Small Edges Research Articles

MFMENet: multi-scale features mutual enhancement network for change detection in remote sensing images

ABSTRACT Change detection, an important task in remote sensing image analysis, has been extensively studied in recent years. However, change detection still faces problems such as difficulty in detecting small targets and incomplete edge detection. Furthermore, pseudo changes such as seasonal changes can also lead to many false detections. In response to these challenges, we propose the Multi-scale Features Mutual Enhancement Network (MFMENet), a simple yet efficient network. MFMENet maximizes feature utilization through mutual guidance and supplementation, enhancing the detection capabilities for small targets and edges. First, we use a lightweight feature extraction network to extract features, which mitigates the information loss caused by continuous downsampling of an overly deep network structure. Then, we design a Context Adaptive Interaction Module (CAIM) to realize the complementarity of feature information at different levels. This facilitates shallow features in gaining more semantic information and deep features in acquiring more texture information, thereby enhancing the model’s capability to capture more comprehensive edge features while effectively mitigating interference from pseudo changes. Finally, we introduce a Feature Aggregation Comparison Module (FACM), which uses a combination of aggregation and comparison methods to refine and enhance features. FACM can not only highlight the changed features but also retain more details, improving the model’s detection ability of small targets and edge details. The full utilization of features and effective mutual enhancement of information ensure the improvement of MFMENet’s performance in small target and edge detection. Extensive experiments on three publicly available datasets (LEVIR, DSIFN, and CDD) demonstrate that our approach achieves superior performance with fewer parameters compared to state-of-the-art methods in recent years. In comparison to these baseline methods, our proposed approach achieves improvements of 0.98%, 12.24%, and 2.03% in the IOU metric on the LEVIR, DSIFN, and CDD datasets, respectively, while utilizing only 1.1 M parameters.

Meta-contact analysis based on contact theory in three-dimensional discontinuous deformation analysis

Accurate contact analysis between nonspherical blocks, such as polyhedral blocks, is crucial for modeling jointed rock masses. The Contact Theory provides a rigorous foundation for contact detection. However, designing a robust and efficient algorithm that conforms to the original Contact Theory remains challenging. This study introduces the meta-contact concept for polyhedral blocks, which corresponds to the contact cover on the entrance block of the Contact Theory. Using the meta-contact concept, contact type identification, determination of contact points and plans, and contact force calculation can be formulated while retaining the advantages of Contact Theory in dealing with irregular blocks with small edges and large penetration.

Laser Weld Seam Curved Path Effect on 6063 Aluminum Alloy Strength and Temperature Distributions: COMSOL Numerical Simulation

To improve the welding performance of aluminum alloys, a thermal source model of an irregular weld seam was established. COMSOL software was used for numerical simulation of the weld seam geometry effect on the temperature and stress fields in laser welding, which results were experimentally validated. The results show that the ellipsoidal laser welding melted micropool exhibited quasi-steady-state temperature field characteristics. The temperature gradient and thermal stress showed an increase followed by a decline. The temperature fluctuation amplitude of the square-tooth-shaped weld seam exceeded that of the arc-toothshaped one. The temperature evolution of the broken line tooth-shaped weld seam showed a slightly increasing trend, except for the inflection point. The experimental average tensile strength of the weld seam was the highest, reaching about 210 MPa, i.e., roughly 85% of the base material (245 MPa), which coincided with the COMSOL-based temperature field simulation results. With increasing deformation amplitude and transition radius, the maximum tensile force, tensile strength, and elongation at fracture showed an increasing trend. However, the deformation amplitude should be below a certain limit because its increase elongates the welding path and reduces the distance between weld seams, resulting in serious heat accumulation. The tensile fracture morphology of the 6063-T6 base material was curved shear, with shallow toughness pits, small tearing edges at the edges, and small granular objects, indicating small plastic deformation during the fracture process. The tensile fracture of the welded part spanned the weld seam and the base material, and the fracture occurred along the tangent direction of the weld seam. The fracture surface was smooth, the tearing edges at the edge of the toughness pit shifted along the weld seam direction, forming many co-directional slip bands, with highly pronounced plastic deformation.

