Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS
The accurate spatial positioning of the target in a fixed camera image is a critical sensing technique. Conventional visual spatial positioning methods rely on tedious camera calibration and face great challenges in selecting the representative feature points to compute the position of the target, e...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-08-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/16/3877 |
| _version_ | 1797408117304066048 |
|---|---|
| author | Binbin Liang Songchen Han Wei Li Daoyong Fu Ruliang He Guoxin Huang |
| author_facet | Binbin Liang Songchen Han Wei Li Daoyong Fu Ruliang He Guoxin Huang |
| author_sort | Binbin Liang |
| collection | DOAJ |
| description | The accurate spatial positioning of the target in a fixed camera image is a critical sensing technique. Conventional visual spatial positioning methods rely on tedious camera calibration and face great challenges in selecting the representative feature points to compute the position of the target, especially when existing occlusion or in remote scenes. In order to avoid these deficiencies, this paper proposes a deep learning approach for accurate visual spatial positioning of the targets with the assistance of Global Navigation Satellite System (GNSS). It contains two stages: the first stage trains a hybrid supervised and unsupervised auto-encoder regression network offline to gain capability of regressing geolocation (longitude and latitude) directly from the fusion of image and GNSS, and learns an error scale factor to evaluate the regression error. The second stage firstly predicts regressed accurate geolocation online from the observed image and GNSS measurement, and then filters the predictive geolocation and the measured GNSS to output the optimal geolocation. The experimental results showed that the proposed approach increased the average positioning accuracy by 56.83%, 37.25%, 41.62% in a simulated scenario and 31.25%, 7.43%, 38.28% in a real-world scenario, compared with GNSS, the Interacting Multiple Model−Unscented Kalman Filters (IMM-UKF) and the supervised deep learning approach, respectively. Other improvements were also achieved in positioning stability, robustness, generalization, and performance in GNSS denied environments. |
| first_indexed | 2024-03-09T03:54:50Z |
| format | Article |
| id | doaj.art-b01e8bf5be0f4dcdad33d8e81c2c2bc2 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T03:54:50Z |
| publishDate | 2022-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-b01e8bf5be0f4dcdad33d8e81c2c2bc22023-12-03T14:23:37ZengMDPI AGRemote Sensing2072-42922022-08-011416387710.3390/rs14163877Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSSBinbin Liang0Songchen Han1Wei Li2Daoyong Fu3Ruliang He4Guoxin Huang5School of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, ChinaSchool of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, ChinaSchool of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, ChinaSchool of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, ChinaSchool of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, ChinaNational Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610065, ChinaThe accurate spatial positioning of the target in a fixed camera image is a critical sensing technique. Conventional visual spatial positioning methods rely on tedious camera calibration and face great challenges in selecting the representative feature points to compute the position of the target, especially when existing occlusion or in remote scenes. In order to avoid these deficiencies, this paper proposes a deep learning approach for accurate visual spatial positioning of the targets with the assistance of Global Navigation Satellite System (GNSS). It contains two stages: the first stage trains a hybrid supervised and unsupervised auto-encoder regression network offline to gain capability of regressing geolocation (longitude and latitude) directly from the fusion of image and GNSS, and learns an error scale factor to evaluate the regression error. The second stage firstly predicts regressed accurate geolocation online from the observed image and GNSS measurement, and then filters the predictive geolocation and the measured GNSS to output the optimal geolocation. The experimental results showed that the proposed approach increased the average positioning accuracy by 56.83%, 37.25%, 41.62% in a simulated scenario and 31.25%, 7.43%, 38.28% in a real-world scenario, compared with GNSS, the Interacting Multiple Model−Unscented Kalman Filters (IMM-UKF) and the supervised deep learning approach, respectively. Other improvements were also achieved in positioning stability, robustness, generalization, and performance in GNSS denied environments.https://www.mdpi.com/2072-4292/14/16/3877visual spatial positioninguncalibrated imageglobal navigation satellite systemmulti-sensor fusiondeep learning |
| spellingShingle | Binbin Liang Songchen Han Wei Li Daoyong Fu Ruliang He Guoxin Huang Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS Remote Sensing visual spatial positioning uncalibrated image global navigation satellite system multi-sensor fusion deep learning |
| title | Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS |
| title_full | Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS |
| title_fullStr | Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS |
| title_full_unstemmed | Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS |
| title_short | Accurate Spatial Positioning of Target Based on the Fusion of Uncalibrated Image and GNSS |
| title_sort | accurate spatial positioning of target based on the fusion of uncalibrated image and gnss |
| topic | visual spatial positioning uncalibrated image global navigation satellite system multi-sensor fusion deep learning |
| url | https://www.mdpi.com/2072-4292/14/16/3877 |
| work_keys_str_mv | AT binbinliang accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss AT songchenhan accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss AT weili accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss AT daoyongfu accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss AT rulianghe accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss AT guoxinhuang accuratespatialpositioningoftargetbasedonthefusionofuncalibratedimageandgnss |