scholarly journals NLOS Satellite Detection Using Fish-Eye Camera and Semantic Segmentation for Improving GNSS Positioning Accuracy in Urban Area

2019 ◽  
Vol 2019 (0) ◽  
pp. 212-217 ◽  
Author(s):  
Kenta Horide ◽  
Akihiro Yoshida ◽  
Reo Hirata ◽  
Yukihiro Kubo ◽  
Yoshiharu Koya
Keyword(s):  
Urban Area ◽  
Semantic Segmentation ◽  
Positioning Accuracy ◽  
Gnss Positioning ◽  
Fish Eye

NLOS Satellite Detection Using a Fish-Eye Camera for Improving GNSS Positioning Accuracy in Urban Area

2016 ◽  
Vol 28 (1) ◽  
pp. 31-39 ◽  
Author(s):  
Shodai Kato ◽  
◽  
Mitsunori Kitamura ◽  
Taro Suzuki ◽  
Yoshiharu Amano ◽  
...  
Keyword(s):  
Weather Conditions ◽  
Urban Environments ◽  
Global Navigation Satellite Systems ◽  
Image Feature ◽  
Transport Systems ◽  
Positioning Accuracy ◽  
Positioning Systems ◽  
Satellite Systems ◽  
Gnss Positioning ◽  
Fish Eye

[abstFig src='/00280001/03.jpg' width=""300"" text='NLOS satellites detection method' ]In recent years, global navigation satellite systems (GNSSs) have been widely used in intelligent transport systems (ITSs), and many countries have been rapidly improving the infrastructure of their satellite positioning systems. However, there is a serious problem involving the use of kinematic GNSS positioning in urban environments, which stems from significant positioning errors introduced by non-line-of-sight (NLOS) satellites blocked by obstacles. Therefore, we propose the method for positioning based on NLOS satellites detection using a fish-eye camera. In general, it is difficult to robustly extract an obstacle region from the fish-eye image because the image is affected by cloud cover, illumination conditions, and weather conditions. We extract the obstacle region from the image by tracking image feature points in sequential images. Because the obstacle region on the image moves larger than the sky region, the obstacle region can be determined by performing image segmentation and by using feature point tracking techniques. Finally, NLOS satellites can be identified using the obstacle region on the image. The evaluation results confirm the GNSS positioning accuracy without the NLOS satellites was improved compared with using all observed satellites, and confirm the effectiveness of the proposed technique and the feasibility of implementing its highly accurate positioning capabilities in urban environments.

scholarly journals Semantic segmentation of PolSAR image data using advanced deep learning model

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Rajat Garg ◽  
Anil Kumar ◽  
Nikunj Bansal ◽  
Manish Prateek ◽  
Shashi Kumar
Keyword(s):  
Machine Learning ◽  
Remote Sensing ◽  
Deep Learning ◽  
Urban Area ◽  
Urban Areas ◽  
Learning Algorithms ◽  
Semantic Segmentation ◽  
Learning Model ◽  
Machine Learning Algorithms ◽  
Deep Learning Model

AbstractUrban area mapping is an important application of remote sensing which aims at both estimation and change in land cover under the urban area. A major challenge being faced while analyzing Synthetic Aperture Radar (SAR) based remote sensing data is that there is a lot of similarity between highly vegetated urban areas and oriented urban targets with that of actual vegetation. This similarity between some urban areas and vegetation leads to misclassification of the urban area into forest cover. The present work is a precursor study for the dual-frequency L and S-band NASA-ISRO Synthetic Aperture Radar (NISAR) mission and aims at minimizing the misclassification of such highly vegetated and oriented urban targets into vegetation class with the help of deep learning. In this study, three machine learning algorithms Random Forest (RF), K-Nearest Neighbour (KNN), and Support Vector Machine (SVM) have been implemented along with a deep learning model DeepLabv3+ for semantic segmentation of Polarimetric SAR (PolSAR) data. It is a general perception that a large dataset is required for the successful implementation of any deep learning model but in the field of SAR based remote sensing, a major issue is the unavailability of a large benchmark labeled dataset for the implementation of deep learning algorithms from scratch. In current work, it has been shown that a pre-trained deep learning model DeepLabv3+ outperforms the machine learning algorithms for land use and land cover (LULC) classification task even with a small dataset using transfer learning. The highest pixel accuracy of 87.78% and overall pixel accuracy of 85.65% have been achieved with DeepLabv3+ and Random Forest performs best among the machine learning algorithms with overall pixel accuracy of 77.91% while SVM and KNN trail with an overall accuracy of 77.01% and 76.47% respectively. The highest precision of 0.9228 is recorded for the urban class for semantic segmentation task with DeepLabv3+ while machine learning algorithms SVM and RF gave comparable results with a precision of 0.8977 and 0.8958 respectively.

scholarly journals Helmert Variance Component Estimation for Multi-GNSS Relative Positioning

Sensors ◽  
2020 ◽  
Vol 20 (3) ◽  
pp. 669 ◽  
Author(s):  
Mowen Li ◽  
Wenfeng Nie ◽  
Tianhe Xu ◽  
Adria Rovira-Garcia ◽  
Zhenlong Fang ◽  
...  
Keyword(s):  
Variance Component ◽  
Computation Time ◽  
Satellite System ◽  
Unit Weight ◽  
Variance Component Estimation ◽  
Positioning Accuracy ◽  
Gnss Positioning ◽  
Component Estimation ◽  
Weighting Strategy ◽  
Gnss Constellation

