Layer Separation Research Articles

Advanced Adaptive Protein Membrane with Super-High Flux and Precise Molecular Sieving of Dyes from Salts.

Precise separation of small organic molecules and electrolyte salts is critical for various industrial processes, necessitating advanced membranes with uniform pores. Proteins, as one typical nature polymer having the exactly same structure and molecular weight, offer a promising material for making such membranes. Here, hemoglobin (BHb) and lysozyme (Lyz) are utilized to fabricate precise nanofiltration membranes through amyloid-like co-assembly, triggered by Tris(2-carboxyethyl) phosphine. Molecular dynamics simulations reveal that BHb intercalates between Lyz molecules, enhancing tight assembly and reducing defects. The Lyz/BHb membrane exhibits a high void volume of 27.2% and achieves an exceptional permeance of 335 Lm-2h-1bar-1. Amazingly, the Lyz/BHb membrane achieves ultra-selective molecular sieving of dyes and salts with an unparalleled high separation factor of 1.25 × 103 for the NaCl/Brilliant blue. The unique adaptive separation layer is formed by anchoring dye molecules into the larger pores and thus narrowing down the pore size distribution and enhancing electrostatic repulsion, endowing the membrane with a distinctive molecular sieving of dyes and salts. This study offers valuable insights to finely tailor the pore structure of the membrane by co-assembly of proteins and to significantly enhance the membrane perm-selectivity by creating an adaptive protein separation layer with a unique molecular sieving property.

Lung and Colon Cancer Classification Using Multiscale Deep Features Integration of Compact Convolutional Neural Networks and Feature Selection

The automated and precise classification of lung and colon cancer from histopathological photos continues to pose a significant challenge in medical diagnosis, as current computer-aided diagnosis (CAD) systems are frequently constrained by their dependence on singular deep learning architectures, elevated computational complexity, and their ineffectiveness in utilising multiscale features. To this end, the present research introduces a CAD system that integrates several lightweight convolutional neural networks (CNNs) with dual-layer feature extraction and feature selection to overcome the aforementioned constraints. Initially, it extracts deep attributes from two separate layers (pooling and fully connected) of three pre-trained CNNs (MobileNet, ResNet-18, and EfficientNetB0). Second, the system uses the benefits of canonical correlation analysis for dimensionality reduction in pooling layer attributes to reduce complexity. In addition, it integrates the dual-layer features to encapsulate both high- and low-level representations. Finally, to benefit from multiple deep network architectures while reducing classification complexity, the proposed CAD merges dual deep layer variables of the three CNNs and then applies the analysis of variance (ANOVA) and Chi-Squared for the selection of the most discriminative features from the integrated CNN architectures. The CAD is assessed on the LC25000 dataset leveraging eight distinct classifiers, encompassing various Support Vector Machine (SVM) variants, Decision Trees, Linear Discriminant Analysis, and k-nearest neighbours. The experimental results exhibited outstanding performance, attaining 99.8% classification accuracy with cubic SVM classifiers employing merely 50 ANOVA-selected features, exceeding the performance of individual CNNs while markedly diminishing computational complexity. The framework’s capacity to sustain exceptional accuracy with a limited feature set renders it especially advantageous for clinical applications where diagnostic precision and efficiency are critical. These findings confirm the efficacy of the multi-CNN, multi-layer methodology in enhancing cancer classification precision while mitigating the computational constraints of current systems.

Magnetohydrodynamics (MHD) flow of ternary nanofluid and heat transfer past a permeable cylinder with velocity slip

The increasing demand for energy-efficient and environmentally sustainable solutions has driven the exploration of ternary nanofluids for enhanced heat transfer applications. This numerical study investigates the magnetohydrodynamics (MHD) flow and heat transfer characteristics of a ternary hybrid nanofluid (copper-alumina-titania/water) over a shrinking/stretching cylinder, considering the effects of velocity slip, suction, and curvature parameters. The variable wall temperature is also included in the physical model. The model in PDEs is transformed into a simplified version of differential equations (ODEs) which can be solved using the bvp4c solver. Validation against previously published data confirms the accuracy of the model. Two solutions are possible if the shrinking surface is permeable (with suction effect) while only unique solution for the stretching case. Surprisingly, the copper-alumina-titania/water exhibits a lower heat transfer rate but higher skin friction as compared to the copper-alumina/water. Specifically, the heat transfer coefficient decreases by 5-10% when incorporating the titania nanoparticle, while an increase of 8-12% for the skin friction coefficient. Additionally, the velocity slip and curvature parameters accelerate the boundary layer separation, whereas increased suction delays this process. These findings provide insights into optimizing thermal management systems using ternary hybrid nanofluids, particularly under MHD effects.

