Journal of University of Chinese Academy of Sciences

Review of studies on electrical conductivity of hydrous minerals

WANG Libing, WANG Duojun, SHEN Kewei

2022, 39 (4): 433-448. DOI: 10.7523/j.ucas.2022.027

Hydrous minerals play a vital role in the Earth's interior water cycle. The path of water cycle and the change of composition in the Earth's interior can be estimated by the electrical conductivity of typical hydrous minerals. In this paper, we summarized the electrical conductivities, conduction mechanisms, dehydration mechanisms and geophysical implications for typical hydrous minerals before and after dehydration under high temperatures and pressures. The electrical conductivities of most hydrous minerals before dehydration are generally distributed at 10^-4S/m, and increase significantly after dehydration. There are differences in the electrical conductivities of different hydrous minerals. The coupling of dehydration mechanisms and conduction mechanisms may determine the electrical conductivity of hydrous minerals during dehydration. The electrical conductivity of hydrous minerals after dehydration generally depends on the composition and connectivity of the fluid. The changes of electrical conductivities of hydrous minerals before and after dehydration provide experimental constraints for the high-conductivity anomalies at different depths in subduction zones. In addition, the determination of the composition of the fluid derived from the dehydration of the hydrous minerals enriched our knowledge of material circulation in subduction zones. Based on the recent progress, the potential research directions of the electrical conductivity of hydrous minerals are prospected.

Using copula method with SMAA in decision-making analysis

DENG Wei, YE Wuyi, YANG Feng

2022, 39 (4): 449-462. DOI: 10.7523/j.ucas.2020.0061

Abstract ( 297 ) PDF (0KB) ( 0 )

Stochastic multi-criteria acceptability analysis (SMAA) is a series of methods for multicriteria decision making. It is used to give decision opinions on the problem of uncertain or missing preferences of decision makers, the uncertainty of decision variables and incomplete or missing preference information can be represented by probability distribution. This method improves the past methods by considering inversely what specific parameters each decision outcome is determined by. However, the existing SMAA methods does not take into account the impact of the interdependence between these variables on the analysis results, or simply using a simple model such as a Gaussian distribution to describe these interdependencies, so that the analysis results do not have sufficient scientific basis. In order to fix this fault, this paper create an innovation by combining copula dependency analysis with stochastic multi-criteria acceptability analysis method, and uses vine copula modeling method to describe the uncertainty of decision variables and their interdependence, making the SMAA method complete and more effective. This paper introduces the basic methods of SMAA and vine copula separately, and gives the specific steps of combining the two methods. In simulation experiments, a comparative analysis is conducted between the original SMAA method and the method given in the paper to show their advantages and disadvantages under different dependency structures.

Theoretical study on the inert C-H arylation and alkylation by metallaphotoredox catalysis?

ZHENG Xiaofan, ZHANG Beibei, LI Deqing, CHEN Bozhen

2022, 39 (4): 463-480. DOI: 10.7523/j.ucas.2020.0015

Abstract ( 351 ) PDF (0KB) ( 12 )

Metallaphotoredox, including organic photoredox and transition metal catalysis, has emerged as a forceful method to realize the inert C-H bond functionalization to construct C-C or C-heteroatom bonds. In this research, the detailed mechanisms for sp³ C-H arylation and alkylation of tetrahydrofuran (THF) with aryl halide and alkyl halide catalyzed by the triplet excited diaryl ketone and nickel complex have been investigated using density functional theory calculations. The calculations indicate that the whole reaction includes three reaction processes:the formation of (THF) radical, generation of the product catalyzed by nickel catalysts (nickel catalysis), and regeneration of the Ni catalyst. The (THF) radical could be produced by the triplet ketone extracting the H atom of THF. For the nickel catalysis, NiL (L=5,5'-dimethyl-2,2'-bipyridine), not Niacac, plays an important role in the arylated reaction. In addition, the Na₂CO₃ species should be indispensable to the regeneration of NiL. Moreover, the similar results have been obtained for the sp³ C-H alkylation catalyzed by NiL' (L'=4,4'-diterbutyl-2,2'-bipyridine).

