Search Results (147)

Despite recent advances in deep learning for fire detection, much of the current research prioritizes model-centric metrics over dataset fidelity, particularly from a fire safety engineering perspective. Commonly used datasets are often dominated by fully developed flames, mislabel smoke-only frames as non-fire, or lack intra-video diversity due to redundant frames from limited sources. Some works treat smoke detection alone as early-stage detection, even though many fires (e.g., electrical or chemical) begin with visible flames and no smoke. Additionally, attempts to improve model applicability through mixed-context datasets—combining indoor, outdoor, and wildland scenes—often overlook the unique false alarm sources and detection challenges specific to each environment. To address these limitations, we curated a new video dataset comprising 1108 annotated fire and non-fire clips captured via indoor surveillance cameras. Unlike existing datasets, ours emphasizes early-stage fire dynamics (pre-flashover) and includes varied fire sources (e.g., sofa, cupboard, and attic fires), realistic false alarm triggers (e.g., flame-colored objects, artificial lighting), and a wide range of spatial layouts and illumination conditions. This collection enables robust training and benchmarking for early indoor fire detection. Using this dataset, we developed a spatiotemporal fire detection model based on the mixed convolutions ResNets (MC3_18) architecture, augmented with Convolutional Block Attention Modules (CBAM). The proposed model achieved 86.11% accuracy, 88.76% precision, and 84.04% recall, along with low false positive (11.63%) and false negative (15.96%) rates. Compared to its CBAM-free baseline, the model exhibits notable improvements in F1-score and interpretability, as confirmed by Grad-CAM++ visualizations highlighting attention to semantically meaningful fire features. These results demonstrate that effective early fire detection is inseparable from high-quality, context-specific datasets. Our work introduces a scalable, safety-driven approach that advances the development of reliable, interpretable, and deployment-ready fire detection systems for residential environments. Full article

Current tomato harvesters rely primarily on external color as the sole indicator of ripeness. However, this approach often results in premature harvesting, leading to insufficient lycopene accumulation and a suboptimal nutritional content for human consumption. Such limitations are especially critical in controlled-environment agriculture (CEA) systems, where maximizing fruit quality and nutrient density is essential for both the yield and consumer health. To address that challenge, this study introduces a novel, multimodal harvest readiness framework tailored to nutrient film technology (NFT)-based smart farms. The proposed approach integrates plant-level stress diagnostics and fruit-level phenotyping using wearable biosensors, AI-assisted computer vision, and non-invasive physiological sensing. Key physiological markers—including the volatile organic compound (VOC) methanol, phytohormones salicylic acid (SA) and indole-3-acetic acid (IAA), and nutrients nitrate and ammonium concentrations—are combined with phenomic traits such as fruit color (a*), size, chlorophyll index (rGb), and water status. The innovation lies in a four-stage decision-making pipeline that filters physiologically stressed plants before selecting ripened fruits based on internal and external quality indicators. Experimental validation across four plant conditions (control, water-stressed, light-stressed, and wounded) demonstrated the efficacy of VOC and hormone sensors in identifying optimal harvest candidates. Additionally, the integration of low-cost electrochemical ion sensors provides scalable nutrient monitoring within NFT systems. This research delivers a robust, sensor-driven framework for autonomous, data-informed harvesting decisions in smart indoor agriculture. By fusing real-time physiological feedback with AI-enhanced phenotyping, the system advances precision harvest timing, improves fruit nutritional quality, and sets the foundation for resilient, feedback-controlled farming platforms suited to meeting global food security and sustainability demands. Full article