Spectral proper orthogonal decomposition of turbine tip clearance flow under low-Reynolds number conditions

Spectral proper orthogonal decomposition, a data-based decomposition technique, was performed to extract key spatiotemporal coherent structures from large-eddy simulation data. Both velocity and vorticity were decomposed to identify the dominant unsteady features around the near-tip region. Four operating conditions for two different Reynolds numbers, with and without tip clearance were investigated, i.e., low_Re TC, low_Re, high_Re TC and high_Re. The energy spectra (eigenvalues) and their accumulation at each frequency exhibited the low-rank behavior, indicating the presence of a physically dominating mechanism. The velocity mode shapes (eigenvectors) revealed three typical coherent structures for the low_Re TC condition: vortex shedding from the blade wake and the tip leakage vortex, Kelvin-Helmholtz instability induced by the leakage jet, and complex interactions among the boundary layer separation, the tip leakage vortex and the horseshoe vortex. For the condition without tip clearance, the dominant structure was the vortex shedding from the wake only. The vorticity mode shapes indicated that high dissipation of the vorticity field occurred close to the suction side, where strong secondary vortices existed. The present research could develop a better understanding of turbine tip clearance flow and loss mechanisms at low Reynolds numbers.

Oligomerized Electron Acceptors with Alkynyl Linkages to Suppress Electron-Photon Coupling for Low-Energy-Loss Organic Solar Cells.

Oligomerized electron acceptors, featuring molecular weights akin to polymers and well-defined chemical structures, have emerged as promising candidates for organic solar cells (OSCs) due to their consistent batch-to-batch reproducibility and improved thermal stability. In this study, we developed a series of oligomerized electron acceptors incorporating alkynyl linkages via an efficient Sonogashira coupling reaction between alkyne-substituted Y-type precursors and multi-substituted iodobenzenes. This method produced monomeric (S-Alkyne-YF), dimeric (D-Alkyne-YF), and trimeric (T-Alkyne-YF) configurations, enabling systematic control over molecular size and substituent arms. The alkynyl linkages, characterized by high bond strength and planar geometry, enhanced molecular planarity and aggregation in films, thus facilitating precise control over morphology and phase separation in the photoactive layers. Notably, the inclusion of these linkages effectively suppressed electron-phonon coupling, resulting in reduced non-radiative energy losses and elevated photocarrier lifetime. OSCs based on PM6:T-Alkyne-YF achieved a power conversion efficiency of 17.90%, a low non-radiative energy loss of 0.185 eV, and an open-circuit voltage of 0.943 V. Furthermore, integrating T-Alkyne-YF into the D18:N3 blend yielded an exceptional PCE of 19.52%. These results underscore the potential of alkynyl-linked oligomerized acceptors in advancing highly efficient and stable OSCs, offering a viable pathway for reducing electron-phonon coupling and enhancing device performance.

Engineered ionic polymer metal composites (eIPMCs) under dynamic compression loading conditions: theory and experiments

Abstract Engineered Ionic Polymer Metal Composites (eIPMCs) represent the next generation of IPMCs, soft electro-chemo-mechanically coupled smart materials used as actuators and sensors. Recent studies indicate that eIPMC sensors, featuring unique microstructures at the interface between the ionic polymer membrane and the electrode, exhibit enhanced electrochemical behavior and sensitivity under compression, as compared to traditional IPMCs. However, a complete and experimentally-validated model of how eIPMCs behave under dynamic compression loads is currently missing. In this paper, we develop an analytical model for eIPMC sensors, elucidating the role of the engineered interface, modeled as a separate material layer with unique mechanical and electrochemical properties. Theoretical predictions focus on the mechanical-to-electrochemical transduction response under dynamic compressive loads. Experimental verification is conducted on conventional IPMC and novel eIPMC samples fabricated using the polymer abrading technique. Electrochemical impedance spectroscopy is performed to study the effect of the engineered interface on the electrochemical properties. Open-circuit voltage and short-circuit current are measured under external compressive loads in different loading scenarios to demonstrate sensing performance. Results show good qualitative agreement between experimental trends and model predictions. Experiments over the frequency range 1-18 Hz demonstrate an increase of 220-290% in open-circuit voltage and 17-166% in short-circuit current sensitivity for eIPMCs over IPMCs.

