Technology Category
Intelligent Robots: Multimodal Perception & Sensor FusionComputer Vision: Remote Sensing / Geospatial AIMachine Learning: Multi-instance/Multi-view Learning
Drone-Based Multispectral Imaging and Deep Learning for Timely Detection of Branched Broomrape in Tomato Farms
🤖 AI Summary
To address the challenges of early detection and costly, environmentally harmful chemical control of the parasitic weed *Phelipanche ramosa* in California’s tomato industry, this study proposes an intelligent identification framework integrating unmanned aerial vehicle (UAV)-based multispectral remote sensing with deep learning. Methodologically, we innovatively combine Long Short-Term Memory (LSTM) networks with time-series multispectral data and introduce Synthetic Minority Oversampling Technique (SMOTE) for class imbalance mitigation—enabling robust detection of subtle spectral signatures during early parasitism. Furthermore, we incorporate growing degree days (GDD)-driven phenological staging to enhance temporal modeling accuracy. At the critical parasitism window (897 GDD), the model achieves 79.09% accuracy and 70.36% recall. When trained on full-season data augmented with SMOTE, performance improves to 88.37% accuracy and 95.37% recall—substantially outperforming conventional approaches. This work delivers an efficient, sustainable, and scalable precision agriculture solution for early-warning monitoring of parasitic weeds.
Technology Category
Application Category
📝 Abstract
This study addresses the escalating threat of branched broomrape (Phelipanche ramosa) to California's tomato industry, which supplies over 90 percent of U.S. processing tomatoes. The parasite's largely underground life cycle makes early detection difficult, while conventional chemical controls are costly, environmentally harmful, and often ineffective. To address this, we combined drone-based multispectral imagery with Long Short-Term Memory (LSTM) deep learning networks, using the Synthetic Minority Over-sampling Technique (SMOTE) to handle class imbalance. Research was conducted on a known broomrape-infested tomato farm in Woodland, Yolo County, CA, across five key growth stages determined by growing degree days (GDD). Multispectral images were processed to isolate tomato canopy reflectance. At 897 GDD, broomrape could be detected with 79.09 percent overall accuracy and 70.36 percent recall without integrating later stages. Incorporating sequential growth stages with LSTM improved detection substantially. The best-performing scenario, which integrated all growth stages with SMOTE augmentation, achieved 88.37 percent overall accuracy and 95.37 percent recall. These results demonstrate the strong potential of temporal multispectral analysis and LSTM networks for early broomrape detection. While further real-world data collection is needed for practical deployment, this study shows that UAV-based multispectral sensing coupled with deep learning could provide a powerful precision agriculture tool to reduce losses and improve sustainability in tomato production.
Problem
Research questions and friction points this paper is trying to address.
Detecting parasitic weed early in tomato farms
Using drone multispectral imagery for crop monitoring
Applying deep learning to improve detection accuracy
Innovation
Methods, ideas, or system contributions that make the work stand out.
Drone-based multispectral imaging for crop monitoring
LSTM deep learning networks for temporal analysis
SMOTE technique to handle class imbalance
🔎 Similar Papers
No similar papers found.
