Journal of University of Chinese Academy of Sciences

Carbon-nitrogen biogeochemical cycle processes driven by ammonia-oxidizing archaea in marine environment

HONG Yiguo, HE Xiang, WU Jiapeng, LIU Xiaohan

2020, 37 (4): 433-441. DOI: 10.7523/j.issn.2095-6134.2020.04.001

Ammonia-oxidizing archaea (AOA), one kind of the most abundant prokaryotes in the seawater, is widely distributed in the marine environment. As the dominant microorganism of ammonia oxidation in the ocean, AOA fixes carbon with chemoautotroph under oligotrophic conditions and plays an important role in marine nitrogen and carbon cycles. This autotrophic carbon sequestration driven by AOA ammonia oxidation process is an important energy source in deep sea, which is significant for understanding the imbalance between organic carbon input and microbial energy demand in deep sea. At the same time, the energy from this autotrophic microorganism fuels the deep-sea ecological system. In this paper, we review the diversity distribution of AOA and the process and function of AOA in biogeochemical cycles of carbon and nitrogen in order to provide information for the researchers in this field.

Influence of underground coal mining on soil fertility quality in the northwestern arid and semi-arid regions: a review

MA Kang, YANG Fan, ZHANG Yuxiu

2020, 37 (4): 442-449. DOI: 10.7523/j.issn.2095-6134.2020.04.002

Abstract ( 320 ) PDF (KB) ( 2 )

The arid and semi-arid regions are typically ecologically fragile areas in China. Underground coal mining caused geological structure damage and surface deformation, resulting in land subsidence and ground fissures, and further leading to an increase in soil sand content and porosity, as well as a decrease in soil water retention capacity. Moreover, the loss of soil nutrients is triggered, the microbial community is disturbed, and soil enzyme activity is also affected. The soil fertility quality might be reduced ultimately by the mining. This paper mainly focuses on the mechanism of coal mining subsidence and change trend of soil structure, based on the effects of coal mining on soil physical, chemical, and biological properties, the mechanism of underground coal mining on soil fertility quality in the northwestern arid and semi-arid regions was discussed. The results have an important theoretical significance for accelerating the restoration of mining subsidence soil, development of land reclamation technology and ecological environment protection.

Point matching algorithm based on machine learning method

TANG Siqi, HAN Congying, GUO Tiande

2020, 37 (4): 450-457. DOI: 10.7523/j.issn.2095-6134.2020.04.003

Abstract ( 666 ) PDF (KB) ( 1 )

Point matching is an important issue of computer vision and pattern recognition, and it is widely used in target recognition, medical image, pose estimation, etc. In this study, we propose a novel end-to-end model (multi-pointer network) based on machine learning method to solve this problem. We capitalize on the idea of multi-label classification to ameliorate the pointer network. Instead of outputting a member of input sequence, our model selects a set of input elements as output. Considering matching problem as a sequential manner, our model takes the coordinates of points as input and outputs correspondences directly. Using this new method, we can effectively solve the translation of the whole space and other large-scale rigid transformations. Furthermore, experiment results show that our model can be generalized to other combinatorial optimization problems in which the output is a subset of input, like Delaunay triangulation.

First-principle quantum transport simulations of a fully spin-polarized device TiCl₃/RhCl₃/TiCl₃

ZHANG Zhen, HUANG Qiang, SHENG Xianlei, ZHENG Qingrong

2020, 37 (4): 458-464. DOI: 10.7523/j.issn.2095-6134.2020.04.004

Abstract ( 340 ) PDF (KB) ( 1 )

Based on the first-principle quantum transport simulation methods within the non-equilibrium Green function combined with density functional theory (NEGF+DFT), we predict that a magnetic tunnel junction (MTJ), consisting of a TiCl₃ semi-metal electrode and a RhCl₃ semiconducting scattering region, could be used as a spin-polarized transport device. We calculate the I-V curves in the ranges of small bias voltage (0-20 mV) and large bias voltage (0-0.6 V), respectively. In the range of small bias voltage, the current under parallel configuration (PC) is much larger than that under antiparallel configuration (APC). In addition, the tunneling magnetoresistance (TMR) always maintains a stable large value of 100%, and so does the spin injection efficiency (SIE) value under PC. In the range of large bias voltage, the TMR value decreases as the voltage increases, but the SIE value remains a stable value of 100%. The nonequilibrium transport properties are explained by analyzing the projected density of state.

