Advancements in Weed Mapping: A Systematic Review

Weed mapping plays a critical role in precision management by providing accurate and timely data on weed distribution, enabling targeted control and reduced herbicide use. This minimizes environmental impacts, supports sustainable land management, and improves outcomes across agricultural and natural environments. Recent advances in weed mapping leverage ground-vehicle Red Green Blue (RGB) cameras, satellite and drone-based remote sensing combined with sensors such as spectral, Near Infra-Red (NIR), and thermal cameras. The resulting data are processed using advanced techniques including big data analytics and machine learning, significantly improving the spatial and temporal resolution of weed maps and enabling site-specific management decisions. Despite a growing body of research in this domain, there is a lack of comprehensive literature reviews specifically focused on weed mapping. In particular, the absence of a structured analysis spanning the entire mapping pipeline, from data acquisition to processing techniques and mapping tools, limits progress in the field. This review addresses these gaps by systematically examining state-of-the-art methods in data acquisition (sensor and platform technologies), data processing (including annotation and modelling), and mapping techniques (such as spatiotemporal analysis and decision support tools). Following PRISMA guidelines, we critically evaluate and synthesize key findings from the literature to provide a holistic understanding of the weed mapping landscape. This review serves as a foundational reference to guide future research and support the development of efficient, scalable, and sustainable weed management systems.

💡 Research Summary

This paper presents the first comprehensive systematic review that covers the entire weed‑mapping pipeline, from data acquisition through processing to mapping and decision‑support tools. Following PRISMA guidelines, the authors screened two major databases (Elsevier and IEEE Xplore) for publications after the year 2000, initially retrieving 238 records. After title/abstract screening, eligibility checks, and full‑text assessment, 151 studies (135 journal articles, 8 reports, and 8 conference papers) were retained for analysis, spanning 2003‑2025.

The review is organized into three major sections. The first examines data‑acquisition technologies. Ground‑based vehicles equipped with RGB cameras, UAVs (drones) carrying multispectral, hyperspectral, near‑infrared (NIR) and thermal sensors, and satellite platforms are compared. While satellites provide broad coverage, UAVs offer high spatial resolution and rapid revisit capability; ground vehicles deliver the highest image quality for plot‑scale studies. The authors note a trend toward hybrid sensor suites that combine low‑cost RGB with selective hyperspectral bands to balance cost and classification accuracy.

The second section focuses on data processing. Because modern weed‑mapping generates terabytes of multi‑temporal imagery, big‑data pipelines (cloud storage, distributed file systems, and parallel processing frameworks) are essential. Annotation strategies have shifted from fully manual labeling to semi‑automated tools and clustering‑based label propagation, reducing labor while maintaining quality. Machine‑learning models dominate the literature: convolutional neural networks (CNNs) for object detection (YOLO, Faster R‑CNN), encoder‑decoder architectures (U‑Net, DeepLab) for semantic segmentation, and newer transformer‑based networks for spectral‑spatial fusion. Reported accuracies exceed 95 % for species‑level discrimination. Edge‑computing solutions—NVIDIA Jetson, Google Coral, FPGA accelerators—enable on‑site inference with latencies under 100 ms, supporting real‑time weed detection on tractors and drones.

The third section addresses mapping and decision‑support. Spatiotemporal analyses use time‑series of images to model weed population dynamics, employing clustering (DBSCAN, K‑means) and forecasting models (LSTM, Prophet) to predict future infestations. GIS platforms (QGIS, ArcGIS) and web‑based dashboards visualize risk maps, allowing users to toggle layers, define treatment zones, and generate herbicide‑application schedules automatically. Integrated decision‑support systems (DSS) combine rule‑based recommendations with reinforcement‑learning agents to suggest optimal timing and dosage, while also providing cost‑benefit and environmental impact assessments for policymakers.

Meta‑analysis of the literature reveals a steady increase in publications, with a sharp rise in deep‑learning and hyperspectral studies over the past five years. However, the authors identify persistent challenges: high sensor costs, lack of standardized data formats and metadata, limited model generalization across crops and regions, and insufficient real‑time processing capabilities for field deployment.

To guide future work, the paper proposes a roadmap: (1) development of affordable high‑performance sensors and multimodal fusion hardware; (2) creation of open, standardized data repositories and metadata schemas (aligned with ISO/FAO standards); (3) design of lightweight, edge‑optimized AI models through pruning, quantization, and knowledge distillation; (4) integration of multimodal data (RGB, NIR, thermal, LiDAR) into unified deep‑learning frameworks; and (5) coupling of mapping outputs with policy‑aware DSS for sustainable weed‑management strategies.

In conclusion, the review demonstrates that weed‑mapping technologies have matured across sensing, analytics, and decision support, but achieving scalable, cost‑effective, and policy‑aligned solutions will require coordinated advances in hardware, algorithms, data standards, and stakeholder collaboration.