生成式图像编辑研究综述

图像编辑是指保持图像非编辑区域内容不变,编辑区域根据输入条件改变的过程。生成式图像编辑利用深度学习模型,以编辑条件作为指导,通过生成图像的方式来修改图像的内容,包括但不限于修改图像中的对象、风格、颜色、纹理等特征。生成式图像编辑是计算机视觉研究中不可或缺的一环。生成式图像编辑具有重要的理论和实际应用价值,引起了学术界和工业界的广泛关注,然而目前对相关研究的综述工作相对较少。本文针对极具代表性的生成式图像编辑方法进行了综述,首先根据编辑方法不同,将编辑任务分为基于属性、基于文本、基于线稿和基于实物范例引导的图像编辑,结合损失函数、类型以及特点等对不同方法进行分析总结;同时进一步探讨了各方法对生成质量的影响,并详尽阐述了数据集和评价指标;最后,指出该领域未来面临的挑战和可能的研究方向。

一种新型构型卫星模态分析及试验研究

该文中的卫星采用新型框架面板式构型,该构型卫星与运载采用新型四点连接的方式,而不是传统的包带连接方式。对该文中的卫星进行模态分析及试验,采用不同的安装基础进行多次模态试验,找出其中的差异,并进行相关比对分析。根据分析及试验获得框架面板式卫星的模态特性,判读卫星的模态特性是否满足运载要求,同时根据试验结果完善相关航天试验规范。

Dynamic characteristics of cold rolling mill and strip based on flatness and thickness control in rolling process

The dynamic model of cold rolling mill based on strip flatness and thickness integrated control was proposed,containing the following sub-models:the rolling process model,the dynamic model of rolls along axial direction,and the compensation model.Based on the rule of volume flow rate,the dynamic rolling process model was built.The work roll and backup roll were taken as elastic continuous bodies,the effect of shear and moment of inertia were taken into consideration,and then the dynamic model of rolls was built.The two models were coupled together,and the dynamic model of rolling mill was built.In the dynamic model,the thermal expansion of the rolls,the wear of the rolls and other related parameters can not be considered.In order to compensate the dynamic model,the coupled static model of rolls and strip was applied.Then,according to the inner relationship of these models,the dynamic model and the compensation model were coupled,and the dynamic model of rolling mill based on the strip flatness and thickness integrated control was built.The dynamic simulation of the rolling process was made,and the dynamic thickness and the dynamic flatness information were obtained.This model not only provides a theory basis for the virtual rolling,but also provides a platform for the application of advanced control theory.

Time series prediction of mining subsidence based on a SVM

In order to study dynamic laws of surface movements over coal mines due to mining activities,a dynamic prediction model of surface movements was established,based on the theory of support vector machines(SVM) and times-series analysis.An engineering application was used to verify the correctness of the model.Measurements from observation stations were analyzed and processed to obtain equal-time interval surface movement data and subjected to tests of stationary,zero means and normality.Then the data were used to train the SVM model.A time series model was established to predict mining subsidence by rational choices of embedding dimensions and SVM parameters.MAPE and WIA were used as indicators to evaluate the accuracy of the model and for generalization performance.In the end,the model was used to predict future surface movements.Data from observation stations in Huaibei coal mining area were used as an example.The results show that the maximum absolute error of subsidence is 9 mm,the maximum relative error 1.5%,the maximum absolute error of displacement 7 mm and the maximum relative error 1.8%.The accuracy and reliability of the model meet the requirements of on-site engineering.The results of the study provide a new approach to investigate the dynamics of surface movements.