Tgfbr1 regulates lateral plate mesoderm and endoderm reorganization during the trunk to tail transition.

During the trunk to tail transition the mammalian embryo builds the outlets for the intestinal and urogenital tracts, lays down the primordia for the hindlimb and external genitalia, and switches from the epiblast/primitive streak (PS) to the tail bud as the driver of axial extension. Genetic and molecular data indicate that Tgfbr1 is a key regulator of the trunk to tail transition. Tgfbr1 has been shown to control the switch of the neuromesodermal competent cells from the epiblast to the chordoneural hinge to generate the tail bud. We now show that in mouse embryos Tgfbr1 signaling also controls the remodeling of the lateral plate mesoderm (LPM) and of the embryonic endoderm associated with the trunk to tail transition. In the absence of Tgfbr1, the two LPM layers do not converge at the end of the trunk, extending instead as separate layers until the caudal embryonic extremity, and failing to activate markers of primordia for the hindlimb and external genitalia. The vascular remodeling involving the dorsal aorta and the umbilical artery leading to the connection between embryonic and extraembryonic circulation was also affected in the Tgfbr1 mutant embryos. Similar alterations in the LPM and vascular system were also observed in Isl1 null mutants, indicating that this factor acts in the regulatory cascade downstream of Tgfbr1 in LPM-derived tissues. In addition, in the absence of Tgfbr1 the embryonic endoderm fails to expand to form the endodermal cloaca and to extend posteriorly to generate the tail gut. We present evidence suggesting that the remodeling activity of Tgfbr1 in the LPM and endoderm results from the control of the posterior PS fate after its regression during the trunk to tail transition. Our data, together with previously reported observations, place Tgfbr1 at the top of the regulatory processes controlling the trunk to tail transition.

Tgfbr1 regulates lateral plate mesoderm and endoderm reorganization during the trunk to tail transition

During the trunk to tail transition the mammalian embryo builds the outlets for the intestinal and urogenital tracts, lays down the primordia for the hindlimb and external genitalia, and switches from the epiblast/primitive streak (PS) to the tail bud as the driver of axial extension. Genetic and molecular data indicate that Tgfbr1 is a key regulator of the trunk to tail transition. Tgfbr1 has been shown to control the switch of the neuromesodermal competent cells from the epiblast to the chordoneural hinge to generate the tail bud. We now show that in mouse embryos Tgfbr1 signaling also controls the remodeling of the lateral plate mesoderm (LPM) and of the embryonic endoderm associated with the trunk to tail transition. In the absence of Tgfbr1, the two LPM layers do not converge at the end of the trunk, extending instead as separate layers until the caudal embryonic extremity, and failing to activate markers of primordia for the hindlimb and external genitalia. The vascular remodeling involving the dorsal aorta and the umbilical artery leading to the connection between embryonic and extraembryonic circulation was also affected in the Tgfbr1 mutant embryos. Similar alterations in the LPM and vascular system were also observed in Isl1 null mutants, indicating that this factor acts in the regulatory cascade downstream of Tgfbr1 in LPM-derived tissues. In addition, in the absence of Tgfbr1 the embryonic endoderm fails to expand to form the endodermal cloaca and to extend posteriorly to generate the tail gut. We present evidence suggesting that the remodeling activity of Tgfbr1 in the LPM and endoderm results from the control of the posterior PS fate after its regression during the trunk to tail transition. Our data, together with previously reported observations, place Tgfbr1 at the top of the regulatory processes controlling the trunk to tail transition.

