ECM算法的多视SAR影像分割

多视SAR影像像素强度通常建模为Gamma分布,其形状参数为常数(视数)。实验表明,多视SAR影像分割时,设Gamma分布的形状参数为变量可取得更好的分割结果。由于Gamma分布中形状参数以Gamma函数的形式出现,利用EM算法求解时无法获得形状参数的解析解。为此,本文提出了一种基于Expectation/Conditional Maxinnization(ECM)算法的多视SAR影像分割方法。利用ECM算法估计最大化后验概率条件下的Gamma分布参数及表征最优多视SAR影像分割的标号场实现。采用模拟和真实多视SAR影像验证提出算法。实验结果表明,Gamma分布的形状和尺度参数均能快速收敛到稳态值,且以此得到各同质区域的Gamma分布曲线可以很好地拟合其直方图。通过对分割结果的定性和定量分析,可知提出算法具有有效性和可行性,且优于EM算法。     

面向第二次全国地名普查的数字化调查系统建设研究

针对第二次全国地名普查中现有技术路线的不足,提出了基于移动GIS的数字化调查解决方案。结合地名普查的实际要求,详细介绍了数字化调查系统的系统构架、数据组织、系统功能等,并探讨了系统实现的关键技术。通过实际应用,验证了系统建设的合理性和可操作性,为各地名普查工作提供了较好的参考和借鉴。     

面向应用的航天遥感科学论证研究

为适应中国民用航天遥感从科学试验向业务服务模式转变,更好地探索、了解与解决应用需求与航天遥感系统对接等方面遇到的技术问题,促进航天遥感统筹协调可持续发展,中国适时于2004年成立了国家航天局航天遥感论证中心。10余年来,论证中心以航天遥感系统为研究对象,系统开展了面向应用的航天遥感科学论证概念、理论方法、技术工程与应用研究。本文是论证中心团队长期从事航天遥感科学论证研究与实践的系统总结,介绍了遥感论证初步认知、遥感论证关注问题、遥感论证理论体系与模型方法集、遥感论证能力建设及遥感论证实践等方面内容,给出了遥感论证定义并详细分析了研究范围和内容,提出了由知识维、进程维和逻辑维所组成的遥感论证作用域3维空间结构,指出社会发展加快和信息化水平提高,带动整个航天遥感数据信息链向更大规模、更短响应时间周期、更综合数据集成、更高数据质量、更加智能化方向发展,航天遥感系统将进入新的"智慧遥感"发展阶段。得益于十余年来中国民用航天快速发展,我们经历了风云三号新型载荷校飞、多角度多光谱偏振遥感器论证、环境星应用工程论证等实践,取得了多方面理论方法的突破,并应用到2030民用航天发展规划、高分辨率对地观测系统、国家自然灾害空间信息基础设施、国家民用空间基础设施中长期发展、2030中国综合地球观测系统规划等论证当中。经过不断实践,快速迭代,形成了遥感论证理论体系及相应的十大模型方法,包括遥感信息流模型、遥感信息特征模型、遥感信息应用模型、遥感信息量分析模型、遥感数据工程模型、航天遥感系统结构模型、航天遥感系统状态描述模型、航天遥感系统质量模型、航天遥感系统发展动力模型及能力体系模型。这些模型方法全面反映了航天遥感系统特征、结构、状态、发展动力、条件等,可广泛用于对航天遥感系统进行顶层设计、规划、考察、分析、评价、预测,并开展实践探索。     

TEC variation over an equatorial and anomaly crest region in India during 2012 and 2013