Automatic Marker Generation to Similarity Based Region Merging Algorithm using Edge Information

One of the most important tasks in computer vision is image segmentation. Several interactive strategies are utilized for picture segmentation because automated techniques are difficult for this kind of work. The outcome of interactive methods is mostly determined by user input. Obtaining excellent interactions for huge datasets is challenging. However, automated picture segmentation is starting to play a significant role in image analysis and computer vision. Effective and efficient segmentation outcomes are obtained using the interacting region merging method suggested by Maximal Similarity based region merging algorithm. Limitation of MSBRM is it requires some efforts on the part of users and is yet not a fully-automatic approach. Here, we suggest a completely fresh unsupervised image segmentation method that combines edge information with MSBRM. We propose an integrated framework to generate object markers for similarity based region algorithm using edge information. Long edges give rough distribution of objects in image. After retrieving edges using phase congruency, edge processing operations are employed to remove small edges and to group color similar long boundaries. Centroids of long boundaries are used as object markers to the MSBRM algorithm. The generation of object markers is done using edge segment grouping. These object markers guide the region merging process. The proposed method shows its effectiveness in segmenting natural real world color images.

Importance of high-resolution spatial data for the detection of winter wildlife responses to edges

Several wildlife species are thought to avoid edges of large habitat gaps, such as clear-cuts, but detailed evidence is rarely available for edges of smaller gaps. We compared the responses of nine wintering mammal species to forest edges in southern Quebec, Canada, using high-resolution spatial data from light detection and ranging (LiDAR) and low-resolution photo-interpretation. We defined edges of open areas as roads, lakes, rivers, or forest open areas. We geolocated mammal snow tracks along systematic transect lines between 2009 and 2018. We compared distances of snow tracks and reference points along transects to the nearest edge with linear models. LiDAR data revealed five species avoiding forest open area edges, whereas no avoidance was shown using photo-interpretation data. Weasels ( Mustela sp.) were the only species showing a positive association with forest open area edges using photo-interpreted data. No significant response was detected for river or lake edges. Four species were positively associated with road edges. We conclude that avoidance of small forest open area edges is widespread in our study area, but it can only be detected with high-resolution spatial data. Our results imply that edge effect can operate at a fine scale and using appropriate spatial resolution is crucial to detect such effects.

A CASE OF SELF-HARM USING A DEVICE TO STUN FARM ANIMALS

There are several types of devices used to stun cattle before slaughter or stunner (captive bolt pistol, cattlegun, stunner). This device consists of a heavy impact rod made of corrosion-resistant alloys such as stainless steel. It is held inside the case with the help of rubber washers and does not protrude from it. The rod is actuated by the trigger and is advanced by compressed air, a spring mechanism, or a blank cartridge charge that is ignited by the firing pin. After a shallow but strong blow to the animal's forehead, the spring tension causes the rod to roll back into the housing. In forensic medical practice, such devices usually appear as instruments of death in cases of suicidal or accidental injuries. Moreover, when reviewing the Ukrainian-language literature, the authors did not come across reports of deaths caused by stunners for stunning animals. Such situations are described in English-language professional periodicals, but they are extremely rare. Killings are generally isolated, apparently due to the fact that fatal wounds when fired from these devices occur from a distance of less than 10 cm, which follows from the design, namely the length of the rod. That is why the authors decided to report on a case from their expert work.
 A 58-year-old man was found unconscious in a beam and was taken to the hospital with a severe open penetrating brain injury, a hole-multifragmentary fracture of the left temporal bone, where he died 10 hours later. No additional data was provided by the police
 According to the autopsy data, a wound of a slightly irregular, rounded shape, with a diameter of about 0.9 cm, with small scalloped edges, which were unevenly beveled due to the "minus-tissue" defect towards the center, was found. Around the wound, there was an irregular, rectangular-shaped, brown-brown, dried-up scab, which was more pronounced on the lower edge. On the bow of the right temporal bone there is a hole-fragmentary fracture with the formation of a bone defect in the central part. The defect had a rounded shape, its edges are finely serrated, beveled like a trapezoid in the direction of the skull cavity, around the defect there is a funnel-shaped depression of the bone, which is formed by three incomplete fragments of an irregular shape. The total length of the wound canal was about 9 cm, and there were small bone fragments in the brain detritus along its entire length. No striking elements of the projectile or the projectile itself were found within the wound channel.
 According to the medical criminalistics investigation, the following were found: intermittent overlays of small dust-like dull particles of dark gray and black color, similar to soot shot, within the sedimentation belt, on the walls of the hole, and on the area of an irregular oval shape, located below and to the right of the above-mentioned damage. A diphenylamine test with dust-like particles gave a positive result. Traces of compounds of iron (characteristic form), lead, nickel and copper were also detected by the contact-diffusion method.
 Conclusions:
 