Half-Covered ‘Glitter-Cake’ AM@SE Composite: A Novel Electrode Design for High Energy Density All-Solid-State Batteries

All-solid-state batteries (ASSBs) are pursued due to their potential for better safety and high energy density. However, the energy density of the cathode for ASSBs does not seem to be satisfactory due to the low utilization of active materials (AMs) at high loading. With small amount of solid electrolyte (SE) powder in the cathode, poor electrochemical performance is often observed due to contact loss and non-homogeneous distribution of AMs and SEs, leading to high tortuosity and limitation of lithium and electron transport pathways. Here, we propose a novel cathode design that can achieve high volumetric energy density of 1258 Wh L-1 at high AM content of 85 wt% by synergizing the merits of AM@SE core-shell composite particles with conformally coated thin SE shell prepared from mechanofusion process and small SE particles. The core-shell structure with an intimate and thin SE shell guarantees high ionic conduction pathway while unharming the electronic conduction. In addition, small SE particles play the role of a filler that reduces the packing porosity in the cathode composite electrode as well as between the cathode and the SE separator layer. The systematic demonstration of the optimization process may provide understanding and guidance on the design of electrodes for ASSBs with high electrode density, capacity, and ultimately energy density.

Dual-function terahertz metamaterial absorber based on microfluidic structures

In recent years, terahertz metamaterial sensors have shown great potential in label-free biosensing; yet, the detection of high-absorption liquid samples that are sensitive to terahertz waves remains a significant challenge. In this study, a dual-function absorber capable of dynamically switching between broadband absorption and high-sensitivity sensing is proposed based on the microfluidic technology and phase change characteristics of vanadium dioxide (VO2). Compared with traditional terahertz microfluidic sensors, this structure differs in that it incorporates a VO2 film as a separation layer in the sensor cover plate and a VO2 square resonator on top. This configuration not only exhibits high-sensitivity sensing but can also function as an absorber for broadband absorption. When VO2 is in the metallic state, the structure acts as a broadband absorber with an absorption rate exceeding 90% across the 1.09–3.02 THz range. When VO2 is in the insulating state, the structure functions as a microfluidic sensor, achieving an absorption rate above 99.9% at 1.438 and 2.068 THz, with nearly perfect absorption and refractive index sensitivities of 532 and 785 GHz/RIU, respectively; the quality factor is 17.6 and 23.5, respectively, indicating excellent sensing performance. Moreover, due to the symmetry of the metal micro-structured layer and the VO2 square resonator, the device exhibits polarization insensitivity and stability at large incident angles. In summary, this structure significantly broadens the applications of traditional absorbers and sensors and holds promise for future applications in electromagnetic cloaking, energy harvesting, and biomedical detection.

Cosmic jurisdictions: quod lege naturae, moribus et consuetudine inductum est

The study is devoted to organizing the legal space of the Universe based on “ius naturale,” morals, and consuetudes of space activities, with the prospects for the existence of extraterrestrial intelligent beings in the Universe. On the grounds of the results of the study, 7 natural factors and 2 technical factors are identified, the influence of which determines the spaces to which the sovereignty and jurisdiction of both states and all of humanity extend. Based on these factors, the NMC Concept “natura, moribus et consuetudines” is proposed, according to which, due to natural and other factors, all outer space above the Earth`s surface is divided into a “domestic room” (our Solar System) and an “alien room” (outside the Solar System), the boundary between which are the Kuiper Belt and the center of the Sun. At the same time, it is proposed to apply the principle of “Res Communis Humanitatus” to the “domestic room”, and the new principle of “Res Nullius Civitatis et Res Communis Animal Rationale” to the “alien room”. Additionally, the authors emphasize the fact that the above factors have already formed the structure of the “domestic room” regardless of the existing intentions and suggestions, leaving the world to recognize and accept it. Within the structure of the “domestic room”, the following two types of spaces have been formed: the “unified and sovereign spatial-territorial domains of states”, to which the exclusive jurisdiction of states extends, and “Res Communis Humanitatus”, to which generally recognized international law extends. In these circumstances, considering all natural and technical factors, the authors take notice of the allocation of two separate layers in “Res Communis Humanitatus”: “Orbital layer” and “Ballistic space”. The authors also propose developing a “sanitary atmospheric zone” to ensure humanity`s safety from the effects of the “X” factor. Moreover, the authors suggest applying different principles of international regulation to the “Ballistic space,” the “Orbital layer” and the space beyond it, based on the mechanisms of “tacit consent,” “silent disapproval”, “active consent”, and “active disapproval”. The result of this research is a draft Pact for the Cosmos.