Determination of methylated arginines in serum by a dispersive solid-phase extraction coupled with capillary electrophoresis-laser induced fluorescence detection

XIE Shijie, YANG Zecheng, ZHOU Zhou, LIN Yahui, SU Baoman, DING Yongsheng

2020, 37 (4): 465-472. DOI: 10.7523/j.issn.2095-6134.2020.04.005

Abstract ( 245 ) PDF (KB) ( 1 )

We describe a dispersive solid-phase ion exchange extraction method combined with capillary electrophoresis-laser induced fluorescence detection for the determination of asymmetric dimethylarginine, symmetric dimethylarginine, and monomethylarginine in serum samples. The methylated arginine-extracted resin beads were added into a mixed solution of 4-chloro-7-nitrobenzofuran acetonitrile solution and alkaline sodium borate solution (pH=10. 5) for the derivatization in a 60℃ water bath. The three derivatives were subjected to baseline separation in a fused silica capillary under the conditions of a separation electrolyte of 80 mmol/L phosphate (pH=2. 0) and a separation voltage of 20 kV. On the basis of the FDA bioanalytical method validation guidance, the method validation was conducted and the results were satisfactory. This method was used to determine the three methylated arginines in the serum samples from local hospital.

Challenges and experiences of eco-city construction in rapid urbanization area in China: a case study of Xiamen City

LIU Jiakun, LIN Tao, ZHANG Xiao, DENG Fuliang, ZHANG Guoqin, ZHAO Yu, YE Hong, LI Xinhu

2020, 37 (4): 473-482. DOI: 10.7523/j.issn.2095-6134.2020.04.006

Abstract ( 511 ) PDF (KB) ( 2 )

Taking Xiamen City as a typical representative of China's eco-city construction, we addressed and summarized the challenges and successful experiences of eco-city construction during rapid urbanization, through reviewing and analyzing the historical and natural process of Xiamen's eco-city construction in the past four decades. We found that the administrative division, the expansion of built-up areas, and the limitation of natural resources played the important role in the eco-city construction process. According to the major challenges faced during the rapid urbanization, the Xiamen's eco-city construction process was divided into three stages:the material resource constraints stage (from the 1980s to the 1990s), the environmental pollution and ecological restoration stage (from 1990s to the early 21^st century), and the optimization and management of urban spatial pattern stage (from the early 21^st century to the present). Three aspects of experiences can be summarized from the Xiamen's eco-city construction process:1) attention of the government and relevant departments integrate leadership, 2) parallel development of city construction and ecological construction according to natural and environmental capacity, and 3) top-level design of multi-conformity urban development path. The paper provides valuable information for eco-city construction of rapid urbanization areas in China and other developing countries.

FD-RCF-based boundary delineation of agricultural fields in high resolution remote sensing images

LI Sen, PENG Ling, HU Yuan, CHI Tianhe

2020, 37 (4): 483-489. DOI: 10.7523/j.issn.2095-6134.2020.04.007

Abstract ( 586 ) PDF (KB) ( 2 )

High spatial resolution (high resolution) remote sensing images are reliable data sources for creating and updating field graphics databases. However, manual vectorization is a tedious process which costs much time and effort. In order to solve this problem, a method called full dilated-RCF (FD-RCF), which is based on dilated convolution in deep learning, is proposed. Compared with holistically-nested edge detection (HED) and richer convolutional features (RCF), FD-RCF mainly differentiates in the way of combining different results from different layers. FD-RCF consists of 6 stages and takes more care of the way of fusing these outputs from convolution layers into side-outputs. With dilated convolution, FD-RCF decreases the loss of information in deep layer. These methods can totally be used in detecting the boundary of agricultural fields. All of them get F₁-values of over 0.8 in ODS and OIS. FD-RCF gets the highest F₁-values of 0.848 1 and 0.850 2 in ODS and OIS, respectively, and the average precision of 0.795 7. The results gotten from FD-RCF are clearer than other methods and FD-RCF costs less time than manual vectorization.