 Unfortunately, police very often do not provide important, and sometimes critically important, information, especially before the examination, so it is necessary to request the provision of criminal proceedings materials in any case with undetermined or unspecified circumstances.
 Forensic medical experts must use all available verified sources of information, within the framework allowed by legislation of Ukraine.
 In this case, the injuries were caused by a device with a firearm mechanism of operation, therefore, in unclear cases, it is necessary to use the term "device" and not "weapon" in the results.

Convergence analysis of virtual element method for the electric interface model on polygonal meshes with small edges

A conforming virtual element method is studied for approximating the solution of a pulsed electric interface model on polygonal meshes with small edges or faces, which traditional virtual and finite element methods cannot easily handle. One of the significant advantages of this virtual element method is to generate interface-conforming meshes efficiently exploiting polygonal meshes and hanging nodes. The virtual element method employs specific projection operators to establish optimal error estimates with reasonable assumptions regarding the solution's regularity. The voltage potential of the pulsed electric model through the physical media is approximated by using a fully discrete virtual element approach based on Crank-Nicolson temporal discretization. Numerical examples are used to illustrate the expected order of convergence. The efficiency and robustness of the proposed method are displayed on interface-independent/dependent background-fitted meshes with small edges and large magnitudes of discontinuities across the interface by these experiments.

Linear energy transfer (LET) distribution outside small radiotherapy field edges produced by 6 MV X-rays

In modern radiotherapy with photons, the absorbed dose outside the radiation field is generally investigated. But it is well known that the biological damage depends not only on the absorbed dose but also on LET. This work investigated the dose-average LET (LΔ,D) outside several small radiotherapy fields to provide information that can help for better evaluating the biological effect in organs at risk close to the tumour volume. The electron fluences produced in liquid water by a 6 MV X-rays Varian iX linac were calculated using the EGSnrc Monte Carlo code. With the electron spectra, LΔ,D calculations were made for eight open small square fields and the reference field at water depths of 0.15 cm, 1.35 cm, 9.85 cm and 19.85 cm and several off-axis distances. The variation of LΔ,D from the centre of the beam to 2 cm outside the field’s edge depends on the field size and water depth. Using radiobiological data reported in the literature for chromosomal aberrations as an endpoint for the induction of dicentrics determined in Human Lymphocytes, we estimated the maximum low-dose relative biological effectiveness, (RBEM) finding an increase of up to 100% from the centre of the beam to 2 cm from the field's edge.

An improved supervised and attention mechanism-based U-Net algorithm for retinal vessel segmentation

The segmentation results of retinal blood vessels are crucial for automatically diagnosing ophthalmic diseases such as diabetic retinopathy, hypertension, cardiovascular and cerebrovascular diseases. To improve the accuracy of vessel segmentation and better extract information about small vessels and edges, we introduce the U-Net algorithm with a supervised attention mechanism for retinal vessel segmentation. We achieve this by introducing a decoder fusion module (DFM) in the encoding part, effectively combining different convolutional blocks to extract features comprehensively. Additionally, in the decoding part of U-Net, we propose the context squeeze and excitation (CSE) decoding module to enhance important contextual feature information and the detection of tiny blood vessels. For the final output, we introduce the supervised fusion mechanism (SFM), which combines multiple branches from shallow to deep layers, effectively fusing multi-scale features and capturing information from different levels, fully integrating low-level and high-level features to improve segmentation performance. Our experimental results on the public datasets of DRIVE, STARE, and CHASED_B1 demonstrate the excellent performance of our proposed network.