Hardware-Aware Color-Distorted Image Classifier for Intelligent White Balance

Automatic white balance (AWB) is an important module for cameras and classification of the color-distorted image is critical to realize intelligent AWB. Though accurate classifiers usually can be achieved via deep neural network models, they cannot fit into embedded hardware due to their complexity. To increase the classification accuracy and decrease latency, a lightweight convolutional neural network (CNN) with a histogram layer for AWB (AWBHNet) is constructed, which consists of one histogram layer, one regular convolutional layer, three depth separable convolutional layers, four pooling layers, two fully connected layers and two dropout layers. One-tenth of ImageNet is utilized as the normal image dataset. To generate various distorted colors, histogram shifting and matching are proposed to randomly adjust the histogram position or shape. Furthermore, the extent of shifting or matching is randomly generated to ensure the diversity of color distortion. Subsequently, the proposed AWBHNet and other CNNs are successively trained. Experiments show that the accuracy of the classifier trained by AWBHNet is 0.9150, which is at least 1.33% higher than each regular or lightweight network. Finally, intelligent AWB is realized on smartphones, the inference latency of AWBHNet is 47.6% lower than the best existing network.

Investigation on Flow Features and Combustion Characteristics in a Boron-Based Solid-Ducted Rocket Engine

Numerical and experimental approaches are conducted to investigate the flow features and secondary combustion performance induced by different air–fuel ratios in a boron-based solid-ducted rocket engine. The results indicated that the afterburning chamber flow features become more complicated owing to the multiple nozzles of the gas injector, and a number of recirculation zones are generated. Because of this, the mixing of the fuel gas and incoming air is enhanced. When the air–fuel ratio decreases, the heat release in the afterburning chamber increases continuously, which causes the pre-combustion shock train to continue to propagate upstream in the subsonic diffuser of the inlet isolator, along with the boundary layer separation zone also moving forward, and the stability margin of the direct-connect inlet decreasing gradually. Furthermore, the direct-connect inlet works at a critical state with an air–fuel ratio of 11.5. As the mass flow rate of the fuel-rich gas rises gradually, the engine thrust gradually increases, and the number of vortexes in the afterburning chamber and the corresponding region affected by the vortexes generally decrease. Meanwhile, the mixing and combustion of the fuel-rich gas and incoming flow were not substantially changed. Additionally, the combustion efficiency and specific impulse are proportional to the air fuel ratio.

Challenges in the Design for Disassembly of Light Timber Framing Panelized Components

The construction sector generates more than one-third of global waste. Although there is a consensus on the need to reduce it, empirical research evaluating current systems to develop circular solutions remains limited. Using a full-scale model, this article evaluates the disassemblability of the corner joint between two prefabricated lightweight timber-framed walls, a system widely adopted in residential construction in North America. The analysis deconstructed the disassembly actions, identified their level of difficulty, and classified the recovered materials into three categories: reusable, recyclable, and waste. The results reveal that the lack of design criteria for disassembly significantly limits the system’s circularity, as it prioritizes assembly speed and energy performance. The predominant use of nails as fasteners complicates the separation of layers, damages materials, and restricts their reuse. This highlights the urgent need to redesign construction solutions that enable efficient disassembly, promote component recovery, and extend their time in circulation. This study establishes a foundation for the evolution of lightweight timber-framed panel design toward systems more aligned with circularity principles.

Exploring Moiré Superlattices and Memristive Switching in Non-van der Waals Twisted Bilayer Bi2O2Se.

