Search Results (10)

Object detection in aerial images can provide great support in traffic planning, national defense reconnaissance, hydrographic surveys, infrastructure construction, and other fields. Objects in aerial images are characterized by small pixel–area ratios, dense arrangements between objects, and arbitrary inclination angles. In response to these characteristics and problems, we improved the feature extraction network Inception-ResNet using the Fast Architecture Search (FAS) module and proposed a one-stage anchor-free rotation object detector. The structure of the object detector is simple and only consists of convolution layers, which reduces the number of model parameters. At the same time, the label sampling strategy in the training process is optimized to resolve the problem of insufficient sampling. Finally, a decoupled object detection head is used to separate the bounding box regression task from the object classification task. The experimental results show that the proposed method achieves mean average precision (mAP) of 82.6%, 79.5%, and 89.1% on the DOTA1.0, DOTA1.5, and HRSC2016 datasets, respectively, and the detection speed reaches 24.4 FPS, which can meet the needs of real-time detection. Full article

The purpose of this research is to solve the complex long-distance and high-lift water supply engineering accident water hammer protection problem. Taking the Zhaojinzhuang water supply project as an example, based on the method of characteristics (MOC), the water hammer of the pumping station under the combined action of a water hammer relief valve, hydraulic-control butterfly valve, air vessel, air valve, and other water hammer protection measures is numerically simulated and calculated, and the effectiveness of the range method is analyzed, to ensure a waterproof hammer in pump stop accidents. The results show that the main factors affecting the effect of water hammer protection under the two-stage valve-closing parameters of the hydraulic-control butterfly valve are the fast-closing angle and the slow-closing time. The arrangement of the air vessel behind the pump can effectively increase the minimum water hammer pressure in the climbing section, and with the increase of the volume of the air vessel, the pump reverse speed and the maximum positive pressure increase slightly, but the overall water hammer protection effect is better. With the increase of the moment of inertia of the motor, the maximum positive pressure and minimum negative pressure of the pipeline still do not meet the requirements of the specification, and the modification cost is relatively large. The combination of the one-stage hydraulic-control butterfly valve, the air valve, the air vessel, and the water hammer relief valve can effectively reduce the volume of the air vessel. Under the optimal method, the maximum positive pressure head is 236.61 m, and the minimum negative pressure head is −3.18 m. Compared with the original method, the maximum positive pressure head is increased by 1.18%, the minimum negative pressure head is reduced by 95.78%, the maximum reverse speed of the pump is reduced by 100%, and the maximum reverse flow of the pump is reduced by 70.27%, meeting the requirements of water hammer protection. This is a safe and economical protection method. Full article

Background: Short implants have been used in the restoration of edentulous jaws in the past several years. However, some studies have suggested that short implants are less successful than standard implants. The aim of this study is to investigate the outcome of short implants placed in the posterior maxilla or mandible following one-stage or immediate-function protocols with a follow-up of 7 years (clinically) and 5 years (radiographically). Methods: This study included 127 patients rehabilitated with 217 implants measuring 7 mm and supporting 157 fixed prostheses in the posterior segments of both jaws. Final abutments were delivered at the surgery stage and were loaded after 4 months in 116 patients (199 implants). The primary outcome measure was implant survival measured through life tables. Secondary outcome measures were marginal bone loss and the incidence of biological and mechanical complications at the patient level and implant level (evaluated through descriptive statistics). Results: Twenty-four patients (18.9%) with 45 implants (20.7%) were lost to the follow-up. In total, 32 implants failed (14.8%) in 22 patients (17.3%), resulting in a cumulative survival rate at 7 years of 81.2% for 7 mm implants in the rehabilitation of the posterior regions of the maxilla and mandible. The average (standard deviation) marginal bone loss was 1.47 mm (0.99 mm) at 5 years. The incidence rate of biological complications was 12.6% and 10.6% at the patient and implant levels, respectively. The incidence rate of mechanical complications was 21.3% for patients and 16.1% for implants. A higher failure rate was registered in smokers and in implant arrangements with a sequence of three fixtures in proximity. Conclusions: Within the limitations of this study, it can be concluded that the placement of 7 mm long implants for the partial implant-supported rehabilitation of atrophic posterior jaws is possible in the long term, judging by the survival rate and stable average marginal bone loss. Nevertheless, strict case selection should be performed, especially in smokers and with implant arrangements that provide a minimum of one unit in inter-implant distance. Full article

