Irregular Regions Research Articles

Improved EfficientNet for corn disease identification.

Corn is one of the world's essential crops, and the presence of corn diseases significantly affects both the yield and quality of corn. Accurate identification of corn diseases in real time is crucial to increasing crop yield and improving farmers' income. However, in real-world environments, the complexity of the background, irregularity of the disease region, large intraclass variation, and small interclass variation make it difficult for most convolutional neural network models to achieve disease recognition under such conditions. Additionally, the low accuracy of existing lightweight models forces farmers to compromise between accuracy and real-time. To address these challenges, we propose FCA-EfficientNet. Building upon EfficientNet, the fully-convolution-based coordinate attention module allows the network to acquire spatial information through convolutional structures. This enhances the network's ability to focus on disease regions while mitigating interference from complex backgrounds. Furthermore, the adaptive fusion module is employed to fuse image information from different scales, reducing interference from the background in disease recognition. Finally, through multiple experiments, we have determined the network structure that achieves optimal performance. Compared to other widely used deep learning models, this proposed model exhibits outstanding performance in terms of accuracy, precision, recall, and F1 score. Furthermore, the model has a parameter count of 3.44M and Flops of 339.74M, which is lower than most lightweight network models. We designed and implemented a corn disease recognition application and deployed the model on an Android device with an average recognition speed of 92.88ms, which meets the user's needs. Overall, our model can accurately identify corn diseases in realistic environments, contributing to timely and effective disease prevention and control.

Observation and verification of surface electrical explosion driven by radial‐distributed pulsed current in laboratory lightning strike test

AbstractThe laboratory lightning test is essential for assessing the effectiveness of lightning strike protection (LSP). Particularly, direct lightning strike damage can be performed with pulsed current injection into the specimen. This paper focuses on the dynamic process of arc plasma and shock wave behaviour in the vicinity of the ‘strike’ point. A rod‐plate discharge load is built for testing aluminium and coated plate under 40‐kA‐level pulsed current. The visualisation of the luminous discharge plasma and its flow field via high‐speed photography (from different angles) is meticulously designed and implemented, synchronised with electro‐physical diagnostics. The results indicate some new mechanisms for lightning strike damage, apart from the impulse heat loading from the thermal arc. The transient current injection through the arc root concentrates on a thin skin layer (skin‐depth effect), with the radial‐attenuated current density, driving asynchronously electrical explosions on the plate surface. The inhomogeneous Joule heating of the plate leads to outwardly propagating phase transition and shock wave along the conductive surface. In addition, the electro‐thermal instability is observed and regarded as the seed of irregular erosion region. Spectroscopic information reveals two different plasma states of main discharge arc channel and adjacent surface electrical explosion. The correspondence of the physical mechanism of electrical explosion and optical radiation is established. Microscopic images for different regions depict erosion characteristics and summarise influencing factors, further confirming the mechanism above. The research clarifies the role of skin‐depth effect in transaction arc erosion for electrode, complements the electrical explosion theory with unevenly distributed current and helps optimise strategies of LSP.

Invariant Attribute-Driven Binary Bi-Branch Classification of Hyperspectral and LiDAR Images

The fusion of hyperspectral and LiDAR images plays a crucial role in remote sensing by capturing spatial relationships and modeling semantic information for accurate classification and recognition. However, existing methods, such as Graph Convolutional Networks (GCNs), face challenges in constructing effective graph structures due to variations in local semantic information and limited receptiveness to large-scale contextual structures. To overcome these limitations, we propose an Invariant Attribute-driven Binary Bi-branch Classification (IABC) method, which is a unified network that combines a binary Convolutional Neural Network (CNN) and a GCN with invariant attributes. Our approach utilizes a joint detection framework that can simultaneously learn features from small-scale regular regions and large-scale irregular regions, resulting in an enhanced structural representation of HSI and LiDAR images in the spectral–spatial domain. This approach not only improves the accuracy of classification and recognition but also reduces storage requirements and enables real-time decision making, which is crucial for effectively processing large-scale remote sensing data. Extensive experiments demonstrate the superior performance of our proposed method in hyperspectral image analysis tasks. The combination of CNNs and GCNs allows for the accurate modeling of spatial relationships and effective construction of graph structures. Furthermore, the integration of binary quantization enhances computational efficiency, enabling the real-time processing of large-scale data. Therefore, our approach presents a promising opportunity for advancing remote sensing applications using deep learning techniques.