Node scattering manipulation based on trajectory model in wireless sensor network

To deploy sensor nodes over the area of interest,a scheme,named node scattering manipulation,was proposed.It adopted the following method:during node scattering,the initial states of every node,including the velocity and direction,were manipulated so that it would land in a region with a certain probability;every sensor was relocated in order to improve the coverage and connectivity.Simultaneously,to easily analyze the process of scattering sensors,a trajectory model was also proposed.Integrating node scattering manipulation with trajectory model,the node deployment in wireless sensor network was thoroughly renovated,that is,this scheme can scatter sensors.In practice,the scheme was operable compared with the previous achievements.The simulation results demonstrate the superiority and feasibility of the scheme,and also show that the energy consumption for sensors relocation is reduced.

Using Successive Multi-Goals Method in the Individual-Based Model to Simulate Crowd Behavior during Evacuation Processes

The fact of proportional population growth in many countries drags the attention of researchers in the field of crowd dynamics to the need for developing reliable models to predict the behavior of human crowds in emergency situations such as evacuation processes. Computer based models that simulate human crowd dynamics prove to offer the optimum way to predict the crowd realistic behavior especially in emergency situations. This paper presents a vital extension of my previous work in which an individual-based model to simulate the behavior of human crowd was developed using the artificial potential fields to describe the interaction forces between each crowd member and the environment on one side and amongst the crowd members on the other side to add realistic flavor to the predicted crowd behavior. In this paper, the successive multi-goals （SMG） method, which is a new method to represent the environment in which the crowd moves, is developed. Rather than using the traditional static potential field, the successive multi-goals method uses a dynamic potential field which is vital to solve the reactive problem that is considered as a drawback of the model when simulating the human crowd behavior during evacuation of buildings whose structures are complex such as bottlenecks and narrow corridors. Numerical results that match the real behavior of human individuals in emergency situations prove the efficiency of the new method to solve the problem on an individual basis as well as its applicability.

A damage-based constitutive model for rock under impacting load

Using the Splitting Hopkinson Pressure Bar （SHPB） experimental system, investigations were made into the dynamic mechanical performances of underground soft rocks. The experiments proved that the measured stress-strain curves display the characteristics of plastic deformation. By making use of a revised overstress constitutive formula for the stress model and by taking into account that the strain rate and strain are a function of I - E（t）/Eo, a revised overstress constitutive formula for the stress model was simplified by applying dimensional analysis and consequently, a simplified overstress formula was obtained for the stress model. Then, by taking into consideration the effects of damage under a dynamic load on the dynamic loading strength of the rock, the continuous damage theory and the statistical strength theory were introduced into the development of the simplified overstress constitutive formula for the stress model. Hence, a damage-based constitutive formula for an overstress model, which can be appropriately applied to the analysis of full dynamic stress-strain curves, was developed. By using the simplified damage-based constitutive formula for an overstress model, the actually measured curves are fitted, indicating that the fitting curves and those actually measured are in good agreement.

The Construction of a PPT-Based College English Class in Multimodal Background

This paper starts from other scholars＇ studies of multimodality and multimodal discourse and discusses PPT--PowerPoint, as a pragmatic and effective means in teaching English to non-English major students. By using multiple modalities of language, pictures, sound; and videos together with the facilities of slide-projector and computer, the multimodal background provides top teaching auxiliary for teachers employing PPT to achieve most desirable effects in real class. This paper discusses the feasibility and effectiveness of PPT as teaching materials and PPT-based class construction in order to show examples to other English teachers how to conduct English teaching and promote students to take more interests in English learning. The writer gets to the conclusion that PPT, as a multimodal measure in teaching, can help the students to learn English more effectively and teachers to achieve better quality in lectures.

