Abstract Rotameters are reliable primary sensors for measuring the flow rate of fluids in pipelines. However, they have the limitation of being local, indicative sensors that rely on vertically mounted, breakable, transparent, or semi-transparent tapered tubes, suitable only for transparent liquids. To enable remote indication of flow rates or for control purposes, a non-contact secondary sensor is necessary. This paper investigates a modified non-contact rotameter transducer using a magneto-resistive (MR) element as the secondary sensor. A signal conditioner circuit is designed to transmit the rotameter reading and transducer output to a remote location for monitoring and control. The theoretical equations explaining the sensor, transducer, and transmitter’s performance are derived and discussed. Experimental studies on the prototype transmitter have been conducted, with results showing good linearity and repeatability, validating the theoretical model. Furthermore, artificial neural networks (ANN) have been applied to optimize the flow rate measurement by enhancing accuracy and reducing any environmental impacts on sensor performance. ANN-based models provide better calibration and fault tolerance, ensuring improved precision in flow rate detection even under varying environmental conditions. The experimental results and their correlation with the theoretical equations are presented, confirming the effectiveness of the proposed system.
Abstract Hyper spectral image has drawn great attention in applications like crop-disease detection, geological mapping, remote-sensing, environmental monitor, and defense surveillance, etc. Nonetheless, the high dimensionality and big data volume due to many spectral bands present an important obstacle for efficient storage and transmission. To address this issue, we propose a hybrid compression framework that decreases both spatial and spectral redundancies and is capable of accurately modeling non-linearities. The discrete wavelet transform provides a frequency domain representation of the image features, minimizing the spatial redundancy. Tucker Decomposition exploiting spectral redundancy through higher-order singular value decomposition by exploiting inter-band correlations. These techniques assume linear relationships between parameters, and to get around this limitation, the proposed method includes a non-linear iterative partial least squares model, which effectively models the non-linear relationships. We validate the proposed method on four public hyperspectral datasets such as Cuprite, Pavia University, Indian Pines Corrected, and Salinas. Experimental results show a very high compression factor of 280, which is better than Tucker and CANDECOMP/PARAFAC (CP) decomposition methods by margins of 199 and 61, respectively. Peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM)—High PSNR and SSIM values are attained despite the high compression. In addition, the proposed method outperforms traditional tensor decomposition approaches such as CP and Tucker, and recent methods for hyperspectral image compression, including hybrid transform with wavelet-based transform coding, long short-term memory autoencoder based compression as well as convolutional neural network-based lossy compressive methods under consistent experimental conditions.
Abstract The rapid growth of user-generated content on social media platforms has reinforced the urgency of detecting abusive language automatically. This includes identifying cyber-bullying, hate speech, and toxic sentiment. This paper introduces PCA-QLSTM-stacked ensemble (SE), a unified hybrid framework that combines principal component analysis (PCA) based quantum long short-term memory (QLSTM) neural network. The framework is optimized by black-winged kite optimization (BWKO) and augmented by a stacked ensemble of support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost [XGB]) classifiers. Firstly, the Text features are extracted using a hybrid term frequency–inverse document frequency (TF-IDF) and multi-scale hash vectorization (HV) pipeline. The features are then reduced by PCA retaining 95% cumulative variance, before classification by the hybrid model. Experiments across five benchmark datasets (internet movie dataset [IMDB], Yelp Polarity, Cyber-bullying Classification, Kaggle Hate Speech, and a combined corpus) demonstrate the accuracy of 96.41%, 93.87%, 98.12%, and 96.80% respectively. Ablation studies confirm the independent contribution of each component, and Friedman-Nemenyi statistical tests validate significance at alpha = 0.05.
Abstract Acoustic signals play a vital role in animal communication and offer a valuable, non-invasive means of assessing animal health and welfare. Insights from bioacoustic analysis allow livestock farmers to adopt sound-based monitoring systems that improve farm management while minimizing direct interaction with animals. The use of audio and acoustic sensors provides a practical and animal-friendly approach for monitoring large populations, enabling continuous, real-time observation that is less intrusive than conventional physiological methods. However, several challenges persist, including limited availability of high-quality datasets, variability in recording environments, and difficulties in interpreting model outputs, which collectively hinder large-scale deployment in farm settings. To address these limitations, the incorporation of advanced data augmentation strategies and multi-modal sensor integration has been suggested to enhance data diversity and improve classification performance. Current research efforts are focused on developing cost-effective, real-time, and robust bio-acoustic monitoring systems that are adaptable to varying farm conditions and species. Such systems have the potential to transform livestock management by enabling early detection of health and welfare issues, thereby supporting more sustainable and ethically responsible farming practices. This paper provides a comprehensive review of bioacoustic-based approaches for animal welfare monitoring, encompassing a wide range of techniques including individual identification, stress detection, estrus recognition, and respiratory condition analysis. Spanning over a decade of research, the current study underscores the evolving use of machine learning and deep neural networks, including convolutional and transformer-based models, in decoding complex animal vocalizations. This review develops a comprehensive taxonomy and provides a critical synthesis of sensor-enabled bioacoustic systems for livestock and wildlife monitoring. It further outlines major research gaps and practical deployment challenges, proposing a forward-looking framework to facilitate scalable and real-world implementation of these systems.
Abstract The grading electrodes in the water-cooling system is crucial for the stable operation of the converter valve in ultra-high voltage converter station. However, incrustation can significantly affect the performance of the electrode. Scaling accumulation in these systems can lead to reduced heat dissipation efficiency, increased energy consumption, and even equipment failure. Traditional scaling detection methods, such as manual inspection and chemical analysis, are often inefficient, subjective, and unable to provide real-time monitoring. To address these challenges, this study proposes an intelligent scaling recognition method based on X-ray imaging and deep learning. A YOLOv8_OBB based model is designed to automatically detect and classify scaling on grading electrodes. The proposed approach utilizes electrode diameter regression and corrected rules to assess the severity of scaling. The accuracy of this model in detecting scaling reaches 98.5%, and the average error in diameter prediction is 0.016mm. Experimental results demonstrate that the model achieves high precision in identifying scales and measure diameter. This research contributes to the development of automatic online monitoring equipment in the future.