Robust indoor robot navigation typically demands either costly sensors or extensive training data. We propose a cost-effective RGB-D navigation pipeline that couples feature-based relative pose estimation with a lightweight multi-layer-perceptron (MLP) policy. RGB-D keyframes extracted from human-driven traversals form nodes of a topological map; edges are added when visual similarity and geometric–kinematic constraints are jointly satisfied. During autonomy, LightGlue features and SVD give six-DoF relative pose to the active keyframe, and the MLP predicts one of four discrete actions. Low visual similarity or detected obstacles trigger graph editing and Dijkstra replanning in real time. Across eight tasks in four Habitat-Sim environments, the agent covered 190.44 m, replanning when required, and consistently stopped within 0.1 m of the goal while running on commodity hardware. An information-theoretic analysis over the Multi-Illumination dataset shows that LightGlue maximizes per-second information gain under lighting changes, motivating its selection. The modular design attains reliable navigation without metric SLAM or large-scale learning, and seamlessly accommodates future perception or policy upgrades. Full article

This work demonstrates an ultra-compact optical gas-sensing system, consisting of a pigtailed laser diode emitting at 1392.5 nm for water vapor (H₂O) detection, a silicon nitride (Si₃N₄) optical waveguide to guide the laser light, and a custom-designed, low-frequency, and T-shaped Quartz Tuning Fork (QTF) as the sensitive element. The system employs both Quartz-Enhanced Photoacoustic Spectroscopy (QEPAS) and Light-Induced Thermoelastic Spectroscopy (LITES) techniques for trace gas sensing. A 3.8 mm-wide, S-shaped waveguide path was designed to prevent scattered laser light from directly illuminating the QTF. Both QEPAS and LITES demonstrated comparably low signal-to-noise ratios (SNRs), ranging from 1.6 to 3.2 for a 1.6% indoor H₂O concentration, primarily owing to the reduced optical power (~300 μW) delivered to the QTF excitation point. These results demonstrate the feasibility of integrating photonic devices and piezoelectric components into portable gas-sensing systems for challenging environments. Full article

Electric wheelchairs significantly enhance mobility for individuals with disabilities, but navigating confined or crowded spaces remains a challenge. This paper presents a mechatronic anticollision system based on Time-of-Flight (ToF) sensors designed to improve wheelchair navigation in such environments. The system performs eight-plane 3D environmental scans in 214–358 ms, with a vertical field of view of 12.4° and a detection range of up to 4 m—sufficient for effective obstacle avoidance. Unlike existing solutions like the YDLIDAR T-mini Plus, which has a narrow vertical field of view and a longer detection range that may be excessive for indoor spaces, or the xLIDAR, which struggles with shorter detection ranges, our system balances an optimal detection range and vertical scanning area, making it especially suitable for wheelchair users. Preliminary tests confirm that our system achieves high accuracy, with a standard deviation as low as 0.003 m and a maximum deviation below 0.05 m at a 3-m range on high-reflectivity surfaces (e.g., white and light brown). Our solution offers low power consumption (140 mA) and USB communication, making it an energy-efficient and easy-to-integrate solution for electric wheelchairs. Future work will focus on enhancing angular precision and robustness for dynamic, real-world environments. Full article

The production of duckweed (Lemnaceae) as a novel protein source could make a valuable contribution to human nutrition. The greatly reduced habitus of duckweed enables simple cultivation with extremely low space requirements, making this free-floating freshwater plant ideal for substrate-free and vertical cultivation in controlled environment agriculture. Of particular importance in the design of a plant-producing Indoor Vertical Farming process is the determination of light intensity, as artificial lighting is generally the most energy-intensive feature of daylight-independent cultivation systems. In order to make the production process both cost-effective and low emission in the future, it is, therefore, crucial to understand and mathematically describe the primary metabolism, in particular the light utilization efficiency. To achieve this, a growth model was developed that mathematically describes the combined effects of plant density and light intensity on the growth rate of Lemna minor L. and physiologically explains the intraspecific competition of plants for light through mutual shading. Furthermore, the growth model can be utilized to derive environmental and process parameters, including optimum harvest quantities and efficiency-optimized light intensities to improve the production process. Full article