Optimizing forestry industrial structure based on the framework of rationalization,advancement and ecologicalization

Under the target of the sustainable development,optimization of forestry industrial structure contains rationalization and advancement of forestry industrial structure in the traditional sense,as well as ecologicalization of forestry industrial structure.The connotation of optimization of forestry industrial structure is analyzed at first,and then the paper chooses the typical evaluation indexes from the three sides of rationalization,advancement and ecologicalization to evaluate the current situation of the regional forestry industrial structure.With the aid of the theory of fuzzy mathematics and information entropy,the paper establishes a fuzzy evaluation model which is based on information entropy for optimization of forestry industrial structure,and the evaluation model is applied in the study for the current situation of forestry industrial structure from 2000 to 2006 in Hubei province.The results show that the grade of forestry industrial structure is bad basically during the research,of which the overall level of forestry industrial structure of Hubei province is poor,the degree of utilizing forest resource is lower,and ecological environment in the forest region has been damaged to some extent.On this basis,the author presents the countermeasure to promote the optimization of forestry industrial structure.

Development of Sensor Oriented Dynamic Rotor Model as the Basis of Machine Diagnostics

Today, digital signal processing based on new microprocessor technology allows for successful machine protection and is acceptably cost-effective. Furthermore, by means of vibration-diagnostics monitoring of machines assisted by analytical methods it is possible to monitor and quantify dynamic parameters of a machine. By means of the analysis of dynamic behavior of rotor, it is possible to diagnose and solve the defects relating to rotor and other parts of machine. Recently, vibrations analysis and monitoring have become the most widespread method of machine diagnostics. Detection, identification and elimination of problems related to machine work can be simplified by modeling some dynamics solutions. The paper presents a mathematical model of elastic rotor with concentrated weight designed for the purpose of detecting defects with the help of vibrations measured by sensors based in bearings. Experimental results from test tables of the world＇s leading company, Bentley Nevada Co. USA, were used as the basis for development of this model. Accordingly, mathematical model for most of the cases was based on radial oscillation of rotor while torsion oscillations were used as the basis in some cases. Torsion vibrations cannot be detected without special sensors even at their highest magnitudes. Reliability of this model and accuracy of vibrodiagnostic analysis has been experimentally confirmed when applied on machines in operation.

A Knowledge Management Framework of the Information Technology-Based Approach for Improving Occupational Safety Training of Young Workers

This paper aims to develop a novel knowledge management framework for improving the system of pro-safety attitudes among young worker population in Poland. This is done through an efficient use of existing sources of explicit and tacit knowledge in the field of vocational occupational safety and health （OSH） training. This project identifies three specific research steps for the development of the knowledge management framework of the information technology-based approach for improving occupational safety training of young workers, including： （1） the development of the conceptual model of the flow of explicit knowledge of OSH; （2） evaluation of the effectiveness of formal and explicit knowledge sources for OSH and training methods using these sources in the context of attitudes towards health and safety at work; and （3） verification of the proposed model of vocational training with the use of combined explicit and hidden knowledge. The proposed framework includes consideration of the effectiveness of the formal and informal sources of safety knowledge. A mapping from the formal sources of explicit knowledge about the occupational safety to expected attitudes （hidden knowledge） towards OSH outcomes, including emotional, cognitive, and behavioral aspects has been used. The above framework should help to improve the system of vocational training for young workers in Poland.

遥感大模型:进展与前瞻

近年来,人工智能领域大语言模型和视觉基础模型的显著进展引发了学者们对遥感领域通用人工智能技术的关注,推动了遥感信息处理大模型研究的新范式。遥感大模型也称为遥感预训练基础模型,是一种利用大量的未标注遥感图像来训练大规模深度学习模型的方法,目的是提取遥感图像中的通用特征表示,进而提高遥感图像分析任务的性能、效率和通用性。遥感大模型的研究涉及3个关键因素:预训练数据集、模型参数量和预训练技术。其中,预训练数据集和模型参数量能够随着数据和计算资源的增加而灵活地扩大,预训练技术则是提升遥感大模型性能的关键因素。以遥感大模型的预训练技术为主线,归纳分析了现有的有监督单模态预训练遥感大模型、无监督单模态预训练遥感大模型和视觉-文本联合多模态预训练遥感大模型。最后,对遥感大模型在结合遥感领域知识与物理约束、提高数据泛化性、扩展应用场景以及降低数据成本4个方面,对遥感大模型进行了展望。