Remote sensing identification of tourism land use based on object-oriented technology

LUO Kaisheng

2020, 37 (4): 490-497. DOI: 10.7523/j.issn.2095-6134.2020.04.008

Abstract ( 355 ) PDF (KB) ( 1 )

Using the Chinese GF-1 satellite images and taking Nanjing City as an example, we used the object-oriented technology to thematically extract the tourism land use information, and developed the method and process of extracting tourism land use information based on remote sensing method. The results are shown as follows. The average NDVI variance among the 4 layers of GF-1 summer images (VI-Summer) is an effective indicator to identify tourism land. In this study, the mapping accuracy and user accuracy of tourism land extraction are 83.33% and 73.53% respectively, and the research results meet the needs of practice. This shows that GF-1 satellite image and object-oriented technology can be effectively used to identify and monitor tourism lands. The area of tourism land use in Nanjing is 137.34 km². Tourism land mainly concentrates on the south of Yangtze River region including Yuhuatai, Jianye, Xuanwu, Qinhuaihe districts, and their surroundings, while the north of Yangtze River is relatively sparse. The results provides technical support for the rapid and effective supervision of tourism land use using GF-1 satellite images.

An ISM frequency band reliable wireless communication system based on channel clearing technology

JI Xingjian, LIANG Guang, SUN Siyue, JIANG Quanjiang, YU Jinpei

2020, 37 (4): 498-506. DOI: 10.7523/j.issn.2095-6134.2020.04.009

Abstract ( 289 ) PDF (KB) ( 1 )

The design of reliable wireless communication systems in the unlicensed frequency band has become an emerging research hotpot to improve the utilization efficiency of spectrum resources and shorten the development period with the increasing shortage of spectrum resources. This work proposes a reliable wireless communication system in ISM band based on new wireless channel clearing technique and a priori self-interference suppression method. In order to address the inherent self-interference problem of the proposed system, this work proposes a priori information-assistant radio frequency AD clearing strategy, which avoids the influences of self-interference on the bit error rate and loop tracking performance of the communication system. Finally, the performance of the proposed system is verified through software and hardware simulations and experiments.

An underwater mining navigation method based on an improved particle filter

ZHANG Zhihui, FENG Yingbin, LI Zhigang, ZHAO Xiaohu

2020, 37 (4): 507-515. DOI: 10.7523/j.issn.2095-6134.2020.04.010

Abstract ( 166 ) PDF (KB) ( 1 )

An underwater mining navigation method based on an improved particle filter (PF) is proposed to solve the problems of non-Gaussian and intense measurement noise during underwater mining, and a new resampling algorithm is designed, as an improvement, to eliminate the influences of particle degeneration and particle impoverishment of PF. Compared to the resampling algorithms, the proposed algorithm avoids particle impoverishment and improves estimation accuracy. Finally, the estimation accuracies of underwater mining navigation algorithms based on the improved PF and the unscented Kalman filter (UKF) are compared by combining the lake trial data and underwater mining navigation model. The results of simulation experiments manifest that the proposed method has more accurate estimation and remarkable robustness.

Study of deep transfer learning for SAR ATR based on simulated SAR images

WANG Zelong, XU Xianghui, ZHANG Lei

2020, 37 (4): 516-524. DOI: 10.7523/j.issn.2095-6134.2020.04.011

Abstract ( 503 ) PDF (KB) ( 2 )

Using deep convolutional neural networks to realize automatic target recognition of SAR requires a large amount of labeled data. In order to solve the problem caused by the scarcity of SAR real images, we propose a method for improving the target recognition performance of SAR by using simulated SAR images on convolutional neural networks improved by CReLU activation function and batch normalization. The method transfers the effective knowledge learned from a large number of simulated SAR images onto the real SAR images. In the training, the pre-trained convolutional neural networks can be obtained by training by using the simulated SAR images firstly, and the deep transfer learning method is used to effectively solve the problem caused by the insufficiency of SAR image data. The validation experiment is carried out on the MSTAR dataset. The highest recognition accuracy reaches 99.78%, and good recognition results are obtained based on a small amount of SAR image data.