A Novel Adaptively Optimized PCNN Model for Hyperspectral Image Sharpening

Hyperspectral satellite imagery has developed rapidly over the last decade because of its high spectral resolution and strong material recognition capability. Nonetheless, the spatial resolution of available hyperspectral imagery is inferior, severely affecting the accuracy of ground object identification. In the paper, we propose an adaptively optimized pulse-coupled neural network (PCNN) model to sharpen the spatial resolution of the hyperspectral imagery to the scale of the multispectral imagery. Firstly, a SAM-CC strategy is designed to assign hyperspectral bands to the multispectral bands. Subsequently, an improved PCNN (IPCNN) is proposed, which considers the differences of the neighboring neurons. Furthermore, the Chameleon Swarm Optimization (CSA) optimization is adopted to generate the optimum fusion parameters for IPCNN. Hence, the injected spatial details are acquired in the irregular regions generated by the IPCNN. Extensive experiments are carried out to validate the superiority of the proposed model, which confirms that our method can realize hyperspectral imagery with high spatial resolution, yielding the best spatial details and spectral information among the state-of-the-art approaches. Several ablation studies further corroborate the efficiency of our method.

Supervised learning and meshless methods for two-dimensional fractional PDEs on irregular domains

Recently, several numerical methods have been developed for solving time-fractional differential equations not only on rectangular computational domains but also on convex and non-convex non-rectangular computational geometries. On the other hand, due to the existence of integrals in the definition of space-fractional operators, there are few numerical schemes for solving space-fractional differential equations on irregular regions. In this paper, we develop a novel numerical solution based on the machine learning technique and a generalized moving least squares approximation for two-dimensional fractional PDEs on irregular domains. The scheme is constructed on the monomials, and this is the strength of this technique. Moreover, it will be used to approximate the space derivatives on convex and non-convex non-rectangular computational domains. The numerical results are extended to solve the fractional Bloch–Torrey equation, fractional Gray–Scott equation, and fractional Fitzhugh–Nagumo equation.

Meshfree methods for the nonlinear variable-order fractional advection–diffusion equation

The nonlinear variable-order fractional advection–diffusion equation is applied to simulate complex engineering problem that exhibits time memory and space global dependence, for example, anomalous diffusion. In this paper, Kansa’s method is used to solve the nonlinear time–space variable-order fractional advection–diffusion equation, in which the variable order of space fractional derivative relies on both time and space while the variable order of the time fractional derivative is determined by space. Moreover, the convection coefficients, diffusion coefficients and source term are nonlinear functions that depend on exact solution. The finite difference method is contracted for the time fractional derivative and Wendland’s C6 compactly supported radial basis functions are used to substitute the space fractional derivative in the problem. The nonlinear terms are approximated by using an explicit difference scheme. The Gauss-Jacobi quadrature rule is used to approximate the weakly singular integral of space fractional derivative after radial basis function approximation. To illustrate the effectiveness and accuracy of the method, three numerical examples are solved, in which 2D experiments include the cases of regular and irregular regions. The results show that Kansa’s method has obvious advantages for variable-order fractional advection–diffusion equation.

