Volume 16 Issue 6
Published on June 2025This study takes City S, a mega-city in North China, as the research object. Based on a thorough review of relevant literature and theoretical foundations, it employs the Principal Component Analysis (PCA) method to construct a multidimensional indicator system encompassing population, economy, society, and ecology. Using statistical data from 2013 to 2020, the study quantitatively analyzes the degree of influence exerted by various driving factors on urban landscape changes. The results show that natural factors, population factors, economic development factors, and social policy factors are the primary drivers of landscape change. Social development and ecological constraints also play a role in the adjustment of urban spatial structure to a certain extent. The study further reveals the comprehensive driving mechanism underlying urban landscape evolution and provides a theoretical basis and methodological support for urban land use optimization and landscape planning. PCA demonstrates strong applicability in identifying multifactor coupling mechanisms and can serve as a scientific reference for the formulation of urban sustainable development strategies.
View pdf
Environmentally friendly lead-free metal halide scintillators have attracted significant research interest in recent years in the field of radiation detection due to their low toxicity and outstanding radioluminescent properties. However, enhancing the luminescence performance of lead-free metal halide scintillators and fabricating thin films that integrate high light output, high spatial resolution, and excellent compatibility with photodetectors remain major challenges. To address these issues, this study employs Cs₃Cu₂I₅ as the scintillating material and introduces Terbium (Tb) doping to tune its emission to match current photodetectors. A high-refractive-index flexible UV-curable adhesive, NOA170F, is used as the matrix to synthesize a Cs₃Cu₂I₅: Tb scintillating thin film featuring a large Stokes shift, high luminescence intensity, low cost, environmental friendliness, and good stability. The luminescence mechanism and X-ray scintillation properties of the film are also investigated.
View pdf
In recent years, Vision-Language Models (VLMs) have emerged as a significant breakthrough in multimodal learning, demonstrating remarkable progress in tasks such as image-text alignment, image generation, and semantic reasoning. This paper systematically reviews current VLM pretraining methodologies, including contrastive learning and generative paradigms, while providing an in-depth analysis of efficient transfer learning strategies such as prompt tuning, LoRA, and adapter modules. Through representative models like CLIP, BLIP, and GIT, we examine their practical applications in visual grounding, image-text retrieval, visual question answering, affective computing, and embodied AI. Furthermore, we identify persistent challenges in fine-grained semantic modeling, cross-modal reasoning, and cross-lingual transfer. Finally, we envision future trends in unified architectures, multimodal reinforcement learning, and domain adaptation, aiming to provide systematic reference and technical insights for subsequent research.
View pdf
This paper proposes a waste heat recovery system that integrates a small-scale Organic Rankine Cycle (ORC) power generation unit with an air compressor using lubricating oil. An innovative design directly connects the capacitor to the scroll expander, replacing the traditional generator. Experimental results show that when the lubricating oil is at 105°C and the cooling water at 19°C, the system achieves a maximum power output of 1.754 kW and a thermal-electric efficiency of 3.91%. Once the lubricating oil flow rate exceeds a certain threshold, its impact on efficiency improvement becomes limited, whereas lowering the cooling water temperature significantly enhances efficiency (3.91% at 19°C, dropping to 2.7% at 30°C). Increasing the capacitor size allows a maximum power output of 1.667 kW and a thermal efficiency of 3.66%. This system offers an efficient solution for recovering waste heat from industrial air compressors.
View pdf
In the era of digital communication, the exponential growth of user-generated content across social media and online platforms has intensified the demand for effective emotion analysis tools. Traditional text-based sentiment analysis methods, however, often fall short in accurately capturing the nuances of human emotions due to their reliance on a single modality. Motivated by the need for more comprehensive and context-aware emotion recognition, this study systematically reviews the literature on both unimodal and multimodal aspect-level sentiment analysis. By comparing different approaches within the multimodal domain, we identify existing challenges and emerging trends in this research area. Our findings highlight the potential of integrating multiple modalities—such as text, images, and audio—to enhance the precision of sentiment detection and suggest future directions for advancing multimodal sentiment analysis.
View pdf
Since their emergence, Two-Dimensional (2D) materials have garnered significant attention due to their unique crystal structures and electronic properties, which offer distinct advantages for various applications. As a result, the study of 2D materials has become a crucial area within materials science. This paper introduces four representative 2D materials: graphene, hexagonal boron nitride, two-dimensional transition metal dichalcogenides, and phosphorene. It also briefly discusses their applications in electronic and optoelectronic devices, batteries, supercapacitors, and photocatalytic reactions, analyzing the advantages they hold over traditional materials.
View pdf
Accurate viewport prediction is crucial for enhancing user experience in 360-degree video streaming. However, due to significant behavioral differences among user groups, traditional single LSTM models tend to fall into local optima and fail to achieve precise predictions. To address this, this paper proposes a hybrid prediction model based on user clustering. First, a Density-Based Clustering Algorithm (DBSCAN) is used to group users with similar behavioral patterns. Then, a hybrid prediction model combining Generative Adversarial Networks (GANs) and Long Short-Term Memory networks (LSTMs) is designed to effectively mitigate data imbalance and overfitting through collaborative training. Experiments conducted on three real-world datasets from YouTube demonstrate that this approach significantly outperforms existing methods based on user trajectories or video saliency in terms of prediction accuracy and stability.
View pdf
Metamaterials, engineered with artificially designed microstructures, transcend the performance limits of natural materials and exhibit disruptive potential in mechanics, electromagnetism, thermodynamics, and aerospace engineering. This paper first introduces the key characteristics of metamaterials and their applications across various domains. It then discusses the structural failure of metamaterials under thermal environments, including the physical mechanisms of failure and typical failure modes, along with recent advances in the development of novel heat-resistant materials. Finally, it examines the stability and failure analysis of metamaterials in extreme environments, addressing both single-factor stability and failure as well as response and solutions under multi-field coupling conditions.
View pdf
This paper presents a comprehensive review of the key technologies and recent developments in the field of Civil Unmanned Aerial Vehicle (CUAV), systematically analyzing the latest developments in flight platform technology, propulsion and energy technology, and navigation and control systems. Special attention is given to the performance characteristics and application scenarios of fixed-wing, multi-rotor, and composite flight platforms, as well as to the advantages and limitations of lithium-ion and hydrogen fuel cells in terms of endurance, and the role of multi-sensor data fusion algorithms in navigation and control. By examining practical cases from typical fields such as aerial photography, logistics, and agriculture, this paper summarizes current technological achievements and application trends. Furthermore, it analyzes challenges in areas such as regulations and airspace management, safety, and technical limitations, while proposing future directions including technology integration, quantum communication, bio-inspired drones, and privacy and data protection. The research indicates that with the deeper application of emerging technologies like artificial intelligence and quantum communication, CUAV will achieve new breakthroughs in intelligence, autonomy, and safety, providing strong technological support for low-altitude economy and smart society.
View pdf
With the rapid development of industries such as artificial intelligence and big data, the demand for liquid cooling in data centers is continuously increasing. Among various technologies, cold plate liquid cooling has become one of the most widely applied methods. The performance and quality of components in the cooling system directly affect its operational and maintenance costs. This paper focuses on fluid connectors in liquid cooling systems, exploring key technical factors influencing their performance. It introduces process methods covering the entire production cycle from part machining to assembly and testing, and elaborates on critical control points at each manufacturing stage. The aim is to improve the technological level of fluid connectors. Finally, the paper briefly introduces new materials and technologies applied to fluid connectors and presents prospects for the future development of the liquid cooling industry.
View pdf