A Virtual Element Method for the Elasticity Spectral Problem Allowing for Small Edges

A Virtual Element Method for the Elasticity Spectral Problem Allowing for Small Edges

VEM discretization allowing small edges for the reaction–convection–diffusion equation: source and spectral problems

In this paper we analyze a lowest order virtual element method for the classic load reaction–convection–diffusion problem and the convection–diffusion spectral problem, where the assumptions on the polygonal meshes allow to consider small edges for the polygons. Under well defined seminorms depending on a suitable stabilization for this geometrical approach, we derive the well posedness of the numerical scheme and error estimates for the load problem, whereas for the spectral problem we derive convergence and error estimates fo the eigenvalues and eigenfunctions. We report numerical tests to asses the performance of the small edges on our numerical method for both problems under consideration.

Habitat area and edges affect the length of trophic chains in a fragmented forest.

The food chain length represents how much energy reaches different trophic levels in food webs. Environmental changes derived from human activities have the potential to affect chain length. We explore how habitat area and edges affect chain length through: (1) a bottom-up effect of abundance ('pyramid hypothesis'); (2) the truncation of the highest trophic level ('trophic-rank hypothesis'); and (3) changes in species connectivity patterns ('connectivity hypothesis'). We built plant-leaf miner-parasitoid food webs in 19 remnants of a fragmented Chaco forest from central Argentina. On each remnant, we constructed food webs from different locations at the forest interior and edges. For each food web, we registered the abundance of species, the species richness of each trophic level, estimated the connectivity of their networks, and the average food chain length. We used structural equation models to evaluate the direct and indirect effects of habitat area and edge/interior location on food chain length mediated by species richness, abundance and connectivity. We found no direct effects of habitat area on chain length but chains were longer at forest edges than at their interior. The three mechanisms were supported by our results, although they showed different strengths. First, we found that the interior favours a bottom-up abundance effect from herbivores to parasitoids that positively affected chain length; second, we found that the forest area positively affects plant richness, which has a strong effect on the number of resources used by consumers, with a positive effect on chain length. Third, the remnant area and interior position favoured plant richness with a negative effect on the abundance of parasitoids, which had a positive effect on chain length. In general, the strongest effects on chain length were detected through changes in abundance rather than species richness although abundance was less affected by habitat fragmentation. We evaluated for the first time the effects of human-driven habitat fragmentation on the length of trophic chains in highly diverse plant-herbivore-parasitoid networks. Despite the loss of species, small habitat fragments and edges embedded in the agricultural matrix can support interaction networks, making them conservation targets in managed landscapes.

A virtual element method for the elasticity problem allowing small edges

A virtual element method for the elasticity problem allowing small edges

Mesh quality agglomeration algorithm for the virtual element method applied to discrete fracture networks

We propose a quality-based optimization strategy to reduce the total number of degrees of freedom associated with a discrete problem defined over a polygonal tessellation with the Virtual Element Method. The presented Quality Agglomeration algorithm relies only on the geometrical properties of the problem polygonal mesh, agglomerating groups of neighboring elements. We test this approach in the context of fractured porous media, in which the generation of a global conforming mesh on a Discrete Fracture Network leads to a considerable number of unknowns, due to the presence of highly complex geometries (e.g. thin triangles, large angles, small edges) and the significant size of the computational domains. We show the efficiency and the robustness of our approach, applied independently on each fracture for different network configurations, exploiting the flexibility of the Virtual Element Method in handling general polygonal elements.

