Indoor object detection presents unique challenges such as occlusions, varying lighting conditions, and cluttered environments. While several object detection frameworks, including RetinaNet, Faster R-CNN, SSD, and EfficientDet, have been proposed, they often suffer from high computational cost, reduced inference speed, and limited accuracy in terms of mean Average Precision (mAP), particularly in real-time scenarios. In this study, lightweight YOLO variants, namely YOLOv7, YOLOv8s, YOLOv9s, and a fine-tuned YOLOv9s which considers the optimized training strategy based on albumentations. All the models are evaluated for indoor object detection using the RGB TUT Indoor dataset. The models are assessed using precision, recall, mAP@0.5, and mAP@0.5:0.95. The experimental results demonstrate that the fine-tuned YOLOv9s consistently outperforms the baseline YOLOv9s model across all evaluation metrics, confirming the effectiveness of proposed training optimizations. Specifically, the fine-tuned YOLOv9s achieves a precision of 97.9%, recall of 96.1%, mAP@0.5 of 99.1%, and mAP@0.5:0.95 of 88.7%. These improvements highlight the impact of systematic training refinement beyond standard model configuration. Among the evaluated models, YOLOv8s achieves the highest inference speed of 90 FPS in 11.1 ms, making it suitable for ultra-low-latency applications such as smart homes, assistive systems, and robotics. In contrast, the fine-tuned YOLOv9s provides a superior balance between accuracy and efficiency, making it more suitable for accuracy-sensitive indoor environments where detection reliability is critical. Overall, the study demonstrates that carefully optimized training strategies can significantly enhance the performance of YOLOv9s without architectural modifications, providing practical insights for real-time indoor object detection systems.
In this study, the effects of temperature and corrosion product porosity on the micro-galvanic corrosion behavior of the u03B2-Li phase in Mg-8Li alloy are systematically investigated using COMSOL Multiphysics numerical simulations. A two-dimensional micro-galvanic corrosion model incorporating mass transport, electrochemical reactions, and level set-based interface tracking is established to simulate the corrosion evolution over 72 h under varying temperature and porosity levels. The results indicate that temperature can significantly accelerate the corrosion process and the exchange current density increases exponentially. As the temperature increases from 35u00B0C to 55u00B0C, the electrolyte potential shifts negatively, and the maximum electrode thickness change rises from 8.2 to 34.0 mm, indicating that the localized corrosion approximately doubles when the temperature increases by 10u00B0C, and the peak local current density increases from 250 to 1100 A/m2. When the corrosion product porosity increases from 3% to 4.5%, the corrosion current density increases from 60.2 to 162 A/m2, and the thickness of the corrosion product layer increases from 0.5 to 2.0 mm. Simultaneously, under high porosity conditions, the current distribution becomes more uniform, and corrosion products form a more evenly distributed deposition layer, mitigating excessive localized corrosion. The combined effects of temperature and porosity significantly alter the interfacial ion transport and current density distribution, thereby governing the corrosion evolution path and interface morphology of the u03B2-Li phase.
Degradation data in practical reliability engineering are often scarce and heterogeneous, originating from multiple sources with varying degrees of uncertainty and conflict. Accordingly, this study proposes a hybrid framework that integrates Dempster–Shafer (D-S) evidence theory with the Wiener process for small-sample reliability assessment using multi-source heterogeneous data. First, a probabilistic non-uniform sampling method regularizes varied data sources and computes basic probability assignments (BPA). Second, a weight synthesis mechanism is constructed, where prior weights derived from prior knowledge are updated by evidence similarity quantified through the Expectation–Width (EW) distance, yielding posterior weights. Quantile sequences from each source are then fused via weighted aggregation to generate a time-series probability box. A Wiener process is subsequently employed to model the probability box (P-box) sequence, enabling interval reliability evaluation. Numerical simulations and a satellite gyroscope case study show that the method effectively fuses multi-source data, provides more comprehensive reliability intervals than single-source approaches, and significantly enhances evaluation robustness under small samples. The framework offers a systematic solution for multi-source data fusion and establishes a novel reliability assessment pathway under uncertainty, with broad applicability to aerospace and precision instrumentation systems.