The discovery of moiré physics in two-dimensional (2D) materials has opened new avenues for exploring unique physical and chemical properties induced by intralayer/interlayer interactions. This study reports the experimental observation of moiré patterns in 2D bismuth oxyselenide (Bi2O2Se) nanosheets grown through one-pot chemical reaction methods and a sonication-assisted layer separations technique. Our findings demonstrate that these moiré patterns result from the angular stacking of the nanosheets at various twist angles, leading to the formation of moiré superlattices (MSLs) with distinct periodicities. The presence of these superlattices was confirmed using transmission electron microscopy (TEM) images. The observation of moiré patterns in 2D Bi2O2Se nanosheets highlights the potential of tuning the band structures of the non-van der Waals material and thus unlocking new material properties through precise control of intralayer/interlayer interactions. Furthermore, the stacked 2D Bi2O2Se nanosheets show interesting memristive switching characteristics, presenting a promising candidate for artificial synapses and neuromorphic computing. Traditional memristors typically utilize a vertical metal-insulator-metal (MIM) structure, which relies on the formation of conductive filaments for resistive switching (RS). This configuration, however, often results in abrupt switching during various cycles and significant variation from device to device. Herein, defective BOSe moiré material exhibits a nonfilamentary RS switching characteristic in a two-terminal lateral device configuration. This design reveals an RS mechanism driven by the modulation of the Schottky barrier height (SBH) due to the movement of Se vacancies (VSe) under an external electric field. The fabricated device exhibits excellent RS behavior, achieving an RS ratio of ∼20 with a high degree of control and consistency across multiple cycles and from device to device. Interestingly, the device shows a stable negative differential resistance effect in the high-voltage region due to the carrier trapping process. Finally, we studied the stability of the MSL in BOSe through TEM imaging and electrical characterization on different device configurations to evaluate the repeatability of the switching characteristics.

Spragues Road Rehabilitation with an Unbonded Concrete Overlay Region of Waterloo, Ontario, Canada.

The work of this project includes the rehabilitation and widening of a 1.2 km (0.75 mile) section of Spragues Road (Regional Road No. 75) from Wrigley Road to the Region of Waterloo boundary in the Township of North Dumfries. The existing road is a rural two-lane composite pavement with partially paved and gravel shoulders. An initial geotechnical investigation determined that conventional road resurfacing strategies of the asphalt / concrete composite pavement were no longer cost-effective and additional assessments were completed to evaluate alternate rehabilitation strategies. Through this process, non-destructive testing was conducted to determine the structural adequacy of the concrete base and suitability for pavement preservation. Comprehensive field investigations determined the 80-year old concrete base was generally in good condition and should be preserved as part of the project. The final pavement design consists of re-profiling or cold planing the existing asphalt surface to meet proposed design gradients, installation of an asphalt intermediate separation layer and placement on an Unbonded Concrete Overlay, a unique rehabilitation treatment in the Province of Ontario.

The effects of wall proximity on the turbulent flow field in a square duct structured with detached divergent ribs on one wall

The effect of the wall proximity of detached 60∘ divergent ribs applied on one wall in a square duct on the turbulent flow field was investigated. Laser Doppler anemometry (LDA) measurements were conducted for different clearance-to-rib-height ratios in the range of 0.1–1.0 at a Reynolds number (based on the rib height and mean bulk velocity) of 5000. Mean velocities, Reynolds stresses, triple velocity correlations as well as skewness and kurtosis were determined and yield deep insights into the turbulent flow field. The results showed that a geometry-induced secondary fluid motion occurred above and below the rib. The variations in the highly three-dimensional flow field close to the ribs and the geometry-induced secondary flow motion with the clearance-to-rib-height ratio determined the development of the wall-bounded flows and separating shear layers. Large recirculation regions on the bottom duct wall were prevented by the fluid exiting the gap below the detached ribs and separated shear layers pivoted upward in lateral direction. With decreasing wall proximity, lower mean vertical flow velocities above the ribs and the increasing upward fluid flow originating from the flow in the gap attenuated an intense interaction of the separated shear layers with the wall-bounded flow within the inter-rib spacing. Since turbulent structures originated in the shear layers, distributions of high-order statistic moments depend strongly on the shear layer development. Reynolds stresses and triple velocity correlations increased in the direction of the side walls near the rib due to the lateral flow motion, and their peak regions moved away from the wall with increasing clearance-to-rib-height ratios.

An improved ShuffleNetV2 method based on ensemble self-distillation for tomato leaf diseases recognition.