Local processes with a global impact: unraveling the dynamics of gas evasion in a step-and-pool configuration

Abstract. Headwater streams are important sources of greenhouse gases to the atmosphere. The magnitude of gas emissions originating from such streams, however, is modulated by the characteristic microtopography of the riverbed, which might promote the spatial heterogeneity of turbulence and air entrainment. In particular, recent studies have revealed that step-and-pool configurations, usually found in close sequences along mountain streams, are important hotspots of gas evasion. Yet, the mechanisms that drive gas transfer at the water–air interface in a step-and-pool configuration are not fully understood. Here, we numerically simulated the hydrodynamics of an artificial step-and-pool configuration to evaluate the contribution of turbulence and air entrainment to the total gas evasion induced by the falling jet. The simulation was validated using observed hydraulic features (stage, velocity) and was then utilized to determine the patterns of energy dissipation, turbulence-induced gas exchange and bubble-mediated transport. The results show that gas evasion is led by bubble entrainment and is mostly concentrated in a small and irregular region of a few square decimeters near the cascade, where the local gas transfer velocity (k) peaks at 500 m d−1. The enhanced spatial heterogeneity of k in the pool does not allow one to define a priori the region of the domain where the outgassing takes place and makes the value of the spatial mean of k inevitably scale-dependent. Accordingly, we propose that the average mass transfer velocity should be used with caution to describe the outgassing in spatially heterogeneous flow fields, such as those encountered in step-and-pool rivers.

The Neural Correlates of the Social Perception Dysfunction in Schizophrenia: An fMRI Study.

Patients with schizophrenia show deficits in facial emotion recognition and emotional intensity assessment, and also exhibit structural and functional irregularities in specific brain regions. In this study, we aimed to examine differences in active brain regions involved in processing the Emotion Intensity Recognition Task (EIRT), which can serve as an indicator of emotion recognition and ability to perceive intensity, between patients with schizophrenia and healthy controls (HCs). The purpose of this study was to investigate dysfunctional brain regions and investigate the role of the amygdala in social cognition deficits in patients with schizophrenia by focusing on alterations in amygdala activity linked to facial emotion recognition. Twenty-two patients who met a diagnostic criteria for schizophrenia according to DSM-IV and 27 HCs participated in an MRI scan while completing the EIRT. Behavioral and MRI data were collected and analyzed. Behavioral results showed that patients with schizophrenia made significantly more errors in recognizing surprise, happiness, sadness, fear, and neutral expressions, and patients with schizophrenia exhibited significantly slower response times in recognizing happy facial expressions. Imaging results showed that schizophrenia patients found hypoactivation in several inferior parietal and temporal regions, in the cerebrum and anterior cingulate; and decreased amygdala activation in individuals with schizophrenia was associated with impaired recognition of fear in facial expressions. Facial emotion processing deficits are emotion-specific (surprise, happiness, sadness, fear, and neutral expressions) in schizophrenia. Hypoactivation in several inferior parietal and temporal regions, in the cerebrum and anterior cingulate, was thought to contribute to symptom formation in schizophrenia. Reduction in amygdala activation in schizophrenia patients was associated with impairment of the fear-emotional process.

Study of the stability of a meshless generalized finite difference scheme applied to the wave equation

When designing and implementing numerical schemes, it is imperative to consider the stability of the applied methods. Prior research has presented different results for the stability of generalized finite-difference methods applied to advection and diffusion equations. In recent years, research has explored a generalized finite-difference approach to the advection-diffusion equation solved on non-rectangular and highly irregular regions using convex, logically rectangular grids. This paper presents a study on the stability of generalized finite difference schemes applied to the numerical solution of the wave equation, solved on clouds of points for highly irregular domains. The stability analysis presented in this work provides significant insights into the proper discretizations needed to obtain stable and satisfactory results. The proposed explicit scheme is conditionally stable, while the implicit scheme is unconditionally stable. Notably, the stability analyses presented in this paper apply to any scheme which is at least second order in space, not just the proposed approach. The proposed scheme offers effective means of numerically solving the wave equation, particularly for highly irregular domains. By demonstrating the stability of the scheme, this study provides a foundation for further research in this area.

First Report of Colletotrichum siamense Causing Leaf Spot on Jasminum sambac in China.