Using the global positioning system (GPS) measurements, the total electron content (TEC) at station Bangalore (13.02°N, 77.57°E geographic; 04.44°N, 150.84°E geomagnetic), lying at the equatorial region, and station Lucknow (26.91°N, 80.95°E geographic; 17.96°N, 155.24°E geomagnetic), lying at equatorial ionospheric anomaly (EIA) crest region, have been estimated for the year 2012–2013. In order to evaluate the International Reference Ionosphere (IRI) model regarding simulation/modeling of ionospheric studies specially at equatorial and EIA crest regions, we have compared the TEC derived from the recent version of the IRI-2012 model and the older IRI-2007 with its three topside options, namely IRI-NeQuick (IRI-NeQ), IRI-2001, and IRI01-corr, with that of GPS-TEC over Bangalore and Lucknow. For the EIA station Lucknow, the IRI-2012 model with IRI-NeQ and IRI01-corr topside is found in good agreement with GPS-TEC during summer and equinox season, while the IRI-2012 model for all three topside options significantly overestimates the GPS-TEC during winter season. The IRI-2001 topside overestimates the GPS-TEC over both the stations during all seasons. The anomalous difference between the IRI-2012 model prediction and ground-based GPS-TEC in daytime hours during the winter season observed at Lucknow could be attributed to discrepancies in the slab thickness predicted by the model, which is more during the winter season as compared to summer and equinox. These large discrepancies in the slab thickness predicted by the IRI-2012 as well as the IRI-2007 model during the winter season have been supported by using the foF2 data from Constellation Observing System for Meteorology, Ionosphere, and Climate radio occultation-based measurements. We also observed that the discrepancies in the recent IRI-2012 model with respect to GPS-TEC are found to be slightly larger than those with the older IRI-2007 model over the EIA region Lucknow. However, over the equatorial region Bangalore, the discrepancy with the older model IRI-2007 was found to be larger than with the recent IRI-2012 model. This suggests that the performance of the IRI-2012 model is poorer than the IRI-2007 model at the EIA region while better at equatorial region, and that further improvements in the IRI-2012 models are required particularly in the low-latitude and EIA regions. The GPS-TEC showed disappearance of the winter anomaly during 2012–2013, while the IRI model failed to predict the disappearance of winter anomaly.     

Long-term soil moisture dynamics derived from GNSS interferometric reflectometry: a case study for Sutherland,South Africa

Soil moisture is a geophysical key observable for predicting floods and droughts, modeling weather and climate and optimizing agricultural management. Currently available in situ observations are limited to small sampling volumes and restricted number of sites, whereas measurements from satellites lack spatial resolution. Global navigation satellite system (GNSS) receivers can be used to estimate soil moisture time series at an intermediate scale of about 1000 m². In this study, GNSS signal-to-noise ratio (SNR) data at the station Sutherland, South Africa, are used to estimate soil moisture variations during 2008–2014. The results capture the wetting and drying cycles in response to rainfall. The GNSS Volumetric Water Content (VWC) is highly correlated (r ² = 0.8) with in situ observations by time-domain reflectometry sensors and is accurate to 0.05 m³/m³. The soil moisture estimates derived from the SNR of the L1 and L2P signals compared to the L2C show small differences with a RMSE of 0.03 m³/m³. A reduction in the SNR sampling rate from 1 to 30 s has very little impact on the accuracy of the soil moisture estimates (RMSE of the VWC difference 1–30 s is 0.01 m³/m³). The results show that the existing data of the global tracking network with continuous observations of the L1 and L2P signals with a 30-s sampling rate over the last two decades can provide valuable complementary soil moisture observations worldwide.     

Using a regional numerical weather prediction model for GNSS positioning over Brazil

The global navigation satellite system (GNSS) can provide centimeter positioning accuracy at low costs. However, in order to obtain the desired high accuracy, it is necessary to use high-quality atmospheric models. We focus on the troposphere, which is an important topic of research in Brazil where the tropospheric characteristics are unique, both spatially and temporally. There are dry regions, which lie mainly in the central part of the country. However, the most interesting area for the investigation of tropospheric models is the wet region which is located in the Amazon forest. This region substantially affects the variability of humidity over other regions of Brazil. It provides a large quantity of water vapor through the humidity convergence zone, especially for the southeast region. The interconnection and large fluxes of water vapor can generate serious deficiencies in tropospheric modeling. The CPTEC/INPE (Center for Weather Forecasting and Climate Studies/Brazilian Institute for Space Research) has been providing since July 2012 a numerical weather prediction (NWP) model for South America, known as Eta. It has yield excellent results in weather prediction but has not been used in GNSS positioning. This NWP model was evaluated in precise point positioning (PPP) and network-based positioning. Concerning PPP, the best positioning results were obtained for the station SAGA, located in Amazon region. Using the NWP model, the 3D RMS are less than 10 cm for all 24 h of data, whereas the values reach approximately 60 cm for the Hopfield model. For network-based positioning, the best results were obtained mainly when the tropospheric characteristics are critical, in which case an improvement of up to 7.2 % was obtained in 3D RMS using NWP models.     

