Deep learning has become popular and the mainstream technology in many researches related to learning, and has shown its impact on photogrammetry. According to the definition of photogrammetry, that is, a subject that researches shapes, locations, sizes, characteristics and inter-relationships of real objects from optical images, photogrammetry considers two aspects, geometry and semantics. From the two aspects, we review the history of deep learning and discuss its current applications on photogrammetry, and forecast the future development of photogrammetry. In geometry, the deep convolutional neural network (CNN) has been widely applied in stereo matching, SLAM and 3D reconstruction, and has made some effects but needs more improvement. In semantics, conventional methods that have to design empirical and handcrafted features have failed to extract the semantic information accurately and failed to produce types of “semantic thematic map” as 4D productions (DEM, DOM, DLG, DRG) of photogrammetry. This causes the semantic part of photogrammetry be ignored for a long time. The powerful generalization capacity, ability to fit any functions and stability under types of situations of deep leaning is making the automatic production of thematic maps possible. We review the achievements that have been obtained in road network extraction, building detection and crop classification, etc., and forecast that producing high-accuracy semantic thematic maps directly from optical images will become reality and these maps will become a type of standard products of photogrammetry. At last, we introduce our two current researches related to geometry and semantics respectively. One is stereo matching of aerial images based on deep learning and transfer learning; the other is precise crop classification from satellite spatio-temporal images based on 3D CNN.

Territorial water is a significant part of national sovereignty of China,thus the infrastructures of national space datum and location services should cover the sea areas except for the land areas.China has established relatively complete geodetic coordinate frame in land areas over the past decades,including the lastest developed China Geodetic Coordinate System 2000 (CGCS 2000) with its reference frame and the national gravity datum 2000.However,the currently used geodetic infrastructures have not well covered the sea areas of China.The marine geodetic datum and marine navigation technologies need to be further developed and extended to satisfy the national demands of marine environment and resources detection,scientific investigation as well as marine economy development in new era of China.This paper mainly reviews the development and the progress of Chinese marine geodetic datum and marine navigation,analyses related key technologies in establishing our national marine geodetic datum.Some current trends and future directions for independently developing our national marine geodetic datum and marine navigation technologies are discussed.

The Satellite Positioning and Orbit Determination System (SPODS) is a software package for GNSS positioning/orbit determination, developed by the Xi’an Research Institute of Surveying and Mapping. So far it has been able to analyse GPS data and has the capability of high precision GPS positioning and orbit determination. The underlying theory and the performance evaluation are briefly addressed in this paper. The experiments are carried out with GPS data collected from about 127 IGS stations during 4~10 January 2009. The results show that the RMS 1D difference is 1.1cm between SPODS orbits and final IGS combined orbits, and that the repeatability of daily solutions of station coordinates is 1.5mm for horizontal components, and 4.5mm for vertical component, and that the consistency of ERP solutions with IGS final products is 0.025mas, 0.093 mas and 0.013ms/d respectively for pole coordinates and LOD.

Firstly, the relationship between the accuracy of low altitude aerial photogrammetry and the field angle of camera is made by a quantitative analysis from the theory. The conclusion that the low altitude photogrammetry should use wide-angle camera as much as possible is done. Then, the limitation of the single lens camera to expand field angle and the combined wide-angle camera existing on the market not suitable for light load of low altitude UAV(Unmanned Aerial Vehicle) due to excessive weight are pointed out. The characteristics of combined wide-angle low altitude light camera with self-calibration and self-stabilization developed by the author are described, especially the principle of self-calibration for the combination of static error and dynamic error. Based on the practice of large scale mapping, a technical procedure in aerial photography by taking with wide-angle camera and large overlap simultaneously for improving the accuracy of low altitude photogrammetry is proposed. The typical engineering produced data is used to verity the above theoretical analysis. A technical route for increasing accuracy of low altitude photogrammetry with combined wide-angle camera is expounded.