A full-polarimetric SAR tomography method based on hierarchical sparseness

YANG Mudan, WEI Zhonghao, XU Zhilin, ZHANG Bingchen, HONG Wen

2020, 37 (4): 525-531. DOI: 10.7523/j.issn.2095-6134.2020.04.012

Abstract ( 316 ) PDF (KB) ( 2 )

SAR tomography employs the multiple-pass data to achieve the elevation location reconstruction of the observation target, while the fully polarimetric data owns rich scattering information. We combine the full-polarimetric data with SAR tomography. By considering the same characteristic of the sparsity of elevation scatters in urban building and the sparse support set in full-polarimetric data, a solution model based on group sparse constraint and sparse constraint, solved by hierarchical sparse method, is proposed. The performance of the method has been compared with those of the single-polarimetric tomography model and the group sparse-based solution method by Monte Carlo simulation experiments. Meanwhile, the method is also applied to semi-simulation of point target experiments based on real data. The results show that the proposed method improves the accuracy of elevation reconstruction and has better robustness, and it accurately recovers the elevation position and backscatter coefficient of the target at low SNR.

Multi-satellite imaging task planning algorithms based on gene expression programming

MING Weipeng, MA Guangbin, ZHANG Wenyi

2020, 37 (4): 532-538. DOI: 10.7523/j.issn.2095-6134.2020.04.013

Abstract ( 287 ) PDF (KB) ( 2 )

The constraint-satisfaction model is established by analyzing the constraints of multi-satellite imaging mission planning for regional targets, and the mathematical complexity of the model is analyzed. In order to improve the weak global searching ability of the genetic algorithm in multi-satellite imaging mission planning, the gene expression programming (GEP) is first proposed in this work to solve the problem. In the process of algorithm implementation, the inverted genetic operator is designed to enhance the search ability for the optimal solution, and the repository is introduced to preserve elite individuals in the iteration process. The results show that the gene expression programming (GEP) is effective and reasonable in solving multi-satellite imaging planning problems and greatly improves the accuracy of the solution.

Semi-supervised airplane detection in remote sensing images using generative adversarial networks

CHEN Guowei, LIU Lei, GUO Jiayi, PAN Zongxu, HU Wenlong

2020, 37 (4): 539-546. DOI: 10.7523/j.issn.2095-6134.2020.04.014

Abstract ( 524 ) PDF (KB) ( 1 )

Airplane detection in remote sensing images is a challenging task and researchers have been sparing no efforts in making breakthroughs in this topic. The methods based on artificial neural network are the main methods for airplane detection in remote sensing images. However, a large number of data must be annotated for training, which is time-consuming. This is one of the main bottlenecks to limit large-scale data to be effectively used. To address this problem, a generative adversarial networks (GAN)-based semi-supervised object detection method is proposed in this work. In airplane detection in remote sensing images, the method does not need to annotate all of the images, but only a small part of them, and then trains them together with a large number of unannotated data to achieve excellent detection performance. The proposed method combines a traditional detection network with a semi-supervised learning network based on GAN. During the adversarial training procedure, the generator learns the data distribution and generates fake samples, while the discriminator distinguishes between the true samples and the generated fake samples. At the same time, the discriminator is also trained on labeled data. Finally, the discriminator learns the decision boundary that not only separates labeled samples, but also parallels with the distribution boundary of the whole data. Experiments show that, with a large number of images, reducing the proportion of labeled data will significantly reduce the performance of the full-supervised learning method, while the semi-supervised learning method proposed in this work maintains the detection performance due to the use of unlabeled data.