Jasmine (Jasminum sambac (L.) Aiton) is cultivated as a commercial floricultural crop in many countries around the world (Gao et al., 2020). From June to August 2020, leaf spots on jasmine were observed on a jasmine plantation in Hengzhou of Guangxi province. Over 40% of the plants in 6 ha fields were infected. This disease was prevalent in jasmine production area of China (Chen et al., 2012; Du et al., 2020). Symptoms began as chlorotic regions (from 5 to 10 mm in diameter) with light brown necrotic centers, which gradually expanded to the entire leaf. Eventually, the disease leaded to defoliation and dieback. The edges of the affected parts from diseased leaves were cut into pieces (3 mm2). Pieces were treated with 75% ethanol for 10 s, soaked in 2% NaClO solution for 1 min, washed three times with sterile water, and then incubated on potato dextrose agar (PDA) plates at 28℃ for 5 days in the dark. Fungal cultures that showed similar morphological characteristics were isolated, and three representative isolates (HL6-1 to HL6-3) were purified following Mo et al. (2018). The cultures on PDA changed from white to dark grey after 7 days and produced conidiomata after 14 days. Conidia were hyaline, one-celled, guttulate, cylindrical, of 12.07 to 18.09 × 4.04 to 8.05 μm, 13.17 to 16.35 × 4.22 to 6.13 μm and 10.11 to 22.17 × 3.65 to 8.1 μm for HL6-1, HL6-2 and HL6-3, respectively. Gray-brown or dark brown appressoria formed from conidia were subglobose or elliptical. Conidial appressoria and mycelial appressoria were 5.53 to 13.96 × 3.58 to 13.95 μm and 4.24 to 14.01 × 2.4 to 10.86 μm. Genomic DNA was extracted from three isolates and the partial internal transcribed spacer (ITS) regions, intergenic region of apn2 and MAT1-2-1 (ApMAT), and fragments of actin (ACT), glyceraldehydes-3-phosphate dehydrogenase (GAPDH), chitin synthase (CHS-1), and β-tubulin (TUB2) genes were amplified, sequenced and submitted to GenBank (ITS, ON115173 to ON115175; ApMat, ON156517 to ON156519; ACT, ON146469 to ON146471; GAPDH, ON156502 to ON156504; CHS-1, ON156507 to ON156509; TUB-2, ON156512 to ON156514). Phylogenetic tree was constructed with MrBayes v. 3.2.6 and MEGA v. 10.1.5 based on the concatenation of multiple sequences. Three isolates were grouped with strain C. siamense ICMP 18578. Results indicated three isolates were identified as Colletotrichum siamense Prihastuti, L. Cai & K.D. Hyde. To confirm the pathogenicity of the three isolates, four sets (five plants per set) of 160 healthy leaves of 2-year-old plants (J. sambac, eight leaves per plant) were slightly scratched with a sterilized toothpick at each of eight locations. Conidial suspension (1×106 conidia/mL) in 0.1% Tween 20 were inoculated onto each wounded spot of three sets as the treatment groups, while wounded leaves treated with sterile water as the control. All plants were covered with plastic bags and cultivated in phytotron (12 h/12 h light/dark, 28°C). After 7 days, irregular chlorotic regions with brown lesions were observed on inoculated leaves while no symptoms on controls. The same fungi were reisolated from inoculated leaves and confirmed by morphological and molecular identification, fulfilling Koch's postulates. Colletotrichum siamense has been associated with leaf anthracnose of J. sambac in Vietnam (Wikee et al., 2011) and J. mesnyi in China (Zhang et al., 2019). To our knowledge, this is the first report of C. siamense causing jasmine anthracnose in China, which provides a reference for the management of this disease.