To address the challenges of low pesticide utilization, insufficient automation, and health risks in orchard plant protection, we developed an autonomous spraying vehicle using ArduPilot firmware and a robot operating system (ROS). The system tackles orchard navigation hurdles, including global navigation satellite system (GNSS) signal obstruction, light detection and ranging (LIDAR) simultaneous localization and mapping (SLAM) error accumulation, and lighting-limited visual positioning. A key innovation is the integration of an extended Kalman filter (EKF) to dynamically fuse T265 visual odometry, inertial measurement unit (IMU), and GPS data, overcoming single-sensor limitations and enhancing positioning robustness in complex environments. Additionally, the study optimizes PID controller derivative parameters for tracked chassis, improving acceleration/deceleration control smoothness. The system, composed of Pixhawk 4, Raspberry Pi 4B, Silan S2L LIDAR, T265 visual odometry, and a Quectel EC200A 4G module, enables autonomous path planning, real-time obstacle avoidance, and multi-mission navigation. Indoor/outdoor tests and field experiments in Sun Village Orchard validated its autonomous cruising and obstacle avoidance capabilities under real-world orchard conditions, demonstrating feasibility for intelligent plant protection. Full article

Current research on aged housing prioritizes community planning and environmental enhancement over older adults’ needs, creating a retrofit mismatch amid population aging. To investigate the relationship between indoor environmental quality and residential satisfaction among elderly occupants, this study examines 72 households in aged residential buildings, analyzing four environmental indicators (thermal, lighting, acoustic environments, and air quality). The environmental measurements reveal that 81.9% of thermal environment parameters fall below the ASHRAE-55 comfort range, with winter average temperatures reaching only 13.94 °C. Insufficient illumination exists in kitchen and bedroom areas. Lifestyle patterns including infrequent air conditioning use (87%) and window ventilation substituting range hoods (32%) may deteriorate thermal comfort and air quality. An ordered logistic regression analysis demonstrates significant correlations between all four environmental indicators and elderly satisfaction levels. Thermal comfort emerges as the priority focus for aging-adapted retrofitting. Air quality improvement shows particularly significant potential for enhancing residential satisfaction. Although prolonged window opening (73%) exacerbates low-temperature/high-humidity conditions and noise exposure, it still contributes positively to overall satisfaction. This research provides crucial insights for aligning aged residential retrofitting with home-based elderly care requirements, promoting housing development that better accommodates the lifestyle patterns of older populations, thereby improving quality of life for aging-in-place residents. Full article

Visual inspections of concrete infrastructure in low-light environments require external lighting to ensure adequate visibility. Directional lighting sources, where an image scene is illuminated with an angled lighting source from one direction, can enhance the visibility of surface defects in an image. This paper compares directional and diffused scene illumination images for pixel-level concrete crack segmentation. A novel directional lighting image segmentation algorithm is proposed, which applies crack segmentation image processing techniques to each directionally lit image before combining all images into a single output, highlighting the extremities of the defect. This method was benchmarked against two diffused lighting crack detection techniques across a dataset with crack widths typically ranging from 0.07 mm to 0.4 mm. When tested on cracked and uncracked data, the directional lighting method significantly outperformed other benchmarked diffused lighting methods, attaining a 10% higher true-positive rate (TPR), 12% higher intersection over union (IoU), and 10% higher F1 score with minimal impact on precision. Further testing on only cracked data revealed that directional lighting was superior across all crack widths in the dataset. This research shows that directional lighting can enhance pixel-level crack segmentation in infrastructure requiring external illumination, such as low-light indoor spaces (e.g., tunnels and containment structures) or night-time outdoor inspections (e.g., pavement and bridges). Full article

