2015 Vol. 40, No. 1
The rapid development of network technology and the semantic web has hastened the development of spatial information sharing and application services. However，the functions provided by the existing spatial information service models still cannot meet the needs of some complex applicadons，and how to achieve a combination of on-demand services is becoming a hot research issue. Currently，multiple services are combined in a semi-automatic or automatic manner to meet various applicanon requirements but low time efficiency remains a common problem.This study surveyed the advantages and disadvantages of related service composition methods，and developed an intelligent approach based on neural network theory to correctly and efficiently combine some commonly used spatial information processing services. This method adopts an extended ontology representation model for ad hoc services to provide semantic description，and utilizes the principle of synaptic neural net-work for searching core services. Finally，service composition is realized by an integrated indexing function. Experiments show that the method not only effectively improved the computation efficiency，but also lowered the barriers for intelligent service composition for non-professionals，confirming its practical value in handling various application requirements.
In view of the geographic annotation evaluation problem under the network map service environment，this paper extended the content of traditional geography annotation with the unstructured text form，and defined a broad geographic annotation. Using the putting process to model the annotation behavior，first of all，need confirm the geographic annotation text classification according to the frequency class，and establish a quantitative convergence description about the spatial correlation in the putting neighborhood according to the Voronoi k order adjacency relationship，proceed to the next step，build a putting model based on the geographical annotation types and type transformation combined with the geographic annotation text and the existing geographical annotation neighborhood. Through experimental verification，in the actual case of already known two typological geographic annotation sets，putting model can effectively evaluate the newly increased geographic annotation in a reasonable way.
As an important geometric feature，the uncertainty of polygon is the extension and the depth of line uncertainty. As for this problem，the uncertainty estimation of three-dimensional polygon is studied here，and based on the confidence domain model of line，the uncertainty of three-dimensional polygon is described and estimated from two aspects of the model and the measurement of confidence domain.And then，The acceptance specification for urban construction is chosen to prove that the research of this article is correct.Several conclusions can be got as follows:①The uncertainty model of three-dimensional polygon is proposed here，referring to the derivation of line.②The uncertainty model of three-dimensional polygon is estimated based on integral calculus.③The uncertainty estimation model of three-dimensional polygon can be used to control quality in the field of engineering.
Severe atmosphere，optic，and other negative effects will result in low brightness and contract problem and there makes remote sensing image into low quality. In this paper，two kinds of algorithms based on human eye feature are analyzed with their advantages and limits.A novel optimized Retinex approach is proposed. It fuses Retinex theory and image enhancement algorithm that strengthens brightness and contrast via a color space transform. Brightness and contrast are shifted with additional image edge features while holding image hue being constant. The results from image enhancement can be more comfortable for human eye features，provide significant improvement in brightness and contrast，delivers richer image information，and avoids cross color phenomenon. The experimental data resource was a low-light-level image processed to illustrate the efficiency of our method via fineness，the hue bias exponent，entropy，and several other indexes.
The High Resolution Stereo Camera(HRSC) on the European spacecraft Mars Express has high spatial resolution and global coverage. Thus，global high resolution topographic data from the Martian surface can be derived through photogrammetric processing of HRSC data.，butthe accuracy and efficiency of image matching can be improved with geometric constraints of the epipolar line. A approximate epipolar resampling method which is fit for HRSC linear pushbroom imagery is proposed. Firstly，the epipolar geometry of HRSC image is analyzed using a projection trajectory method.Then，the original HRSC images were differentially rectified into a normalized image. Next，the epipolar geometry in the normalized image is analyzed and the approximate epipolar line direction determined. Lastly，the epipolar image is generated through epipolar resampling in normalized image. Experimental results demonstrate the feasibility of the proposed methodThe parallax in the y direction is less than one pixel in the epipolar resampled image.Moreover，the geometric distortion is decreased when image matching is performed on normalized images.
SIFT descriptor is widely used for local feature extraction. However，some problems such as large numbers of extracted key points and its high dimension appear when using SIFT to extract local features from remote sensing imagery directly. To solve these problems and improve the retrieval results，we use a visual attention model to extract objects using their saliency from remote sensing images. The visual attention model is used to extract salient objects through their saliency from remote sensing images firstly，then we use a K-means algorithm to cluster local features，these results are then used as feature vectors for similarity measures. Some experimental results show that our method not only decreases the number of key points and the dimension of local features，but also improves retrieval results at the same time. It also accords with the human visual system.
In order to retrieve color images of a salient object，especially for those images with simple events and scenes，such as trademark images and landmark images，a new image retrieval method with combinational features is proposed.On the basis of the analysis of spectral energy distribution model，the color boundary is used to build a histogram of contour and color invariants to extract a descriptor of key points，thus to present a kind of feature retrieval method combining contour and local features，called CIFO. Experiments on sets of images were carried out by applying CIFO.Experimental results show that，with regard to the retrieving of color images with salient object，the CIFO method has obvious advantages over the other three classical feature-based retrieval methods.