Spatial distribution and ecological risk assessment of heavy metals in a lead smelting site in Central China

HANG Yufei, TAN Jingqiang, DENG Min, HU Guoqing, GUO Zhaohui, LI Chuxuan, XUE Shengguo

2022, 39 (4): 481-489. DOI: 10.7523/j.ucas.2021.0080

Abstract ( 398 ) PDF (0KB) ( 0 )

Taking a lead smelting site in Central China as the research object, the contents of 8 heavy metals, including As, Cd, Cr, Cu, Hg, Ni, Pb, and Zn, in the soil of the smelter were determined in order to explore the spatial distribution of heavy metal pollution in the soil of nonferrous smelting sites and its ecological risks. Using the interpolation analysis method of ArcGIS, the spatial distribution characteristics of heavy metals in the soil was studied. The Nemerow index method and potential ecological risk assessment were used to assess the contamination level. The lead smelting site showed severe contamination. The over-standard rates of As, Pb, and Cr in the soil were 80%, 50%, and 50%. The spatial distribution characteristics of surface soil heavy metal pollution were mainly local spotty pollution and the high content of heavy metals was mostly concentrated in the electrolysis workshop due to the accumulation of heavy metal containing waste residue. According to the evaluation results of Nemerow pollution index, Pb and As in the research area reached the level of severe pollution. The results of potential ecological risk assessment showed that As, Cd, Hg, and Pb had a extremely strong risk. Combining the two pollution assessment methods, Pb, As, Hg, Cd were the target pollutants of soil remediation in this area. The research results have important scientific significance and practical guidance value for heavy metal remediation in lead smelting site.

Hot spots tracking of nighttime light data application in research of urbanization and its resource and environmental effects

ZHANG Xiaoping, GAO Shanshan, CHEN Mingxing, ZHAO Yanyan

2022, 39 (4): 490-501. DOI: 10.7523/j.ucas.2021.0010

Abstract ( 784 ) PDF (0KB) ( 0 )

Being closely related to human socioeconomic activity and its footprints, nighttime light (NTL) data shows great advantages in urbanization and socioeconomic development research, especially in densely populated cities. Based on CiteSpace software and the core databases of CNKI (China National Knowledge Infrastructure) and WOS (Web of Science), this paper tracked the hot spots of NTL data in the study of urbanization and related resource consumption and environmental effects from 2000 to 2019. The main results are as follows. 1) Urbanization was the main focus of the application of NTL data, but the researches on the resource consumption and environmental effects caused by urbanization were slightly weak, which was more obvious in Chinese literature. 2) Researches of urban expansion and urban form evolution focused on process of land expansion based on different features of NTL datasets, while in researches of population, socioeconomic development, electricity consumption and carbon emissions, NTL data usually played the role as a supporting tool to explore spatiotemporal characteristics and mechanism. 3) In regards of air pollution and urban heat island induced by urbanization, NTL datasets were usually used to represent factors related to human activities and their impacts. 4) Urbanization process and its impacts on resource and environment are complex, the improved spatial resolution and integrated multi-source data, along with new methods as machine learning, will make the urbanization related research be more precise and scientific. Finally, the paper summarizes the possible new directions of the application of NTL data in urban geography.

Analysis of the influence of different algorithms of GEDI L2A on the accuracy of ground elevation and forest canopy height

LIU Lijuan, WANG Cheng, NIE Sheng, ZHU Xiaoxiao, XI Xiaohuan, WANG Jinliang

2022, 39 (4): 502-511. DOI: 10.7523/j.ucas.2021.0076

Abstract ( 459 ) PDF (0KB) ( 1 )

The global ecosystem dynamics investigation(GEDI) is a full-waveform LiDAR system launched by United States in December 2018, which can provide data support for global ground elevation and forest canopy height. In order to adapt to different surface environments, GEDI L2A products provide 6 different algorithms to estimate ground elevation and forest canopy height. The choice of these algorithms will affect the extraction accuracy of surface parameters. In this paper, the digital terrain model and canopy height model obtained by airborne laser scanning data are used as reference data to evaluate the adaptability of different algorithms of the second version of GEDI L2A data under different vegetation coverage values and the impact on the accuracy of ground elevation and forest canopy height. The results show that when the vegetation coverage is less than 0. 2, the result of algorithm 4 is the best, when the coverage is greater than or equal to 0. 8, the result of algorithm 2 is the best, and the result of algorithm 1 is the best under the remaining coverage values. Comparing the optimal algorithm selected in this paper according to the vegetation coverage with the GEDI L2A default optimal algorithm, the results of this paper are generally better than those of the GEDI L2A default optimal algorithm, especially when the coverage value is less than 0. 8 and the slope is greater than or equal to 10°. The optimal algorithm selected in this paper can effectively improve the accuracy of GEDI L2A products for estimating ground elevation and forest canopy height.