In southern Hunan province, a vital element of China’s architectural cultural legacy, the quality of the indoor lighting environment influences physical performance and the transmission of spatial culture. The province encounters minor environmental disparities and diminishing liveability attributed to evolving construction practices and cultural standards. The three varieties of traditional residences in Shanggantang Village are employed to assess the daylight factor (DF), illumination uniformity (U0), daylight autonomy (DA), and useful daylight illumination (UDI). We subsequently integrate field measurements with static and dynamic numerical simulations to create a multi-dimensional analytical framework termed “measured-static-dynamic”. This method enables the examination of the influence of floor plan layout on light, as well as the relationship between window size, building configuration, and natural illumination. The lighting factor (DF) of the core area of the central patio-type residence reaches 27.7% and the illumination uniformity (U0) is 0.62, but the DF of the transition area plummets to 1.6%; the composite patio type enhances the DF of the transition area to 1.2% through the alleyway-assisted lighting, which is a 24-fold improvement over the offset patio type. Parameter optimization showed that the percentage of all-natural daylighting time (DA) in the edge zone of the central patio type increased from 21.4% to 58.3% when the window height was adjusted to 90%. The results of the study provide a quantitative basis for the optimization of the light environment and low-carbon renewal of traditional residential buildings. Full article

Ensuring secure and efficient water level monitoring is critical for the intelligent management of hydropower plants, especially in challenging indoor environments. Existing methods, which are tailored for open areas with optimal conditions (adequate lighting, absence of debris interference, etc.), frequently falter in scenarios characterized by poor lighting, water vapor, and confined spaces. To address this challenge, this study introduces a robust indoor water level monitoring framework specifically for hydropower plants. This framework integrates a temporal super-resolution technique with an improved Yolov5 model. Specifically, to enhance the quality of indoor monitoring images, we propose a temporal super-resolution enhancement module. This module processes low-resolution water-level images to generate high-resolution outputs, thereby enabling reliable detection even in suboptimal conditions. Furthermore, unlike existing complex model-based approaches, our enhanced, lightweight Yolov5 model, featuring a small-scale feature mapping branch, ensures real-time monitoring and accurate detection across a variety of conditions, including daytime, nighttime, misty conditions, and wet surfaces. Experimental evaluations demonstrate the framework’s high accuracy, reliability, and operational efficiency, with recognition speeds reaching O(n). This approach is suitable for deployment in emerging intelligent systems, such as HT-for-Web analysis software 0.2.3 and warning platforms, providing vital support for hydropower plant security and emergency management. Full article

Silicon substrate golden light LED, as an emerging blue-light-free health lighting technology, has become one of the key technologies for home health lighting environments. This study uses silicon substrate golden light LED as the lighting source for home lighting, and based on the lighting demands of two indoor types, employs DIALux Evo lighting simulation software to simulate the indoor lighting environment. First, the simulated lighting data for various indoor areas are compared with the national lighting standards (GB/T50034-2024) to verify whether the lighting type meets the home lighting requirements. Next, a comparison is made between the lighting efficiency of silicon substrate golden light LED and a reference sample LED to validate whether the silicon substrate golden light LED possesses high lighting efficiency and low power consumption. Finally, long-term exposure to both the silicon substrate golden light LED and reference sample LED is used to record the secretion levels of melatonin in the human body. The experimental results show that the silicon substrate golden light LED not only provides sufficient home lighting but also demonstrates high efficiency and low power consumption. Additionally, under the illumination of silicon substrate golden light LED, the melatonin secretion concentration significantly increases to (960 ± 15) pg/mL after 2.5 h of exposure, which is 8.2 times higher than that of the conventional LED group (t = 12.34, df = 14, p < 0.001). The silicon substrate golden light LED technology provides a feasible solution for home health lighting design by creating a zero-blue-light health lighting environment. Full article

Carbon dioxide (CO₂) is one of the important factors affecting vegetable yield in controlled environments. This study used cucumber as an experimental material to investigate the effects of hanging different amounts of CO₂ generation bags (CGBs) on the growth of temperature-loving vegetables under facility soil cultivation. CGBs of three different densities were set up: no application (TC), eight bags/265 m² (T1), and sixteen bags/265 m² (T2). The results showed the following: (1) Hanging CGBs at different densities significantly impacted indoor CO₂ concentration. Light, temperature, and humidity also affected CO₂ concentration to a certain extent. (2) The application of CGBs improved cucumber growth, photosynthesis, and quality-related indexes, resulting in a 28.9% increased yield compared to the control group. (3) The economic benefits of CGB application in each group were analyzed, revealing the economic benefits of high-density CGB cultivation on solar greenhouse cucumber. This study explored a low-cost and effective CO₂ generation application mode. Full article