High-precision surface elevation information can be obtained efficiently by airborne InSAR systems. Airborne InSAR systems have been an important means for remote sensing data acquisition and mapping in the areas where surveying and mapping is currently difficult.A three dimensional positioning and error model for airborne InSAR block adjustment were researched.Observations，weights determination and adjustment parameters were analyzed.In an experiment，76 airborne In-SAR images at 0.5 m resolution were used generate a block of 4 strips by 19 rows.The results show that the positional accuracy of a DEM and DOM at scale of 1:10 000 were achieved for large areas with sparse control points.
In order to solve the problems like low corresponding points matching rate，poor automatic registration process of multi-source optical satellite imagery in geometric correction. We described the principal direction and feature vector by phase congruency instead of the original SIFT algorithm gray level of pixel gradient to ascending accuracy of feature description.Optimizing the matching results with constraints of phase analysis results under the frequency domain，inhibition of error matching points;meanwhile，put forward a adaptive selection strategy to improve parameter estimation phase level of automation RANSAC-based;the ultimate realization of automatic image registration between the multi-source optical satellite imagery. The experimental results of multiple sets of data show that the validity and applicability of this schema to image registration between multi-source optical satellite image with radiation and nonlinear distortion.
In order to realize automatic nondestructive testing for surface cable damage on a cable-stayed bridge，a distributed machine vision system was developed. It uses four cameras to acquire images around the cable surface. Surface defection may be distributed in several images. An improved scale invariant feature transform(SIFT) feature matching algorithm for image mosaicing is proposed to real time processing to obtain a whole defect effectively. First，feature points are extracted by a Harris operator. Second，according to defect images collected by the system，the steps of the SIFT operator such as the distribution of the main direction for the matching feature points and the matcking image rotation is simplified. The simplified SIFT operator is employed to describe the feature points and match the images. Finally，image fusion is implemented and a complete image of a defect is obtained. Experimental results show that the algorithm complexity is greatly reduced and improves detection integrity for surface cable defects using our improved SIFT to automatically stitch the defect images together.
An image fusion algorithm based on gradient pyramid is proposed which takes advantage of image characteristics of underwater sonar environment .First，the gradient pyramid transform was imposed on registrated sonar images .Second，an activity and match measure were constructed based on the local neighborhood of the images. The local neighborhood feature of images was considered as a fusion strategy. Finally，the fusion image was obtained through synthesis module and multi-scale inverse transform. Experiments revealed that the new algorithm is effective under objective evaluation criteria in the absence of a standard reference image.
Atmospheric precipitable water vapor is an important element of the meteorological and meteorological disaster forecast. Using the ECMWF data for regional meteorological research is one of the current research focuses.For the problem caused by the low spatial resolution of the ECMWF data in studying the small regional meteorologicala，we studied the ECMWF water vapor interpolation idea with the Stretched Boundary Layer，and we established the model and realized it.This model can be used to interpolate the ECMWF water vapor，surface pressure and surface temperature. We can obtain the results in any density based on region DEM resolution. The model is tested with the UPS water vapor observations and surface meteorological data from the UPS-IPW network of the U.S.The results show that the model can improve the accuracy greatly on the ECMWF water vapor，surface pressure and surface temperature compared to the raw ECMWF data.
A detection method for ocean surface currents based on the Doppler shift and Bragg scattering for sea clutter imaging using X-band marine radar is addressed in this research. A preprocessing algorithm for current inversion combined with geometric filtering model is presented. Maximun current velocity is not required for band-pass filter processing，since accurate current velocity can be obtained only with liner wave theory dispersion equation and a least-squares algorithm. As compared with the conventional flow detection method using X-band navigation radar，this new algorithm is simple and efficient.It uses resources effectively,and can be applied to both shore-based X-band and shipborne X-band navigation radar. A validation comparison with sea surface current data from a near-by HFUWR shows that this algorithm effectively improves the accuracy and stability of the radar.
The 3D pose estimation of space object is a significant aspect of space situational awareness.We proposed a method to estimate the 3D pose of satellite as well as its sub-components based on as tronomical observed images.Firstly，a multi-viewpoints pose model database was established by use of equal-area subdivision grid. Secondly，the image restoration and filtering were operated on the observed image with the use of maximum likelihood algorithm based on mixed noises，and then the segment and feature abstraction were performed. Thirdly，a scale-invariant feature indexing method was proposed to encode the geometric features of satellite. Finally，the minimum distance match was used to estimate the 3D pose of space object among the pose database. The experiment result verifies the validity and correction of proposed method.
The 3D reconstruction of building facade can be used for a wide range of applications such as urban planning，traffic navigation and virtual reality. The extraction of the geometric features(facades，windows and doors profiles) is a key step in the three-dimensional reconstruction of a building facade. In this paper，the planar structure of the building facade is first detected from point cloud data based on region growing and principal component analysis (PCA).Then，the boundaries of geometric features are defined by TIN contour extraction combined with a convex hull algorithm. Experiments show that the proposed method can effectively extract geometric features of the building facade from variable density point cloud data.