Building extraction based on UNet++ network with different backbones

GU Yumin, YAN Fuli

2022, 39 (4): 512-523. DOI: 10.7523/j.ucas.2020.0040

Abstract ( 786 ) PDF (0KB) ( 0 )

Automatic building extraction methods based on deep learning theory have the technical characteristics of high accuracy and speed,and are of great significance in industrial applications, such as urban planning,disaster prevention and mitigation. This paper introduces the deep learning modules and the traditional remote sensing validation section in the proposed building extraction method in high-resolution remote sensing imageries, forming an operational deep-learning-theory based building extraction technical system that integrates different backbone modules, UNet + + networks,and remote sensing authenticity verification modules. The basic network is transformed through the traditional convolutional network model backbones,such as VGG,ResNet, and Inception to improve the model operational efficiency,strengthen the model feature learning capabilities,verify the effectiveness and applicability of the algorithm through authenticity validation. Taking the Massachusetts building dataset published by Mnih as the data source,a comparative analysis was carried out with the traditional non-full convolutional network model and full convolutional network model. The results show that an increasing in the depth and width of the model can substantially improve the building extraction results. The InceptionV3-UNet + + backbone model has the best performance in recall rate,accuracy,CSI,F1 score,Kappa coefficients, and total accuracy,reaching 85. 14%,90. 50%,0. 781 6,0. 877 4,0. 850 4, and 95. 57%,respectively,and its robustness is also verified on the WHU datasets. This method has significantly improved the extraction accuracy and the details of the buildings extracted, especially on complex and irregular buildings, which will facilitate the building extraction applications in real, complex, and large scene of high-resolution remote sensing imageries.

Removing highlights from single image via an attention-auxiliary generative adversarial network

ZHAO Xinchi, JIANG Ce, HE Wei

2022, 39 (4): 524-531. DOI: 10.7523/j.ucas.2020.0018

Abstract ( 687 ) PDF (0KB) ( 0 )

The highlights in the image will degrade the image quality to some extent. In this paper, we focus on visually removing the highlights from degraded images and generating clean images. In order to solve this problem, we present an attention-auxiliary generative adversarial networks. It mainly consists of the convolutional long short term memory network with squeeze-and-excitation (SE) block and the map-auxiliary module. Map-auxiliary can instruct the autoencoder to generate clean images. The injection of SE block and map-auxiliary module to the generator is the main contribution of this paper. And our proposed deep learning-based approach can be easily ported to handle other similar image recovery problems. Experiments prove that the network architecture is effective and makes a lot of sense.

Remote sensing satellite ground station antenna intelligent scheduling with LSTM and heuristic search

SUN Wenjun, MA Guangbin, TIAN Miaomiao, LIN Youming, HUANG Peng

2022, 39 (4): 532-542. DOI: 10.7523/j.ucas.2020.0014

Abstract ( 361 ) PDF (0KB) ( 2 )

In order to solve the shortage of remote sensing satellite data receiving antennas and improve the utilization of the ground antennas, an intelligent scheduling method which combines LSTM (long short-term memory network) and heuristic search was proposed. First, LSTM is used to extract the antenna using rules from the historical scheduling data of antennas, and then the initial scheduling scheme is obtained by allocating an antenna for each remote sensing data receiving task with the rules; Second, the heuristic search is used to solve the two problems of joint data reception and resource selection conflict in the initial plan, and obtain a practical and feasible scheduling plan. The experiment results show that the method is useful to deal with ground antenna scheduling, improve resource efficiency and reduce computing time to some extent when compared with genetic algorithm.