Timely and accurate recognition of tomato diseases is crucial for improving tomato yield. While large deep learning models can achieve high-precision disease recognition, these models often have a large number of parameters, making them difficult to deploy on edge devices. To address this issue, this study proposes an ensemble self-distillation method and applies it to the lightweight model ShuffleNetV2. Specifically, based on the architecture of ShuffleNetV2, multiple shallow models at different depths are constructed to establish a distillation framework. Based on the fused feature map that integrates the intermediate feature maps of ShuffleNetV2 and shallow models, a depthwise separable convolution layer is introduced to further extract more effective feature information. This method ensures that the intermediate features from each model are fully preserved to the ensemble model, thereby improving the overall performance of the ensemble model. The ensemble model, acting as the teacher, dynamically transfers knowledge to ShuffleNetV2 and the shallow models during training, significantly enhancing the performance of ShuffleNetV2 without changing the original structure. Experimental results show that the optimized ShuffleNetV2 achieves an accuracy of 95.08%, precision of 94.58%, recall of 94.55%, and an F1 score of 94.54% on the test set, surpassing large models such as VGG16 and ResNet18. Among lightweight models, it has the smallest parameter count and the highest recognition accuracy. The results demonstrate that the optimized ShuffleNetV2 is more suitable for deployment on edge devices for real-time tomato disease detection. Additionally, multiple shallow models achieve varying degrees of compression for ShuffleNetV2, providing flexibility for model deployment.

Performance of Permeable Bases

Critical aspects related to the design, construction, maintenance, and performance of permeable bases are examined, and a case study is presented to put such considerations in a more practical perspective. While the popularity of carefully designed subsurface drainage systems is justified by sound engineering considerations, and their performance has been generally satisfactory, their general use is not free of problems. Disappointing local experiences sometimes discourage agencies from specifying permeable bases. One such example is offered by the U.S. Route 50 test road near Athens, OH, for which the performance of the permeable base in the context of the overall subsurface drainage system is investigated. It is found that almost one-third of the outlets at the site could not be located, and that due to little or no maintenance many of the remainder had been silted in. Software DRIP 2.0 calculations affirm that outlet spacing is adequate, yet a considerably higher permeable base thickness is required, and that the permeability of one of the gradations used is too low. Moreover, the absence of a compatible separator layer between the permeable base and subgrade exposes the system to the danger of clogging.

A Pine Wilt Disease Detection Model Integrated with Mamba Model and Attention Mechanisms Using UAV Imagery

Pine wilt disease (PWD) is a highly destructive worldwide forest quarantine disease that has the potential to destroy entire pine forests in a relatively brief period, resulting in significant economic losses and environmental damage. Manual monitoring, biochemical detection and satellite remote sensing are frequently inadequate for the timely detection and control of pine wilt disease. This paper presents a fusion model, which integrates the Mamba model and the attention mechanism, for deployment on unmanned aerial vehicles (UAVs) to detect infected pine trees. The experimental dataset presented in this paper comprises images of pine trees captured by UAVs in mixed forests. The images were gathered primarily during the spring of 2023, spanning the months of February to May. The images were subjected to a preprocessing phase, during which they were transformed into the research dataset. The fusion model comprised three principal components. The initial component is the Mamba backbone network with State Space Model (SSM) at its core, which is capable of extracting pine wilt features with a high degree of efficacy. The second component is the attention network, which enables our fusion model to center on PWD features with greater efficacy. The optimal configuration was determined through an evaluation of various attention mechanism modules, including four attention modules. The third component, Path Aggregation Feature Pyramid Network (PAFPN), facilitates the fusion and refinement of data at varying scales, thereby enhancing the model’s capacity to detect multi-scale objects. Furthermore, the convolutional layers within the model have been replaced with depth separable convolutional layers (DSconv), which has the additional benefit of reducing the number of model parameters and improving the model’s detection speed. The final fusion model was validated on a test set, achieving an accuracy of 90.0%, a recall of 81.8%, a map of 86.5%, a parameter counts of 5.9 Mega, and a detection speed of 40.16 FPS. In comparison to Yolov8, the accuracy is enhanced by 7.1%, the recall by 5.4%, and the map by 3.1%. These outcomes demonstrate that our fusion model is appropriate for implementation on edge devices, such as UAVs, and is capable of effective detection of PWD.