Innovation in cultivation methods is essential to address the growing challenges in agriculture, including abiotic and biotic stress, soil degradation, and climate change. Aeroponics, a particular type of hydroponics, presents a promising solution by improving yield and resource use efficiency, especially in controlled environments such as plant factories with artificial lighting (PFALs). Additionally, non-thermal plasma (NTP), a partially ionized gas containing reactive oxygen and nitrogen species, can affect plant development and physiology, further enhancing crop production. The objective of this study was to explore the potential of NTP as an innovative method to enhance crop production by treating the nutrient solution in aeroponic systems. During this study, three experiments were conducted to assess the effects of NTP-treated nutrient solutions on baby leaf lettuce (Lactuca sativa L.) aeroponically grown indoors. The nutrient solution was treated with ionized air in a treatment column separated from the aeroponic system by making the ionized air bubble from the bottom of the column. After 2 min of NTP application, a pump took the nutrient solution from the treatment column and sprayed it on the roots of plants. Various frequencies of NTP application were tested, ranging from 2.5% to 50% of irrigation events with nutrient solution activated with NTP. Results indicated that low-frequency NTP treatments (up to 5% of irrigations) stimulated plant growth, increasing leaf biomass (+18–19%) and enhancing the concentration of flavonoids (+16–18%), phenols (+20–21%), and antioxidant capacity (+29–53%). However, higher NTP frequencies (25% and above) negatively impacted plant growth, reducing fresh and dry weight and root biomass, likely due to excessive oxidative stress. The study demonstrates the potential of NTP as a tool for improving crop quality and yields in aeroponic cultivation, with optimal benefits achieved at lower treatment frequencies. Full article

The advancement of artificial intelligence has brought visible-light positioning (VLP) to the forefront of indoor positioning research, enabling precise localization without additional infrastructure. However, the complex interplay between light propagation phenomena and environmental factors in indoor spaces presents significant challenges for VLP systems. Additionally, the pose of the light-emitting diodes is prior unknown, adding another layer of complexity to the positioning process. Dynamic indoor environments further complicate matters due to user mobility and obstacles, which can affect system accuracy. In this study, user movement is simulated using a constructed dataset with systematically varied receiver positions, reflecting realistic motion patterns rather than real-time movement. While the experimental setup considers a fixed obstacle scenario, the training and testing datasets incorporate position variations to emulate user displacement. Given these dataset characteristics, it is crucial to employ robust positioning techniques that can handle environmental variations. Conventional methods, such as received signal strength (RSS)-based techniques, face practical implementation hurdles due to fluctuations in transmitted optical power and modeling imperfections. Leveraging machine learning techniques, particularly regression-based artificial neural networks (ANNs), offer a promising alternative. ANNs excel at modeling the intricate relationships within data, making them well-suited for handling the complex dynamics of indoor lighting environments. To address the computational complexities arising from high-dimensional data, this research incorporates principal component analysis (PCA) as a method for reducing dimensionality. PCA eases the computational burden, accelerates training speeds by normalizing the data, and reduces loss rates, thereby enhancing the overall efficacy and feasibility of the proposed VLP framework. Rigorous experimentation and validation demonstrate the potential of employing principal components. Experimental results show significant improvements across multiple evaluation metrics for a constellation comprising eight LEDs mounted in a rectangular structure measuring a room dimension of 12 m × 18 m × 6.8 m, with a photodiode (PD) receiver. Specifically, the mean squared error (MSE) values for the training and testing samples are 0.0062 and 0.0456 cm, respectively. Furthermore, the R-squared values of 99.31% and 94.74% for training and testing, respectively, signify a robust predictive performance of the model with low model loss. These findings underscore the efficacy of the proposed PCA-ANN regression model in optimizing VLP systems and providing reliable indoor positioning services. Full article