RubberAIID: A lightweight intelligent platform for spatial localization and digital identification of individual rubber trees

To address the challenges of absent digital identities and difficult information traceability for individual trees in conventional rubber plantation management, this paper proposes the RubberAIID (Artificial Intelligence for Individual-tree Digital-identity) framework. At the core of this framework is CIS-YOLO, a lightweight deep learning model designed specifically for detecting rubber tree planting holes. By integrating specialized modules (SCDown, C3k2_DFF, and IEMA), the model achieves a [email protected] of 97.1% on the test set, outperforming mainstream models in performance while significantly reducing model complexity. We also introduce a novel concept, "Geo-native Digital Identity" (GDI), which establishes the precise geographic location of a tree as its permanent and immutable identifier; validation against RTK-GNSS measurements shows a mean localization offset of just 0.29 meters. Ultimately, by fusing UAV remote sensing, deep learning, and geospatial technologies, the RubberAIID platform creates persistent, tag-free digital identities, thereby achieving full-process automation of individual tree digital identity generation and management. This research not only provides rubber plantation managers with a low-barrier digital tool but also offers a promising solution for the sustainable management of other perennial crops.