A modeling study of transport through composite membrane with support pore location distributing randomly based on 2D Voronoi tessellations

Composite membrane consisting of selective layer and support is the mainstream membrane configuration. Support has drawn increasing attention in recent years. In this work, transport through composite membrane with support pore location distributing randomly was studied based on 2D Voronoi tessellations and the corresponding membrane performance calculation method was developed. For the calculation, firstly, irregular Voronoi regions are transformed into regular polygons with fixed number of edges meanwhile maintaining the region area. Then, the membrane performance can be obtained by combining the computation results of regular support surface pore spatial distribution and random support surface pore spatial distribution formulas. Poisson Voronoi diagrams, one of the most representative Voronoi diagrams, were mainly used to describe the support surface pore spatial distribution. The results indicate the importance of the number of support surface pores, which determines the average area of Voronoi regions. Support surface pores should be as many as possible. Another important conclusion is that a certain randomness of support surface pore structure (size and location) may not exert an influence on the composite membrane performance. That is, when calculating the composite membrane performance, the support surface pore spatial distribution and surface pore size distribution can be considered as being regular.

Microstructural Evolution and Mechanical Behaviour of Near-Eutectic High Entropy Alloy

In the current work, the thermal stability of the eutectic regions and precipitation and mechanical behaviour of near eutectic high-entropy alloy is studied. The alloy solidified as FCC and underwent sequential ordering to L12 with order–disorder transition temperature between 700°C and 1000°C. Cooling rate, variation in the concentration boundary layer, thermal gradient and interfacial energy were attributed to define the eutectic morphology (L12 and B2). Thermal exposure resulted in homogeneous precipitation of acicular Al-Ni-rich B2 phase within the proeutectic and irregular eutectic regions. The preferential B2 precipitation within the irregular eutectic regions was attributed to the lower stability and higher lattice strain associated with these regions. The FCC: B2 phase fraction after heat treatment was similar to that of the eutectic alloy. Furthermore, lamellar degradation, globularization and Ostwald ripening of the B2 phase were observed after heat treatment. Lamellar degradation occurred via mechanisms like cylinderization, edge spheroidization, termination migration and boundary splitting. The experimental observations on the phase evolution, stability and order–disorder transformation matched well with the CALPHAD predictions except for the σ phase formation. B2 precipitation during heat treatment enhanced the mechanical behaviour of the alloy. However, the work hardening rate was superior for as-cast alloy.

Irregularities Observed at the Edge of a Mid‐Latitude Ionospheric Depletion Following a Geomagnetic Storm

AbstractThis manuscript presents the analysis of data from multiple ground‐ and space‐based sensors in the North American region before, during, and after the 12 October 2021 geomagnetic storm. Total electron content (TEC) and electron density data show the formation and equatorward propagation of a mid‐latitude trough at ∼50°N followed by the appearance of a wider depletion region (∼15° in latitude) at lower latitudes. During the recovery phase on the 13th, the equatorward edge of this depletion region settled at around 30° latitude and exhibited a steep density gradient. By the 14th, this sharp boundary had disappeared. Near this edge on the 13th, small‐scale irregularities formed. The impact of these was observed within Global Positioning System data as elevated rate of TEC index (ROTI) and presented as strong 35 MHz scintillations of cosmic radio sources as well as spread‐F within ionograms from multiple digisonde systems. GPS and 35‐MHz data demonstrated that the irregularity region was narrowly confined (≲5° wide) near the edge of the depletion region. The 35‐MHz scintillation data also showed that the irregularities were moving relatively slowly at ∼7 m s−1, likely toward the southeast. Density and velocity measurements demonstrated that the conditions near the depletion boundary were highly favorable to the gradient drift instability (GDI) with the one‐dimensional growth rate estimated to be ∼0.01 s−1. Since these conditions persisted for many hours, this growth rate was more than sufficient for the GDI to be considered the primary driver of irregularity formation in this case.