Due to the sharp increase in household waste, its separate collection is essential in order to reduce the huge amount of household waste, since it is difficult to recycle trash without separate collection. However, since it is costly and time-consuming to separate trash manually, it is crucial to develop an automatic system for separate collection using deep learning and computer vision. In this paper, we propose two Anchor-free-based Recyclable Trash Detection Networks (ARTD-Net) which can recognize overlapped multiple wastes of different types efficiently by using edgeless modules: ARTD-Net1 and ARTD-Net2. The former is an anchor-free based one-stage deep learning model which consists of three modules: centralized feature extraction, multiscale feature extraction and prediction. The centralized feature extraction module in backbone architecture focuses on extracting features around the center of the input image to improve detection accuracy. The multiscale feature extraction module provides feature maps of different scales through bottom-up and top-down pathways. The prediction module improves classification accuracy of multiple objects based on edge weights adjustments for each instance. The latter is an anchor-free based multi-stage deep learning model which can efficiently finds each of waste regions by additionally exploiting region proposal network and RoIAlign. It sequentially performs classification and regression to improve accuracy. Therefore, ARTD-Net2 is more accurate than ARTD-Net1, while ARTD-Net1 is faster than ARTD-Net2. We shall show that our proposed ARTD-Net1 and ARTD-Net2 methods achieve competitive performance in mean average precision and F1 score compared to other deep learning models. The existing datasets have several problems that do not deal with the important class of wastes produced commonly in the real world, and they also do not consider the complex arrangement of multiple wastes with different types. Moreover, most of the existing datasets have an insufficient number of images with low resolution. We shall present a new recyclables dataset which is composed of a large number of high-resolution waste images with additional essential classes. We shall show that waste detection performance is improved by providing various images with the complex arrangement of overlapped multiple wastes with different types. Full article

The reverse osmosis method has developed extremely rapidly in recent years and has become the most competitive seawater desalination technology in the world, and it has been widely used in all aspects. Large-scale reverse osmosis desalination plants cannot provide fresh water resources in areas with insufficient water resources and limited space. Therefore, this paper proposes a research plan for a small seawater desalination device based on reverse osmosis, which is mainly suitable for handling emergencies, disaster relief, desert areas and outdoor activities and other needs for timely freshwater resources. It mainly includes pretreatment modules, a reaction infiltration module, a post-processing module and an energy supply module. Detailed design calculations are carried out for the small-scale reverse osmosis membrane system, including the selection and quantity and arrangement of membranes. Subsequently, the one-stage two-stage small-scale reverse osmosis membrane system was modeled, and its energy consumption was analyzed theoretically from the perspectives of specific energy consumption and energy utilization efficiency; the main influencing factors were clarified, and the optimal recovery rate for system operation was determined to be 20%–30%. Finally, an experimental prototype was built to conduct relevant experiments to determine the influence trend of pressure, temperature, concentration, and flow rate on the operating performance of the reverse osmosis system. Full article

Object detection is a challenging task in aerial images, where many objects have large aspect ratios and are densely arranged. Most anchor-based rotating detectors assign anchors for ground-truth objects by a fixed restriction of the rotation Intersection-over-Unit $(IoU)$ between anchors and objects, which directly follow horizontal detectors. Due to many directional objects with a large aspect ratio, the object-anchor $IoU$ is heavily influenced by the angle, which may cause few anchors assigned for some ground-truth objects. In this study, we propose an anchor selection method based on sample balance assigning anchors adaptively, which we name the Self-Adaptive Anchor Selection ($A^2$S-Det) method. For each ground-truth object, $A^2$S-Det selects a set of candidate anchors by horizontal $IoU$. Then, an adaptive threshold module is adopted on the set of candidate anchors, which calculates a boundary of these candidate anchors aiming to keep a balance between positive and negative anchors. In addition, we propose a coordinate regression of relative reference ($C{R}^3$) module to precisely regress the rotating bounding box. We test our method on a public aerial image dataset, and prove better performance than many other one-stage detectors and two-stage detectors, achieving the $mAP$ of 70.64. An efficiency anchor matching method helps the detector achieve better performance for objects with large aspect ratios. Full article

With great significance in military and civilian applications, the topic of detecting small and densely arranged objects in wide-scale remote sensing imagery is still challenging nowadays. To solve this problem, we propose a novel effectively optimized one-stage network (NEOON). As a fully convolutional network, NEOON consists of four parts: Feature extraction, feature fusion, feature enhancement, and multi-scale detection. To extract effective features, the first part has implemented bottom-up and top-down coherent processing by taking successive down-sampling and up-sampling operations in conjunction with residual modules. The second part consolidates high-level and low-level features by adopting concatenation operations with subsequent convolutional operations to explicitly yield strong feature representation and semantic information. The third part is implemented by constructing a receptive field enhancement (RFE) module and incorporating it into the fore part of the network where the information of small objects exists. The final part is achieved by four detectors with different sensitivities accessing the fused features, all four parallel, to enable the network to make full use of information of objects in different scales. Besides, the Focal Loss is set to enable the cross entropy for classification to solve the tough problem of class imbalance in one-stage methods. In addition, we introduce the Soft-NMS to preserve accurate bounding boxes in the post-processing stage especially for densely arranged objects. Note that the split and merge strategy and multi-scale training strategy are employed in training. Thorough experiments are performed on ACS datasets constructed by us and NWPU VHR-10 datasets to evaluate the performance of NEOON. Specifically, 4.77% and 5.50% improvements in mAP and recall, respectively, on the ACS dataset as compared to YOLOv3 powerfully prove that NEOON can effectually improve the detection accuracy of small objects in remote sensing imagery. In addition, extensive experiments and comprehensive evaluations on the NWPU VHR-10 dataset with 10 classes have illustrated the superiority of NEOON in the extraction of spatial information of high-resolution remote sensing images. Full article

