3 April 2020 in IEEE Access
Road markings are one of the most important safety elements in a road network, and they play a critical role in traffic safety. However, the automatic extraction of road markings remains a technical challenge in the fields of smart city construction and automatic driving. This paper presents an image-translation-based method of obtaining the 3D vectors of typical road markings from mobile laser point clouds. First, ground roughness is used as a criterion to extract ground points based on the topological relationship of adjacent scan lines, and the feature images of a road surface are generated using the adapted inverse distance weighted method. Second, by comparing objective functions based on the pix2pix framework, a finely adjusted image-to-image translation model named P2P_L1 is proposed for the segmentation of road markings. The proposed model outperforms the advanced DeepLab V3+ network in terms of precision, F1-score, and mean Intersection over Union indicators in the comparative segmentation results of ten types of road markings in the Shenzhen test area. Third, methods such as node averaging and optimized iterative closest point are developed for the 3D vectorization of road markings. This study presents a new approach for the automatic extraction of road markings to provide effective technical support for the construction of smart cities.
14 February 2022 in IEEE Geoscience and Remote Sensing Letters
The Ziyuan3-03 (ZY3-03) satellite is an optical stereo mapping satellite that is able to synchronously provide triple linear-array stereo images and discrete laser altimetry points (LAPs) for plotting 1:50 000-scale maps. Considering that the relative horizontal error between the ZY3-03 nadir image and synchronously acquired LAPs is very small, this study designed a technical method to obtain accurate image coordinates of LAPs on stereo images, and this method successfully overcomes the shortcomings of previous methods. Subsequently, a combined block adjustment model of stereo images and LAPs was constructed. On this basis, this study proposed a combined block adjustment method using ZY3-03 stereo images and LAPs to improve the vertical accuracy of images. For our research, 19 ZY3-03 triple linear-array stereo images and 87 synchronously acquired LAPs of plain, hilly, and mountainous terrains located in China were selected to conduct the block adjustment experiment. Results indicated that the vertical root-mean-square error of the experimental images decreased from 5.27 to 2.58 m for plain and hilly terrains and decreased from 11.25 to 4.45 m for mountainous terrain. Moreover, the vertical accuracy of the experimental images met the requirement of 1:50 000-scale mapping and also exceeded the accuracy of combined adjustment with Ice, Cloud, and land Elevation Satellite (ICESat)-1 satellite LAPs and the same experimental images.
16 December 2021 in IEEE Geoscience and Remote Sensing Letters
Improving the accuracy of block adjustment with few or no ground control points (GCPs) is one of the core issues for satellite imagery high-precision mapping. The GaoFen-7 (GF-7) satellite laser altimeter system is the first Chinese formal spaceborne laser altimeter system, which is equipped with laser footprint cameras for the first time. Given the very high height accuracy and high planimetric accuracy of the instrument, this letter proposed an adjustment method of ZiYuan-3 (ZY-3) satellite imagery by integrating GF-7 laser altimetry data. The laser control points (LCPs) were automatically extracted according to the registration of the footprint images and stereo images. Furthermore, the combined adjustment was performed with the LCPs as the vertical control and horizontal constraint. The performance of the proposed method was evaluated by using 417 stereo scenes of ZY-3 images and 954 LCPs, covering an area of 158 000 km 2 in Shandong, China. The results demonstrated that, only by using all LCPs as the vertical control, the root mean square error (RMSE) of elevation can be significantly improved from the original 8.82 to 1.40 m. Moreover, the planimetric RMSE decreased from the original 15.31 to 3.58 m with residual randomness when 45 LCPs were used as the horizontal constraint. Our method provides an alternative solution for satellite imagery geopositioning improvement without GCPs.
13 December 2023 in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Lu Tan-1 (LT-1) has finished the in-orbit test and will be delivered to the users in 2023. In this paper, we discussed the satellite characteristics including the satellite mission, the platform as well as the payload. The satellite constellation was designed following two main demands, i.e., digital elevation model (DEM) generation and deformation monitoring. Thus, we provided two kinds of productions. First was the DEM created using the images provided by helix bistatic formation. Second was the deformation products created using that of pursuit monostatic formation. The deformation products were finally classified into three different types given the processing technologies. They were deformation field provided by differential interferometric SAR (InSAR, DInSAR), deformation velocity field provided by stacking, and deformation time series provided by multi-temporal InSAR. DEM data were validated to be better than 5 m compared to ICESat. Deformation field, deformation velocity field, as well as deformation time series were validated to be better than 2.7 mm, 8.6 mm/yr and 3.7 mm, respectively.