The Multi-constellation Global Navigation Satellite System (Multi-GNSS) has become the standard implementation of high accuracy positioning and navigation applications. It is well known that the noise of code and phase measurements depend on GNSS constellation. Then, Helmert variance component estimation (HVCE) is usually used to adjust the contributions of different GNSS constellations by determining their individual variances of unit weight. However, HVCE requires a heavy computation load. In this study, the HVCE posterior weighting was employed to carry out a kinematic relative Multi-GNSS positioning experiment with six short-baselines from day of year (DoY) 171 to 200 in 2019. As a result, the HVCE posterior weighting strategy improved Multi-GNSS positioning accuracy by 20.5%, 15.7% and 13.2% in east-north-up (ENU) components, compared to an elevation-dependent (ED) priori weighting strategy. We observed that the weight proportion of both code and phase observations for each GNSS constellation were consistent during the entire 30 days, which indicates that the weight proportions of both code and phase observations are stable over a long period of time. It was also found that the quality of a phase observation is almost equivalent in each baseline and GNSS constellation, whereas that of a code observation is different. In order to reduce the time consumption of the HVCE method without sacrificing positioning accuracy, the stable variances of unit weights of both phase and code observations obtained over 30 days were averaged and then frozen as a priori information in the positioning experiment. The result demonstrated similar ENU improvements of 20.0%, 14.1% and 11.1% with respect to the ED method but saving 88% of the computation time of the HCVE strategy. Our study concludes with the observations that the frozen variances of unit weight (FVUW) could be applied to the positioning experiment for the next 30 days, that is, from DoY 201 to 230 in 2019, improving the positioning ENU accuracy of the ED method by 18.1%, 13.2% and 10.6%, indicating the effectiveness of the FVUW.

scholarly journals ON THE EVALUATION OF GNSS COMPLEMENTARY BY USING QUASIZENITH SATELLITE OF JAPAN

2016 ◽  
Vol XLI-B1 ◽  
pp. 495-502
Author(s):  
N. Sekiguchi ◽  
M. Shikada ◽  
T. Kanai
Keyword(s):  
Urban Area ◽  
Tall Buildings ◽  
Positional Information ◽  
Satellite System ◽  
Japanese Government ◽  
Positioning Accuracy ◽  
Satellite Positioning

The positional information has an important role in our lifestyle. People need to get positional information by GNSS. The satellite positioning must receive a signal from four or more satellites, however, most of Japanese country is covered with mountain and urban area has a lot of tall buildings. Then Japanese government launched QZS (Quasi Zenith Satellite) which is the first satellite of QZSS (Quasi Zenith Satellite System) in 2010. QZSS including QZS can improve positioning accuracy and reliability. QZS has 6 signals by using four kinds of frequency. These signals are the same frequency of GPS and GLONASS and so on. This paper was reported about the comparison of the positioning between GPS and QZSS.

scholarly journals An SVM Based Weight Scheme for Improving Kinematic GNSS Positioning Accuracy with Low-Cost GNSS Receiver in Urban Environments

Sensors ◽  
2020 ◽  
Vol 20 (24) ◽  
pp. 7265
Author(s):  
Zhitao Lyu ◽  
Yang Gao
Keyword(s):  
Real Time ◽  
Low Cost ◽  
Urban Environments ◽  
Least Square ◽  
Single Frequency ◽  
Line Of Sight ◽  
Support Vector ◽  
Positioning Accuracy ◽  
Gnss Receiver ◽  
Gnss Positioning

High-precision positioning with low-cost global navigation satellite systems (GNSS) in urban environments remains a significant challenge due to the significant multipath effects, non-line-of-sight (NLOS) errors, as well as poor satellite visibility and geometry. A GNSS system is typically implemented with a least-square (LS) or a Kalman-filter (KF) estimator, and a proper weight scheme is vital for achieving reliable navigation solutions. The traditional weight schemes are based on the signal-in-space ranging errors (SISRE), elevation and C/N0 values, which would be less effective in urban environments since the observation quality cannot be fully manifested by those values. In this paper, we propose a new multi-feature support vector machine (SVM) signal classifier-based weight scheme for GNSS measurements to improve the kinematic GNSS positioning accuracy in urban environments. The proposed new weight scheme is based on the identification of important features in GNSS data in urban environments and intelligent classification of line-of-sight (LOS) and NLOS signals. To validate the performance of the newly proposed weight scheme, we have implemented it into a real-time single-frequency precise point positioning (SFPPP) system. The dynamic vehicle-based tests with a low-cost single-frequency u-blox M8T GNSS receiver demonstrate that the positioning accuracy using the new weight scheme outperforms the traditional C/N0 based weight model by 65.4% and 85.0% in the horizontal and up direction, and most position error spikes at overcrossing and short tunnels can be eliminated by the new weight scheme compared to the traditional method. It also surpasses the built-in satellite-based augmentation systems (SBAS) solutions of the u-blox M8T and is even better than the built-in real-time-kinematic (RTK) solutions of multi-frequency receivers like the u-blox F9P and Trimble BD982.

Sign in / Sign up

Export Citation Format

Share Document