This paper is concerned with the study of low-cost, low-power thermoacoustic electricity generators. Based on target electrical output power values of 50 and 100 W, three standing wave prototypes (of both one-stage and two-stage prototypes) integrating a commercial loudspeaker with different coupling arrangements are conceived. Each stage consists of a square-pore stack sandwiched between hot and ambient heat exchangers. The working gas is air at atmospheric pressure. The prototypes’ efficiency in converting heat to electrical power is simulated by the specialized Design Environment for Low-Amplitude ThermoAcoustic Engines (DeltaEC) design tool based on the linear theory of thermoacoustics. At a given operation frequency, the optimal impedance matching between the loudspeaker and the engine is realized by adjusting both the engine parameters (stack location, stack length, heat exchangers length, loudspeaker location) and loudspeaker parameters (load resistance and box volume). Computations reveal that the one-stage engine and two-stage engine with loudspeakers coupled in side-branch mode are able to meet the target output power values with comparable thermal-to-electric efficiency (4.6%). The two-stage engine with the loudspeaker coupled in push–pull mode is unable to reach the desired power output and is characterized by low conversion efficiencies (2%) due to the poor loudspeaker–engine acoustic impedance matching. Full article

The removal of dilute diethylether (DEE, concentration: 150 ppm) from an air stream (flow rate: 1.0 L min⁻¹) using non-thermal plasma combined with different cordierite-supported catalysts, including Mn, Fe, and mixed Mn-Fe oxides, was investigated. The experimental results showed that the decomposition of DEE occurred in a one-stage reactor without the positive synergy of plasma and supported catalysts, by which ca. 96% of DEE was removed at a specific input energy (SIE) of ca. 600 J L⁻¹, except when the mixed Mn-Fe/cordierite was used. Among the catalysts that were examined, Mn-Fe/cordierite, the catalyst that was the most efficient at decomposing ozone was found to negatively affect the decomposition of DEE in the one-stage reactor. However, when it was utilized as a catalyst in the post-plasma stage of a two-part hybrid reactor, in which Mn/cordierite was directly exposed to the plasma, the reactor performance in terms of DEE decomposition efficiency was improved by more than 10% at low values of SIE compared to the efficiency that was achieved without Mn-Fe/cordierite. The ozone that was formed during the plasma stage and its subsequent catalytic dissociation during the post-plasma stage to produce atomic oxygen therefore played important roles in the removal of DEE. Full article

This research presents a unified approach to power limits in power producing and power consuming systems, in particular those using renewable resources. As a benchmark system which generates or consumes power, a well-known standardized arrangement is considered, in which two different reservoirs are separated by an engine or a heat pump. Either of these units is located between a resource fluid (‘upper’ fluid 1) and the environmental fluid (‘lower’ fluid, 2). Power yield or power consumption is determined in terms of conductivities, reservoir temperatures and internal irreversibility coefficient, F. While bulk temperatures T_i of reservoirs’ are the only necessary state coordinates describing purely thermal units, in chemical (electrochemical) engines, heat pumps or separators it is necessary to use both temperatures and chemical potentials m_k. Methods of mathematical programming and dynamic optimization are applied to determine limits on power yield or power consumption in various energy systems, such as thermal engines, heat pumps, solar dryers, electrolysers, fuel cells, etc. Methodological similarities when treating power limits in engines, separators, and heat pumps are shown. Numerical approaches to multistage systems are based on methods of dynamic programming (DP) or on Pontryagin’s maximum principle. The first method searches for properties of optimal work and is limited to systems with low dimensionality of state vector, whereas the second investigates properties of differential (canonical) equations derived from the process Hamiltonian. A relatively unknown symmetry in behaviour of power producers (engines) and power consumers is enunciated in this paper. An approximate evaluation shows that, at least ¼ of power dissipated in the natural transfer process must be added to a separator or a heat pump in order to assure a required process rate. Applications focus on drying systems which, by nature, require a large amount of thermal or solar energy. We search for minimum power consumed in one-stage and multi-stage operation of fluidized drying. This multi-stage system is supported by heat pumps. We outline the related dynamic programming procedure, and also point out a link between the present irreversible approach and the classical problem of minimum reversible work driving the system. Full article