Generalization of the shortest path approach for superpixel segmentation of omnidirectional images

With the growing use of image capture devices using wide angles and the need for fast and accurate image analysis in computer vision, there is a demand for dedicated under-representation methods. Most decomposition methods segment an image into a small number of irregular homogeneous regions, called superpixels. Nevertheless, these approaches are generally designed to process natural 2D planar images, i.e., captured with a 90o angle view without distortion. In this work, we present SphSPS, a new general decomposition method (for Spherical Shortest Path-based Superpixels)11Available code at: https://github.com/rgiraud/sphsps, that is dedicated to wide 360o omnidirectional or spherical images. The produced superpixels respect both the geometry of the 3D spherical acquisition space, and the boundaries of image objects. To fastly extract relevant clustering features, we generalize the shortest path approach between a pixel and a superpixel center. We demonstrate that considering the geometry of the acquisition space to compute the shortest path enables to jointly improve the segmentation accuracy and the shape regularity of superpixels. To evaluate this regularity property, we propose a generalization of a standard 2D regularity metric to the spherical space, addressing the limitations of the only existing spherical compactness measure. Finally, SphSPS is validated on reference 360o images from the PSD (Panorama Segmentation Dataset) and also on synthetic road omnidirectional images. Our method significantly outperforms both planar and spherical state-of-the-art approaches in terms of segmentation accuracy, robustness to noise and regularity, providing a very interesting tool for superpixel-based applications on 360o images.

Resting-state functional MRI in treatment-resistant schizophrenia.

Abnormalities in brain regions involved in the pathophysiology of schizophrenia (SCZ) may present insight into individual clinical symptoms. Specifically, functional connectivity irregularities may provide potential biomarkers for treatment response or treatment resistance, as such changes can occur before any structural changes are visible. We reviewed resting-state functional magnetic resonance imaging (rs-fMRI) findings from the last decade to provide an overview of the current knowledge on brain functional connectivity abnormalities and their associations to symptoms in treatment-resistant schizophrenia (TRS) and ultra-treatment-resistant schizophrenia (UTRS) and to look for support for the dysconnection hypothesis. PubMed database was searched for articles published in the last 10 years applying rs-fMRI in TRS patients, i.e., who had not responded to at least two adequate treatment trials with different antipsychotic drugs. Eighteen articles were selected for this review involving 648 participants (TRS and control cohorts). The studies showed frontal hypoconnectivity before the initiation of treatment with CLZ or riluzole, an increase in frontal connectivity after riluzole treatment, fronto-temporal hypoconnectivity that may be specific for non-responders, widespread abnormal connectivity during mixed treatments, and ECT-induced effects on the limbic system. Probably due to the heterogeneity in the patient cohorts concerning antipsychotic treatment and other clinical variables (e.g., treatment response, lifetime antipsychotic drug exposure, duration of illness, treatment adherence), widespread abnormalities in connectivity were noted. However, irregularities in frontal brain regions, especially in the prefrontal cortex, were noted which are consistent with previous SCZ literature and the dysconnectivity hypothesis. There were major limitations, as most studies did not differentiate between TRS and UTRS (i.e., CLZ-resistant schizophrenia) and investigated heterogeneous cohorts treated with mixed treatments (with or without CLZ). This is critical as in different subtypes of the disorder an interplay between dopaminergic and glutamatergic pathways involving frontal, striatal, and hippocampal brain regions in separate ways is likely. Better definitions of TRS and UTRS are necessary in future longitudinal studies to correctly differentiate brain regions underlying the pathophysiology of SCZ, which could serve as potential functional biomarkers for treatment resistance.

A meshless finite difference scheme applied to the numerical solution of wave equation in highly irregular space regions

A meshless finite difference scheme applied to the numerical solution of wave equation in highly irregular space regions

A wavelet method for large-deflection bending of irregular plates

A very rare sixth-order multiresolution method based on wavelet theory for solving irregular domain problems is proposed and applied to the large-deflection bending of irregular plates. High accuracy approximation of multiple integrals of functions defined in the irregular domain is realized by suggesting a new method of boundary extension of functions, which is associated with a boundary node removal technique to avoid singular matrices in the integration process. Each derivative of the unknown function is defined as a new function in the solution of the governing differential equations of the problem. The original equations containing derivatives of different orders can then be transformed into a system of algebraic equations with only discrete nodal values of the highest-order derivative, based on integral relations between these functions and excellent accuracy features on the wavelet approximation of multiple integrals. Various boundary conditions for irregular domains can be automatically and almost exactly included in the integration operation. The proposed method is applied to the solution of large-deflection bending of various irregular plates. The displacement and stress results demonstrate that the accuracy of the proposed method is still as high as the rare sixth order even when the problems are strongly nonlinear and defined in irregular regions, showing a much better accuracy and efficiency than those of classical methods such as the finite element method.

