SWNet: A Cross-Spectral Network for Camouflaged Weed Detection
Henry O. Velesaca, Luigi Miranda, Angel D. Sappa
TLDR
SWNet is a cross-spectral network using visible and NIR data with a Vision Transformer and edge refinement for accurate camouflaged weed detection.
Key contributions
- Introduces SWNet, a bimodal cross-spectral network for camouflaged weed detection in dense crops.
- Uses a Pyramid Vision Transformer v2 backbone and Bimodal Gated Fusion for V/NIR data integration.
- Leverages NIR chlorophyll reflectance differences to distinguish camouflaged weeds from crops.
- Employs an Edge-Aware Refinement module for sharper boundaries and reduced ambiguity.
Why it matters
This paper addresses the critical challenge of camouflaged weed detection, which is vital for precision agriculture. By integrating cross-spectral data, SWNet significantly improves accuracy where visible-light systems fail. This advancement can lead to more efficient and sustainable weed management practices.
Original Abstract
This paper presents SWNet, a bimodal end-to-end cross-spectral network specifically engineered for the detection of camouflaged weeds in dense agricultural environments. Plant camouflage, characterized by homochromatic blending where invasive species mimic the phenotypic traits of primary crops, poses a significant challenge for traditional computer vision systems. To overcome these limitations, SWNet utilizes a Pyramid Vision Transformer v2 backbone to capture long-range dependencies and a Bimodal Gated Fusion Module to dynamically integrate Visible and Near-Infrared information. By leveraging the physiological differences in chlorophyll reflectance captured in the NIR spectrum, the proposed architecture effectively discriminates targets that are otherwise indistinguishable in the visible range. Furthermore, an Edge-Aware Refinement module is employed to produce sharper object boundaries and reduce structural ambiguity. Experimental results on the Weeds-Banana dataset indicate that SWNet outperforms ten state-of-the-art methods. The study demonstrates that the integration of cross-spectral data and boundary-guided refinement is essential for high segmentation accuracy in complex crop canopies. The code is available on GitHub: https://cod-espol.github.io/SWNet/
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.