高光谱成像
残余物
遥感
预处理器
环境科学
计算机科学
精准农业
像素
人工智能
校准
回归
营养物
数据预处理
温室
原始数据
模式识别(心理学)
深度学习
试验数据
数据挖掘
农业工程
作者
Abigail Cohen,Harsh Muriki,Yuming Sun,Zhihao Qin,Lu Gan,Yongsheng Chen
标识
DOI:10.1016/j.atech.2025.101501
摘要
Hyperspectral imaging (HSI) shows promise for non-destructive measurement of tissue nutrient concentrations in plants. However, many HSI analysis approaches rely on extensive preprocessing, traditional statistical methods, and pixel averaging that require subject matter expertise and sacrifice spatial information. To further prospects for adoption of HSI in precision applications, we explored end-to-end deep learning (DL) to process raw hyperspectral data (400-1100nm). This study presents a novel application of vision transformers (ViT) for a regression task in precision nutrient management, comparing it to popular residual network (ResNet) architectures (ResNet-34, ResNet-50, and ResNet-101) and providing a proof-of-concept for nutrient status estimation from raw proximal hyperspectral imagery. To explore the applicability of these approaches, we conducted a pilot-scale longitudinal hydroponic lettuce growth study with destructive sampling, imaging plants grown with different applied nutrient concentrations under pilot-scale greenhouse conditions. Results demonstrate that even with the same data splits, performance is highly variable within and between models. ViT and ResNet-34 performance in N estimation was particularly noteworthy, achieving test R2 of 0.68 and 0.74 R2 without radiometric calibration or extensive preprocessing. These results demonstrate the potential for developing DL approaches that support end-to-end nutrient control automation in precision agriculture.
科研通智能强力驱动
Strongly Powered by AbleSci AI