Ecological factors related to seed germination and early seedling establishment inUlmus minorMill., an endangered riparian tree species

AbstractNatural populations of Ulmus minor Mill., an outstanding component of European riparian forests, are endangered due to Dutch elm disease, which has led to the development of several breeding programs throughout Europe. However, the reproductive ecology of this species is scarcely known even though sexual regeneration is an essential factor that must be understood before implementing elm stand restoration. In this paper, the main ecological factors affecting seed germination and early seedling establishment in U. minor are studied through a factorial design with five different factors: Light (levels: Full-sun, Shade), Initial water (levels: Initially-flooded, Initially-irrigated), Subsequent water (levels: Semi-submerged, Later-irrigated), Soil cover (levels: Bare-soil, Litter) and Substrate (levels: Silty-clay, Sandy-loam). Seed germination and seedling survival were monitored for 7 weeks. All factors except Initial water and Substrate turned out to be highly significant. Shade increased seed germination and seedling survival. Response to Initial water fluctuated, with an early higher percentage of germinated seeds in Initially-flooded but a higher seedling survival in Initially-irrigated. Subsequent water did not affect seedling establishment and survival, but Semi-submerged treatments increased seed germination. Bare-soilincreased seed germination too, whereas, in the medium term, Litter affected positively seedling survival. Sandy-loam substrates favoured seed germination, but Silty-claysubstrates favoured seedling survival. Together, these results suggest that the most favourable scenario for U. minor sexual regeneration may be a combination of a not-too-severe late-winter or early-spring flood followed by a wet spring and summer, probably with a higher establishment in heavy-textured substrates located in small gaps and forest edges where the overstorey provides some protection against strong sunlight.

Capturing Small Objects and Edges Information for Cross-Sensor and Cross-Region Land Cover Semantic Segmentation in Arid Areas

In the oasis area adjacent to the desert, there is more complex land cover information with rich details, multiscales of interest objects, and blur edge information, which poses some challenges to the semantic segmentation task in remote sensing images (RSIs). In traditional semantic segmentation methods, detailed spatial information is more likely lost in feature extraction stage and the global context information is more effectively integrated into segmentation results. To overcome these land cover semantic segmentation model, FPN_PSA_DLV3+ network, is proposed in an encoder–decoder manner capturing more fine edge and small objects information in RSIs. In the encoder stage, the improved atrous spatial pyramid pooling module extracts the multiscale features, especially small-scale feature details; feature pyramid network (FPN) module realizes better integration of detailed information and semantic information; and the spatial context information at both global and local levels is enhanced by introducing polarized self-attention (PSA) module. For the decoder stage, the FPN_PSA_DLV3+ network further adds a feature fusion branch to concatenate more low-level features. We select Landsat5/7/8 satellite RSIs from the areas of north and south of Xinjiang. Then, three self-annotated time-series datasets with more small objects and fine edges information are constructed by data augmentation. The experimental results show that the proposed method improves the segmentation performance of small targets and edges, and the classification performance increases from 81.55% to 83.10% <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">F</i> 1 score and from 72.65% to 74.82% mean intersection over union only using red–green–blue bands. Meanwhile, the FPN_PSA_DLV3+ network shows great generalization in cross region and cross sensor.

A Full-Scale Feature Fusion Siamese Network for Remote Sensing Change Detection

Change detection (CD) is an essential and challenging task in remote sensing image processing. Its performance relies heavily on the exploitation of spatial image information and the extraction of change semantic information. Although some deep feature-based methods have been successfully applied to change detection, most of them use plain encoders to extract the original image features. The plain encoders often have the below disadvantages: (i) the lack of semantic information leads to lower discrimination of shallow features, and (ii) the successive down-sampling leads to less accurate spatial localization of deep features. These problems affect the performance of the network in complex scenes and are particularly detrimental to the detection of small objects and object edges. In this paper, we propose a full-scale feature fusion siamese network (F3SNet), which on one hand enhances the spatial localization of deep features by densely connecting raw image features from shallow to deep layers, and on the other hand, complements the changing semantics of shallow features by densely connecting the concatenated feature maps from deep to shallow layers. In addition, a full-scale classifier is proposed for aggregating feature maps at different scales of the decoder. The full-scale classifier in nature is a variant of full-scale deep supervision, which generates prediction maps at all scales of the decoder and then combines them for the final classification. Experimental results show that our method significantly outperforms other state-of-the-art (SOTA) CD methods, and is particularly beneficial for detecting small objects and object edges. On the LEVIR-CD dataset, our method achieves an F1-score of 0.905 using only 0.966M number of parameters and 3.24 GFLOPs.