Generating GPS satellite fractional cycle bias for ambiguity-fixed precise point positioning

With the development of precise point positioning (PPP), the School of Geodesy and Geomatics (SGG) at Wuhan University is now routinely producing GPS satellite fractional cycle bias (FCB) products with open access for worldwide PPP users to conduct ambiguity-fixed PPP solution. We provide a brief theoretical background of PPP and present the strategies and models to compute the FCB products. The practical realization of the two-step (wide-lane and narrow-lane) FCB estimation scheme is described in detail. With GPS measurements taken in various situations, i.e., static, dynamic, and on low earth orbit (LEO) satellites, the quality of FCB estimation and the effectiveness of PPP ambiguity resolution (AR) are evaluated. The comparison with CNES FCBs indicated that our FCBs had a good consistency with the CNES ones. For wide-lane FCB, almost all the differences of the two products were within ±0.05 cycles. For narrow-lane FCB, 87.8 % of the differences were located between ±0.05 cycles, and 97.4 % of them were located between ±0.075 cycles. The experimental results showed that, compared with conventional ambiguity-float PPP, the averaged position RMS of static PPP can be improved from (3.6, 1.4, 3.6) to (2.0, 1.0, 2.7) centimeters for ambiguity-fixed PPP. The average accuracy improvement in the east, north, and up components reached 44.4, 28.6, and 25.0 %, respectively. A kinematic, ambiguity-fixed PPP test with observation of 80 min achieved a position accuracy of better than 5 cm at the one-sigma level in all three coordinate components. Compared with the results of ambiguity-float, kinematic PPP, the positioning biases of ambiguity-fixed PPP were improved by about 78.2, 20.8, and 65.1 % in east, north, and up. The RMS of LEO PPP test was improved by about 23.0, 37.0, and 43.0 % for GRACE-A and GRACE-B in radial, tangential, and normal directions when AR was applied to the same data set. These results demonstrated that the SGG FCB products can be produced with high quality for users anywhere around the world to carry out ambiguity-fixed PPP solutions.     

A closed-loop EKF and multi-failure diagnosis approach for cooperative GNSS positioning

Current cooperative positioning with global navigation satellite system (GNSS) for connected vehicle application mainly uses pseudorange measurements. However, the positioning accuracy offered cannot meet the requirements for lane-level positioning, collision avoidance and future automatic driving, which needs real-time positioning accuracy of better than 0.5 m. Furthermore, there is an apparent lack of research into the integrity issue for these new applications under emerging driverless vehicle applications. In order to overcome those problems, a new extended Kalman filter (EKF) and a multi-failure diagnosis algorithm are developed to process both GNSS pseudorange and carrier phase measurements. We first introduce a new closed-loop EKF with partial ambiguity resolution as feedback to address the low accuracy issue. Then a multi-failure diagnosis algorithm is proposed to improve integrity and reliability. The core of this new algorithm includes using Carrier phase-based Receiver Autonomous Integrity Monitoring method for failure detection, and the double extended w test detectors to identify failure. A cooperative positioning experiment was carried out to validate the proposed method. The results show that the proposed closed-loop EKF can provide highly accurate positioning, and the multi-failure diagnosis method is effective in detecting and identifying failures for both code and carrier phase measurements.     

Modified compressive sensing approach for GNSS signal reception in the presence of interference

Pre-processing traditional navigation signals in global navigation satellite system (GNSS) receivers includes the conversion of an analog-to-digital sample and acquisition following the basic principle of Nyquist sampling theory. This condition inevitably increases the system computation time and cost of a modern wideband receiver. In recent years, the compressive sensing (CS) approach has been proven to effectively reduce the number of measurement samples required for digital signal acquisition systems. This method gives new potential to this modern design. In this study, a modified compressive sensing algorithm for the acquisition of a GNSS signal that is contaminated by an interfering signal is presented. The proposed method attempts to combine CS demodulation and the subspace projecting method to enhance GNSS signal acquisition performance with interference present. First, the received signal is sub-sampled and aliased from a compressive sampling process. This operation maintains the restricted isometry property (RIP) condition of the second sampling process using a Toeplitz-structured sensing matrix, which replaces a conventional random sensing matrix. The matrix ensures that distances between desired signals on the set of sparse space are not influenced by the sampling process. Next, the interference is eliminated through the orthogonal feature between the interference signal and the desired signal using the subspace projecting method. This also preserves the RIP of the projecting matrix to ensure that the original structure of the linear projection of the signal is preserved. After this, an iterative least-square method is utilized to recover the correlator output from the reception samples taken by the CS demodulator. In addition, the signal detection performance in the presence of co-channel interference using a CS demodulator is analyzed and evaluated. Finally, the relationships between signal detection probability, compressive factor and signal bandwidth are also illustrated. Several numerical results are presented to verify the theory.