NMF endmember generation method based on abundance distribution constraint

SHI Yue, WANG Hongqi, GUO Xinyi

2020, 37 (4): 547-552. DOI: 10.7523/j.issn.2095-6134.2020.04.015

Abstract ( 304 ) PDF (KB) ( 2 )

In recent years, the endmember generation method based on non-negative matrix factorization (NMF) attracted much attention. The NMF endmember generation method can be used to obtain endmembers and the abundance matrix simultaneously, and the multiplicative update rule works. Because of the non-convexity of the objective function, NMF endmember extraction easily goes into local extrema. Several constraints were imposed on NMF to alleviate the local extremum problem, but they often broke the multiplicative update rules and increased the processing time. In this work, we propose a new method based on abundance distribution constraint, and the multiplicative iterations can be used. The experimental results show that the method improves the efficiency and accuracy of endmember generation.

Intrusion detection method based on entity embedding and long short-term memory networks

LAI Xunfei, LIANG Xuwen, XIE Zhuochen, LI Zongwang

2020, 37 (4): 553-561. DOI: 10.7523/j.issn.2095-6134.2020.04.016

Abstract ( 331 ) PDF (KB) ( 1 )

Due to the inability to effectively deal with the representation of categorical variables in intrusion data, the network intrusion detection has low accuracy and high false negative rate. A method combining entity embedding and long short-term memory network (LSTM) is proposed. First, when the data is preprocessed, the numerical variable data and categorical variable data are separated, and the categorical variable data are mapped into an Euclidean space by using the entity embedding method to obtain a vector representation and then this vector is embedded into the numeric data to get the input data. Then, by inputting the data into the long short-term memory network, the parameters are updated by time back propagation. Thus the optimal embedded vector is obtained as the input feature, and a relatively optimal detection model of the LSTM network is also obtained through training. Experiments are carried out on the data set NSL-KDD, and the results show that entity embedding is an effective method to deal with categorical variables in network intrusion data. The model composed of LSTM network effectively improves the detection rate. In the processing of categorical variables, the accuracy of detection using entity embedding method increases by 1.44 percentage points and the false negative rate decreases by 2.99 percentage points, compared with those using the traditional One-Hot coding method.

Facial landmark detection based on cascade convolutional neural network

SUN Mingkun, LIANG Lingy, WANG Han, HE Wei, ZHAO Luyang

2020, 37 (4): 562-569. DOI: 10.7523/j.issn.2095-6134.2020.04.017

Abstract ( 350 ) PDF (KB) ( 2 )

Current facial landmark detection algorithm has achieved promising recognition rates in constrained environment, but in unconstrained environment it is still susceptible to various factors such as non-uniform ambient illumination, wide range of angles, variations in pose, occlusion, and blur. To deal with these problems, we propose a cascade convolutional network to improve the accuracy and robustness of the landmark detection. In face detection, we propose a DPM-CNN model based on Light-VGGNet, which introduces the location information of the facial features. In this way, the detection accuracy is improved and the impact of face detection on the positioning of landmarks is reduced. In facial landmark detection, this algorithm adopts two different layers of cascade networks progressively to complete the positioning of the internal and external landmarks. Finally, on the FDDB dataset, this new algorithm is proved to have higher accuracy rate and detection speed than other algorithms in face detection or in facial landmark location. This algorithm is also robust in an unconstrained environment.

Internet traffic classification based on the improved one-versus-one method

ZHAO Ze, XU Youyu, TANG Liang, BU Zhiyong

2020, 37 (4): 570-576. DOI: 10.7523/j.issn.2095-6134.2020.04.018

Abstract ( 358 ) PDF (KB) ( 1 )

Accurate traffic classification is an effective guarantee for network management and security. Machine learning-based internet traffic classification became particularly notable in recent years, and feature selection had an important impact on the performance of machine learning. However, the feature selection subset that optimizes the overall classification performance is not the subset that optimizes the classification performance of a particular class, which reduces the upper limit of classification performance. Therefore,a new traffic classification model based on the improved one-versus-one method is proposed. In the new traffic classification model, traffic multi-classification task is split into multiple independent sub-tasks.Then feature selection and traffic classification are performed on any two classes of traffic,and the Stacking strategy is used to combine the results of all sub-tasks. The experiments show that the applications of several machine learning and feature selection algorithms to this model improve accuracies compared with those to the classical model.

Current Issue