Novel method for welding gantry robot scheduling at shipyards

ABSTRACT Welding is the most critical operation in the shipbuilding process and has a significant influence on the production cost and quality of ships. Therefore, the welding operation must be optimised. This paper presents a real-world welding gantry robot scheduling problem at shipyards, in which three gantry robots function in parallel. Welding gantry robots cannot cross each other and should operate over a certain distance to avoid collisions. To minimise the makespan, the welding tasks given by line segments should be evenly distributed among the three gantry robots. The welding tasks assigned to each robot should be optimally sequenced to minimise the completion time, including the waiting time required to prevent collisions with neighbouring robots. In addition, long welding edges are split, and the split small length edges are assigned to the gantry robots. This paper proposes a mixed-integer linear programming model, three-stage solution approach, and variable neighbourhood search algorithm to solve this problem. Experimental tests conducted on 20 real problem instances revealed that the proposed approach can reduce the makespan by 14% on average when compared with the conventional method.

Virtual element methods for general linear elliptic interface problems on polygonal meshes with small edges

In this article, we discuss and analyze a conforming virtual element discretization with boundary stabilization term proposed in Brenner and Sung (2018) [30] (suitable for small edges that appeared in the mesh generation) to approximate general linear elliptic problems with discontinuous diffusivity coefficient across the interface. One of the critical features of the proposed virtual element method is to allow the generation of fitted meshes which are independent of the location of the interface. The fitted polygonal meshes naturally admit elements with small edges, and standard virtual element methods/finite element methods fail to handle meshes with arbitrary small edges. With the help of certain projection operators, nearly optimal error estimates are established under realistic assumptions and low regularity of the solution. Numerical experiments are conducted to show the proposed method's flexibility, robustness, and accuracy and validate the theoretical rate of convergence. • A conforming Virtual Element Method (VEM) is employed to approximate general linear elliptic problems with discontinuous diffusivity constant across the interface. The significant challenges in solving such a problem include the low global regularity of the solutions and suitable mesh generation near the interface. In the proposed discretization, it is easy to generate background-fitted meshes independent of the location of the interface. • In our mesh generation, we allow arbitrary small edges (naturally occurred in fitted mesh) and develop the stable virtual element scheme by using boundary stabilization proposed by S.C. Brenner et al. [Math. Models Methods Appl. Sci. 28 (7) (2018) 1291–1336] for establishing the nearly optimal error estimates. We stress that the standard stabilization may not be suitable when polygonal meshes involve arbitrary small edges. • Sophisticated analysis is carried out to show the well-posedness of the discrete problem and prove nearly optimal rate of convergence with minimum regularity on the continuous solution. • Several numerical experiments are conducted to compare the classical and proposed boundary stabilization behavior on different interfaces to show the flexibility and robustness of the method. Furthermore, the scheme's performance is illustrated with large magnitudes of discontinuities across the interface, which confirms the accuracy and theoretical rates of convergence.

A global direct search method for high-fidelity contact detection between arbitrarily shaped three-dimensional convex polyhedral blocks

High-fidelity contact detection between arbitrarily shaped convex polyhedral blocks is an important problem. The common plane (CP) and GJK algorithms are fast but still have some drawbacks in identifying accurate contact points and planes. The Direct Search (DS) method can ascertain the contact points and planes accurately but is limited by its locality assumption. To be specific, irregular blocks with small angles or small edges can cause false contacts, and large penetration will induce vanishing contacts, which will destroy the simulation using the DS method. In this study, a novel Global Direct Search (GDS) method is proposed to search for meta (vertex-face or cross edge-edge in 3D) contacts globally. Without locality assumption, the GDS method can accurately detect contacts between arbitrarily shaped convex polyhedral blocks without suffering from false and vanishing contacts. The new method inherits the global search idea of Contact Theory. Several case studies are used to verify the method and show that the GDS successfully avoids false and vanishing contacts, as well as demonstrate its accuracy, robustness, and improved efficiency.