innovative approach for in situ field detection and data analysis, ultimately contributing to the optimal utilization of resources and sustainable forestry development.
Keywords: hyperspectral image reconstruction, Deep learning models, Machine learning regression, In situ prediction of pine needle nutrients, precision forestry
Received: 18 May 2025; Accepted: 17 Jul 2025.
Copyright: © 2025 Li, Du, Zeng, Wu, Chen, Long, Long, Lan, Che, Liu and Jing. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Tianyi Liu, South China Agricultural University, Guangzhou, China
Zhao Jing, South China Agricultural University, Guangzhou, China
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.