Accurate detection of citrus fruit maturity is essential for robotic selective harvesting in orchards yet remains challenging due to complex environmental conditions. This paper presents ScaleEdgeFusion-Net (SEF-Net), a lightweight object detection framework built on YOLO11n for efficient citrus maturity recognition in real-world orchard environments, achieved through three key innovations: an Adaptive Multi-scale Edge Enhancement module integrated with the backbone to improve fruit discriminability; an Enhanced Multi-scale Feature Extraction module replacing standard spatial pyramid pooling to strengthen robustness against complex environmental conditions; and an Efficient Feature Fusion and Dynamic Sampling Neck redesigned to leverage dynamic upsampling and channel attention for high detection accuracy with minimal computational overhead. Experimental results demonstrate that SEF-Net achieves superior performance with 93.0% mAP@0.5 while maintaining only 2.0 million parameters and 5.2 G, resulting in a compact model size of 4.3 MB, and delivers the highest inference speed (132 FPS) among all compared models. Compared to state-of-the-art detectors—including general-purpose models (YOLOv5n, YOLOv8n, YOLO11n, etc) and citrus-specific models such as ORD-YOLO and LightSal-DETR—the proposed method achieves higher detection accuracy with significantly lower computational requirements. These results indicate that SEF-Net provides an effective balance between accuracy and efficiency, making it suitable for deployment on resource-constrained harvesting robots in precision agriculture applications.
Lanhui Fu, Zhijie Wu, Yingying Song et al.· Engineering Research Express· 0 citations
Production forecasting is essential for reservoir management, but crosswell multistep prediction remains challenging because production data involve nonlinear temporal dynamics, delayed intervariable dependencies, and strong heterogeneity among wells. To address these issues, we propose a temporal segment (TS)-graph convolutional network (GCN)-CausalMamba, a crosswell petroleum production forecasting framework that integrates Peter-Clark momentary conditional independence (PCMCI)-based causal discovery, GCN feature fusion, TS2Vec-assisted self-supervised representation learning, a Mamba selective state-space backbone, and a proportional-integral-derivative (PID)-inspired trajectory constraint loss. The framework combines dynamic production and operational variables with well-level geological and completion descriptors, enabling the model to capture both temporal evolution and crosswell structural heterogeneity. Causal graph modeling provides interpretable time-lagged structural priors. At the same time, the PID-inspired loss improves multistep trajectory stability by jointly constraining point-wise accuracy, cumulative trend consistency, and step-to-step smoothness. Experiments on real multiblock production data from the Daqing Oilfield, using well-based fivefold cross-validation, show that TS-GCN-CausalMamba consistently outperforms both general-purpose and petroleum-oriented baselines. Under the three-step forecasting setting, the proposed method improves coefficient of determination (R2) from 0.8838 to 0.9005 in Block A, from 0.9329 to 0.9342 in Block B, and from 0.9104 to 0.9147 in Block C. Meanwhile, it reduces normalized mean absolute error (NMAE) to 0.0583, 0.0332, and 0.0378, and normalized root mean square error (NRMSE) to 0.0942, 0.0577, and 0.0831 in Blocks A, B, and C, respectively. These results demonstrate its effectiveness, interpretability, and practical value for crosswell multistep petroleum production forecasting.
Jian Han, Zhihao Wang, Zhimin Cao et al.· SPE Journal· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