An evaluation of the economic,social,and ecological risks of China-Mongolia-Russia high-speed railway construction and policy suggestions

The construction of China-Mongolia-Russia high-speed railways is a strategic move to promote transportation infrastructure inter-connectivity between these countries, which will accelerate the implementation of the China-Mongolia-Russia Economic Corridor. However, well-planned China-Mongolia-Russia high-speed railways demand accurately identifying construction risks, scientifically evaluating risk levels, and mapping the spatial distribution of these risks. Therefore, this study established the integrated risk evaluation model(IREM) to scientifically evaluate the economic, social, and ecological risks of China-Mongolia-Russia high-speed railway construction and determine their magnitude and spatial distribution pattern. Based on this analysis, we propose designs for the east and west China-Mongolia-Russia high-speed railways and policy suggestions to mitigate construction risks. Suggestions include developing innovative cooperation of the "high-speed railway for resources and market", strengthening communication and technology dissemination, and applying innovative engineering techniques and setting buffers; establishing collaborative prevention and control systems to mitigate the three major ecological risks in the China, Mongolia, and Russia trans-border areas; and promoting economic integration by improving strategic coordination. In summary, this study provides scientific support for designing the China-Mongolia-Russia high-speed railways minimizing construction risks.

Science Letters:A new quantitative model of ecological compensation based on ecosystem capital in Zhejiang Province,China

Ecological compensation is becoming one of key and multidiscipline issues in the field of resources and environmental management. Considering the change relation between gross domestic product (GDP) and ecological capital (EC) based on remote sensing estimation, we construct a new quantitative estimate model for ecological compensation, using county as study unit, and determine standard value so as to evaluate ecological compensation from 2001 to 2004 in Zhejiang Province, China. Spatial differences of the ecological compensation were significant among all the counties or districts. This model fills up the gap in the field of quantitative evaluation of regional ecological compensation and provides a feasible way to reconcile the conflicts among benefits in the economic, social, and ecological sectors.

A dynamic model of skeletal muscle based on collective behavior of myosin motors-Biomechanics of skeletal muscle based on working mechanism of myosin motors(I)

A dynalnic model of skeletal muscle is developed to describe its activation kinetics and contraction dynamics based on the collective working mechanism of myosin II motors with a statistical mechanics method. According to the structure of sar- comeres arranged in series and in parallel, the mechanical properties of skeletal muscle are studied. This model reveals the re- lations between action potential and muscle characteristics. It is shown that calcium concentration in sarcoplasmic （SP） in- creases linearly with the increasing stimulation frequency and gradually reaches saturation. Active force and contraction veloc- ity follow the trend of calcium concentration and reach a peak value at the maximum stimulation frequency. Contraction ve- locity is inversely proportional to the load while the contraction power increases to maximum and then reduces to zero with the increasing load. These properties are consistent with published physiological experimental results of skeletal muscle.

Distributed parametric modeling and simulation of light polarization states using magneto-optical sensing based on the Faraday effect

To solve the problems encountered in practical processes of magneto-optical sensing, the infinitesimal distributed-parameter model and finite-element accumulation of different dielectric properties of micromaterials were used to describe the evolution of light polarization states, instead of the previously commonly used method of lumped-parameter simulation, thus essentially explaining the mechanism of sensing, magneto-optical effects, and related factors, and achieving multiphysics coupling using the COMSOL finite-element analysis method. Considering the cases of the Faraday effect without and with line birefringence, the magneto-optical effect and output characteristics of an infinitesimal magneto-optical sensor were simulated and studied. The results verified the effectiveness of the infinitesimal sensor model. Because the magnetic field, stress, and temperature changes alter the dielectric properties of magneto-optical materials, the finite-element accumulation method lays a good foundation for research on theoretical analysis and performance of magneto-optical sensors affected by factors such as the magnetic field, temperature, and stress.