Downlink power allocation scheme for LEO satellites based on deep reinforcement learning

ZHANG Huaming, LI Qiang

2022, 39 (4): 543-550. DOI: 10.7523/j.ucas.2020.0045

Abstract ( 391 ) PDF (0KB) ( 6 )

Most of the current satellite resource allocation schemes are designed for geosynchronous orbit satellites. In view of the highly dynamic characteristics and limitation of frequency and power resources in LEO satellites, a power allocation algorithm based on deep reinforcement learning is proposed. First of all, we model the LEO satellite power allocation scenario, and introduce a time slot division scheme to simplify the dynamic characteristics model of the LEO satellite. Then a power allocation policy is proposed based on deep reinforcement learning algorithm which can reduce the co-channel interference by adjusting the power value of the subcarriers in each beam of a single LEO satellite, thus improving the spectral efficiency of the LEO satellite. Simulation results illustrate that the proposed algorithm can converge and reach a stable state in a relatively short time. Under the condition of constant total power, this scheme can effectively improve the throughput of a single LEO satellite. The spectral efficiency based on deep reinforcement learning algorithm is significantly higher than that of water-filling algorithm and Q-learning algorithm.

Classification models based on generative adversarial networks with mutual information regularization

HU Bingbing, TANG Hua, WU Youlong

2022, 39 (4): 551-560. DOI: 10.7523/j.ucas.2020.0037

Abstract ( 405 ) PDF (0KB) ( 0 )

This paper studies classification models based on generative adversarial networks with mutual information regularization. Traditional machine learning methods rely on a large number of labeled datasets, which are scarce in practice, to train the model and can easily overfit to spurious correlations in the data; while generating adversarial networks can be trained in an unsupervised manner. In addition, mutual information constraint allows the model to generate data of a specified through latent variables, which has a certain significance for data augmentation. Moreover, after adding a small amount of label information, the accuracy of the model can be improved. category, which can be used to expand the data set. This paper proposes the InfoCatGAN and CInfoGAN classification models. The former adds the mutual information term to CatGAN model in order to generate images of higher visual fidelity; the latter uses the InfoGAN model for classification, which can ensure the quality of the generated images and provide a mentionable classification accuracy. Additionally, both two models can control the category of generated images

Finite-size scaling analysis of the Planck's quantum-driven integer quantum Hall transition in spin-1/2 kicked rotor model

ZHANG Jialong, ZHANG Long, ZHANG Fuchun

2022, 39 (4): 561-566. DOI: 10.7523/j.ucas.2022.003

Abstract ( 430 ) PDF (0KB) ( 0 )

The quantum kicked rotor (QKR) model is a prototypical system in the research of quantum chaos. In a spin-1/2 QKR, tuning the effective Planck parameter realizes a series of transitions between dynamical localization phases, which closely resembles the integer quantum Hall (IQH) effect and the plateau transitions. In this work, we devise and apply the finite-size scaling analysis to the transitions in the spin-1/2 QKR model. We obtain an estimate of the critical exponent at the transition point, ν=2. 62(9), which is consistent with the IQH plateau transition universality class.

The recognition and utilization of Cannabis sativa in ancient Xinjiang viewed from plant remains

LIU Yan, GU Man, CHEN Tao, WANG Binghua, JIANG Hong'en

2022, 39 (4): 567-576. DOI: 10.7523/j.ucas.2020.0048

Abstract ( 508 ) PDF (0KB) ( 0 )

Xinjiang is an important area connecting the east and west alongside the Silk Road. As a result of its dry weather and less rainfall, plant remains were well preserved, including hemp (Cannabis sativa). The history of Cannabis recognition and utilization in ancient Xinjiang could be roughly divided into two stages. Due to inherent advantages of the geographical environment, Xinjiang was firstly affected by the ancient civilization of Central Asia in the prehistoric period, mainly retaining the hallucinogenic tradition of Cannabis. During the historical period, the beginning of the Silk Road promoted exchanges between the Western Regions and the interior areas. The edible and fiber values of Cannabis were gradually accepted. Therefore, in terms of Cannabis usage, ancestors generally had a transformation from psychoactive agents to food crops, fiber crops, and even medicinal plants. In this study, we reviewed the Cannabis researches in Xinjiang during the past years, summarizing the ways and differences of Cannabis utilization, especially their causes in different periods. Meanwhile, we restudiedCannabis remains in Turpan with plant identification and radiocarbon dating, comprehensively sorting out the handed-down and unearthed documents, to explore not only the utilization of Cannabis medicinal value, but also to provide new clues for the spread of Chinese medicine in ancient Xinjiang.

Current Issue