This paper uses the characteristics of light spots to determine the uncertainty of laser in the spot，which introduces error entropy into the uncertainty of laser points. Uncertainty in points the laser point density function is determined based on the characteristic of lasers. Laser point information entropy is derived based on the information definition，taking advantage of the relationship between the information entropy and the error entropy to derive laser point error entropy. Error entropy is the linear relationship of spot size based on the error entropy expression. Point cloud error entropy andaverage error entropy of each point was acquired according to point cloud spot actual area. The influence of incident angle on the uncertainty of laser points was analyzed by the relationship of incident angle and error entropy，and the best incident angle of scan was determined. A feasibility of point cloud uncertainty by error entropy was verified according to the analysis of point clouds by setting different scan intervals.
Using the SNR observations from geodetic UPS receiver to monitoring the soil moisture is a new method，which does not suffer from destruction of observed soils，difficult data assimilation and time-spatial resolution limits. However，there are issues need to be solved such as uncertain measuring area，selection rules of parameter as wavelength and satellites，and construction of relationshipmodel between the relative delay phase and soil water content. To solve these problems，the Fresnel zone theory is introduced，and two comparative experiments based on simulated and measured soil moisture are carried out. The process and results of experiments show that the remarkable correlation，which can be described well by exponential function，does exist between the relative delay phase and soil moisture. The maximal effective monitoring distance is within about 45 meters. Meanwhile，selection of an advanced satellites and recording of L2C can increasingly improve the quality of SNR and lead to more reliable results.
Based on the UNSS data processing results from the Sichuan area，following conclusions are obtained:①Before the Ms 8.0 earthquake in Wenchuan，there was almost no dextral strike-slip activity distinguished or compressional deformation on the Longmenshan fault zone and nearly 150 km north of it，and was deformation distinguished in the area of 100 kms west of Sichuan Basin-- the Xianshuihe Anninghe Fault.②Between the Ms 8. 0 earthquake in Wenchuan and Ms 7.0 earthquake in Lushan，there was about a 5 mm/a dextral strike-slip activity in the middle and northern sections of the Longmenshan fault zone，with no 3D deformation distinguished in and near the seismic source region. Meanwhile the sinistral deformation increased on the Xianshuihe Anninghe fault and east of the zone.③The co-seismic deformation field of the Lushan earthquake was basically distributed for 10 kilometers around the seismic source region，with an epicenter bounded，on the east side as a dextral deformation，on the west side with a sinistral deformation.④LS05 was about 12 km to the epicenter; the permanent vertical displacement momentum was around 7 cm，with a horizontal thrust momentum of about 4 cm，the horizontal strike-slip was about 5 cm from the station.
One of key issues in the aspect of UPS-based attitude determination that must be solved is to estimate the unknown integer ambiguities. Single-epoch attitude ambiguity resolution can be achieved by using MC-Lambda method，which fully exploited the known geometry constraints of the multi-an-tennae configuration. This method does not need to consider cycle slip problem. However，the float solution precision of integer ambiguity is low based on UPS pseudo-range and carrier phase observations，which leads to large ambiguity search space and low search efficiency. For this reason，choosing the direction cosine matrix to describe the attitude and UPS/Uyro attitude determination model was established. The float solution of integer ambiguity was calculated by matrix kalman filter algorithm，and then the fixed solution of integer ambiguity was obtained by MC-Lambda method. Simulation experimental results showed that the accuracy of float solution by Kalman filter algorithm with constraints of direction cosine matrix are improved，so the computational efficiency and the fixing success rate of the fixed integer ambiguity are all improved，particularly when UPS observations is in the bad observation condition.
In order to improve the accuracy of earth gravity field measurements，a high order term for analytical continuation must be considered in practical applications. The high order term for analytical continuation solution was derived using the direct derivative method and iterative derivative method，which were compared with the traditional recursion method. The general recursion formula is obtained using the iterative derivative method. The formula shows that the classical recursion method has the same form as iterative derivative method after ignoring the little quantity term. An analysis also showsthat ananalytical continuation solution can be induced into the solution of different order derivatives for gravity anomalies along the plumb direction. This analytical continuation experiment was applied to the 50×50 area. and a n，term，and n2 term were calculated in this experiment. The reason for this difference may stem from computation error in the integral operation.
Regression analysis is often used in deformation analysis modeling，however，it is a static model. When the structure or physical properties of the monitoring object change，the static model will no longer be applicable. However，a varying coefficient regression model is a dynamic model and has greater flexibility and adaptability. So varying coefficient regression is introduced into dam deformation analysis. Local linear estimation is used to fit the varying coefficients. Both a simulation experiment and a real dam deformation monitoring data experiment show that a dam deformation model using varying coefficient regression is better than that using ordinary linear regression;and more precise when predicting deformation.