The affection caused by the colored noises should be taken into account to the adjustment model.As useful signals,these colored noises should be accurately identified and extracted by Fourier analysis.A continuous adjustment model is introduced with respect to the colored noises,and then it can be generalized from the finite space to the infinite space so called as Hilbert space.This extension is to provide a new technique to perform the continuous observational system design,Fourier analysis as well as the parameter estimation.It shows that the Gramer’s determinant provides maximization criteria in the system optimization design as well as a rule in diagnosing the adjustment model.Related with the definition of the integral,the least squares solution of the continuous adjustment model becomes the limit of the traditional least squares solution in finite space.Moreover,the influence caused by the colored noises is systematic,but it can be eliminated or compensated by optimally designing the observational system.

Currently, the broadcast ephemerides used in GEOs are same as those of the MEOs and IGSOs in the BeiDou navigation constellation. However, a trade-off strategy, i.e. an orbital inclination of 5° rotation, is needed in the fitting algorithm to solve the ephemeris parameters as well as the user satellite position computation for GEOs. Based on the standard broadcast ephemerides, the representations of both the orbit and its perturbation were revised according to the second class of nonsingular orbital elements. In this research, a 16-parameter broadcast ephemeris is presented specifically for GEOs,and user satellite position computation formulas were derived correspondingly. Fit simulations show that the root of mean squares (RMS) of user range error (URE) with two hour and three hour data sets are better than 0.05m and 0.1m, respectively.

Block adjustment for satellite images cannot be solved with weak convergence geometric conditions,therefore a plane block adjustment method to improve the targeting precision of images is proposed utilizing DEM as height constraint plane block adjustment method. First,a rational function model with affine transformation is selected as the mathematical model of the satellite image plane block adjustment. Second,to update the ground coordinates of tie points(TPs),the plane coordinates of TPs are only solved in the adjustment process. Elevation values are obtained by using DEM interpolation. Finally,the plane coordinates of all TPs and orientation parameters of all satellite images are solved through plane block adjustment with a few ground control points ZY-3 nadir images for two regions are tested for plane block adjustment while ZY-3 forward-nadir-back images of the same two regions are tested for stereo block adjustment. A comparison indicates that almost the same accuracy can be obtained with plane block adjustment support using a 1∶50000 DEM and stereo block adjustment for ZY-3 images. For ZY-3 nadir images,almost no loss of plane block adjustment accuracy occurred when global DEM with 1km grid and SRTM with 90 m grid replaced the 1∶50000 DEM as elevation control,. Test results demonstrate the effectiveness and feasibility of the plane block adjustment method.

The development of new aeronautics and astronautics technologies has been constrained by strict mathematical rules for data processing among the diverse methods used to obtain spatial information.The acquisition of spatial information has been affected by various choices including the applied technologies (e.g.,push broom sensors),techniques (e.g.,zoom imaging),and equipment settings (e.g.,swing angle,aerial platform attitude,camera angle) in terms of the convergence,efficiency,and accuracy of the data.Based on the principle of the bionic machine parallax angle and pyramidal projection of the aerial space platform to the surface,this study explored solutions for high-resolution image sparsity,ill-conditioned singularity,and non-convergence by building a set of mathematical models to process the polar coordinates of the parallax angular vector.This study also formed a polar information theory for initial spatial information.This method improved the ranges of accuracy,efficiency,and anti-interference in close-range photogrammetry and the free net bundle adjustment model by several orders of magnitude.The open source code was made globally available more than 3 years ago,and has received positive reactions.The method’s effectiveness was verified using aerophotogrammetry and absolute network adjustment model experiments,and its performance was better than that of the Cartesian coordinate processing method.Finally,the higher-order solution characteristics of various applications and spaceflight platforms were provided,which are expected to provide a foundation for construction of a new polar coordinate system for aerospace multi-scale all-attitude spatial information acquisition,organization,management,storage,processing,and application.