Large Data Oriented to Image Information Fusion Spark and Improved Fruit Fly Optimization Based on the Density Clustering Algorithm

The density-based applied spatial clustering algorithm is an algorithm based on high-density interconnected regions, which discovers class clusters of arbitrary shapes in noisy data sets and is widely used. However, it suffers from slow computation speed due to large-scale disk I/O and clustering bias due to uneven density class clusters and poor parameter search ability. To address these problems, a parallel density clustering algorithm based on an improved fruit fly optimization algorithm and Spark memory iteration is proposed. The proposed algorithm first divides the data grid using an irregular dynamic density region partitioning strategy. Then, a hybrid fruit fly particle swarm algorithm based on a genetic optimization mechanism is proposed to achieve dynamic optimization seeking for parameters in local clustering to improve the clustering effect of local clustering. Finally, the local merging of samples in irregularly bounded grid cells under each partition is achieved by designing a custom clustering merging strategy. The experiments show that the improved algorithm is generally applicable to the clustering of different shape class clusters and larger scale data and has obvious improvement in accuracy and parallel efficiency.

Arbitrary-Shaped Text Detection with B-Spline Curve Network

Text regions in natural scenes have complex and variable shapes. Directly using contour coordinates to describe text regions will make the modeling inadequate and lead to low accuracy of text detection. To address the problem of irregular text regions in natural scenes, we propose an arbitrary-shaped text detection model based on Deformable DETR called BSNet. The model differs from the traditional method of directly predicting contour points by using B-Spline curve to make the text contour more accurate and reduces the number of predicted parameters simultaneously. The proposed model eliminates manually designed components and dramatically simplifies the design. The proposed model achieves F-measure of 86.8% and 87.6% on CTW1500 and Total-Text, demonstrating the model's effectiveness.

Multi-Source Data Fusion and Hydrodynamics for Urban Waterlogging Risk Identification.

The complex formation mechanism and numerous influencing factors of urban waterlogging disasters make the identification of their risk an essential matter. This paper proposes a framework for identifying urban waterlogging risk that combines multi-source data fusion with hydrodynamics (MDF-H). The framework consists of a source data layer, a model parameter layer, and a calculation layer. Using multi-source data fusion technology, we processed urban meteorological information, geographic information, and municipal engineering information in a unified computation-oriented manner to form a deep fusion of a globalized multi-data layer. In conjunction with the hydrological analysis results, the irregular sub-catchment regions are divided and utilized as calculating containers for the localized runoff yield and flow concentration. Four categories of source data, meteorological data, topographic data, urban underlying surface data, and municipal and traffic data, with a total of 12 factors, are considered the model input variables to define a real-time and comprehensive runoff coefficient. The computational layer consists of three calculating levels: total study area, sub-catchment, and grid. The surface runoff inter-regional connectivity is realized at all levels of the urban road network when combined with hydrodynamic theory. A two-level drainage capacity assessment model is proposed based on the drainage pipe volume density. The final result is the extent and depth of waterlogging in the study area, and a real-time waterlogging distribution map is formed. It demonstrates a mathematical study and an effective simulation of the horizontal transition of rainfall into the surface runoff in a large-scale urban area. The proposed method was validated by the sudden rainstorm event in Futian District, Shenzhen, on 11 April 2019. The average accuracy for identifying waterlogging depth was greater than 95%. The MDF-H framework has the advantages of precise prediction, rapid calculation speed, and wide applicability to large-scale regions.