Comparative Evaluation of Convolutional and Transformer-Based Detectors for Automated Weed Detection in Precision Agriculture

arXiv cs.CV / 5/5/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper compares CNN-based and transformer-based object detection architectures specifically for early weed detection in precision agriculture under realistic conditions.
It evaluates representative models from each category, including a recent YOLO family variant (YOLOv26-nano) alongside RT-DETR and RF-DETR transformer-based approaches.
Experiments on the GROUNDBASED_WEED dataset measure both detection quality (e.g., precision, recall, average precision) and computational efficiency (inference speed).
The findings show a clear trade-off: CNN detectors typically deliver strong performance with lower computational cost, whereas transformer detectors better capture global context but require more resources.
The study’s results are intended to provide practical selection criteria for choosing detection models for real-world precision agriculture systems.

Abstract

This paper presents a comparative evaluation of convolutional and transformer-based object detection architectures for early weed detection in realistic scenarios. Representative models from each paradigm are considered, including YOLOv26-nano, a recent variant of the YOLO family, and transformer-based approaches such as RTDETR and RF-DETR. Experiments were conducted on the GROUNDBASED_ WEED dataset, allowing performance to be evaluated in terms of detection accuracy and computational efficiency using metrics such as precision, recall, average precision, and inference speed. The results highlight a clear trade-off between efficiency and contextual modeling: CNN-based detectors achieve high performance at a lower computational cost, while transformer-based approaches offer better global context capture at the expense of higher resource demands. These results provide practical criteria for model selection in precision agriculture applications.