The location prediction technology can predict the location of the user at the next moment in advance, and plays an extremely important role in the field of Location-based Service (LBS).Most of the existing location prediction techniques only use the geographical location information and time information of the user's historical trajectory. The geographic trajectory is composed of a series of geographically-pointed time-stamped latitude and longitude points, and the geographic trajectory only mines users. Mobile mode is limited by geographic features. In this paper, we propose a novel approach for predicting the next semantic location of a user's movement based on the geographic and semantic characteristics of the group user trajectory. The semantic location prediction based on group users generally consists of three steps: Firstly, the specific algorithm is used to identify the staying area in the user's trajectory; Next, the semantic matching algorithm is used to associate the user's staying area with the semantic information; Finally, Mining the semantic location pattern of group users, using this pattern to predict the semantic location of the user at the next moment. In the stage of staying area identification, in order to reduce the influence of indoor stay time unfixed on the recognition of stay area, this paper proposes a new type of spatial-temporal agglomerative nesting (ST-AGNES), which can automatically identify the number of staying areas in the user's trajectory using only the distance threshold. In the semantic matching stage, this paper proposes a semantic matching method based on attractance rules, which makes uses all trajectory points in the stay area to be associated with indoor high-density semantic information. In the final forecasting stage, this paper uses Long Short-Term Memory (LSTM) to mine the semantic location patterns of group users and predict the future semantic location of users. The experimental results have achieved a prediction accuracy rate of 61.3%.
The Certainty Factor (CF) is a classical method of sensitivity analysis for geological hazards, which can be used to quantify complex multi-factors in the same range, and CF value directly represents the contribution of each factor to geological hazards. The Support Vector Machine (SVM) is a representative method for machine learning, which can be used to evaluate the susceptibility of geological hazards on the basis of various impact factors. The main purpose of this paper is to combine CF method with SVM method called CF-SVM model for developing geological hazards susceptibility model which is applied to Lushui County of Yunnan Province. To assess geological hazards susceptibility, the paper selects ten impact factors, including the elevation, slope angle, slope aspect, profile curvature, distance to faults, distance to rivers, distance to roads, rock soil mass types, geomorphologic type, the land use, which were calculated as the most important impact factors. The state of each impact factor was graded based on the geological hazards area ratio curve and the grading area ratio curve. The CF method is applied to calculated the sensitivity values of each factor on the basis of 381 geological hazards, and the under-sampling method was used to select 381 non-geological hazards in Lushui county. The 381 geological hazards and 381 non-geological hazards hazards are the classification data of SVM method and geological hazards susceptibility maps were produced. Research shows that the extremely high and the high susceptibility areas are mainly distributed on the both sides of the Nujiang River. The comparision of geological hazards susceptibility map and actual geological hazards distribution showed that extremely high and the high susceptibility areas can account for about 89.76% of the actual geological hazards for CF-SVM model which is better than the single SVM model. The model performance was evaluated by the receiver operation characteristic (ROC) and precision recall (P-R) in the two models. Through ROC and P-R of the results, the prediction power of the CF-SVM model is superior to the single SVM model, and prediction accuracy of the CF-SVM model reached 92% and 88% respectively. This indicates that the combination of the CF and SVM has a higher predictive value. The research provides guidance and technical reference for the nationwide county geological hazards evaluation, and the proposed model can be used assess and manage the risk of geological hazard.
Dissected saddle, as an important terrain control point, is the result of the struggle statue between positive and negative terrains. The width of the dissected saddle ranges from only a few meters to about twenty meters, which represents the critical stage of gully capture. That is, the headward erosion of gullies on both sides of watersheds are going to erode and dissect the boundary of the watershed. Thus, the division of positive terrains and the connection of negative terrains are achieved. The typical dissected saddle is located in the loess landform in the Loess Plateau, also known as the loess dissected saddle. This dissected saddle could act as an important indicator for distinguishing the extent between loess interfluve area and loess gully area during the landform evolution process. In this paper, taking the typical loess landform as an example, and on a basis of DEM data and remote sensing images, the semi-automatic extraction of dissected saddles is conducted. The terrain characteristics of these extracted dissected saddles, i.e., slope, relief, depth of cut, were then calculated based on the DEM data. Moreover, the spatial pattern of the dissected saddle was summarized. The experimental results show that the dissected saddles are distributed at the boundary of the main stream and perpendicular to the widest part of the main channel, indicating an obvious terrain controlling effect. The quantity and distribution of dissected saddles determine the development and the shape of the watershed to some extent. The results of slope, relief, and depth of cut for dissected saddles are all larger than that for the normal saddles. At the same time, the value of the high-hierarchical watershed is greater than the value of the lower-hierarchical watershed, which reflects that the dissected saddle has characteristics of strong erosion and high surface fragmentation. In summary, the dissected saddle is highly eroded by the channel, which could help to demonstrate the development stage of the loess landform. Along with the development of the landform, the dissected saddle could be regarded as a symbol, indicating the development of the loess landform has reached the metaphase of the landform evolution process.
Ecological engineering such as Grain for Green Project have significant impacts on the structure of regional land use and ecosystem service functions. Based on the RUSLE model and RS & GIS spatial analysis methods, this study assessed the impacts of returning farmland on soil conservation function in the western region of Farming-pastoral Ecotone of Northern China (FPENC) during 2000-2015. The results showed that the total area of farmland in the FPENC decreased by 1663.83 km2 from 2000 to 2015, which was mainly converted into forest land, grassland and construction land. The implementation of the Grain for Green Program was the main reason for farmland decrease, and the area of farmland converted into forest and grassland accounted for 66.93% of the total area of farmland reduction. The newly added farmland was mostly converted from grassland and unused land, and mainly concentrated in the northern and central regions. Besides, the soil conservation function had improved significantly in the western region of FPENC during the past 15 years, and the amount of soil conservation increased by 56.50×104t, which mainly resulted from returning farmland to forest and grassland between 2005 and 2010. In addition, the increase in soil conservation caused by ecological restoration had obvious difference in different slope grades, but the increased soil conservation generally showed decrease trend with the slope increase. Nevertheless, in some areas of slope greater than 25 degrees implemented Grain for Green Project have high benefits of soil conservation. The steep slope (slope greater than 25 degrees) area is mostly a contiguous area of extreme poverty, where is the key area implemented by a new round of Grain for Green Project and poverty alleviation projects. This study about the impact mechanism of returning farmland on soil conservation function in western region of FPENC will provide quantitative scientific basis for the planning and construction of regional ecological protection and restoration projects.
Since the land economic density can only characterize the economic utilization efficiency of construction land, the lighting density of construction land is used to characterize the output efficiency of construction land. and the Kernel density analysis, ESDA and the SDE is used to analysis the spatial-temporal pattern of it. The results show that: (1) There is a strong correlation between nighttime lighting and construction land output. It is scientific to use the light density to characterize the output efficiency of construction land. (2)The efficiency is high in the east and low in the west. The output efficiency of construction land in each region continues to increase while the regional differences continue to increase. However, the average annual growth rate between regions is 0.56%, which is relatively balanced. (3)The kernel density curve indicating that the overall level of construction land output efficiency in China is low. and presents the trend converging to middle and low level club. (4) The overall Moran's I index of construction land output efficiency is positive, indicating that there is a positive spatial distribution characteristics of construction land output efficiency and the local spatial pattern changes little, there are stable and dynamic, strong and weak, and weak and strong. (5) The azimuth of construction land output efficiency is always between 72.420° and 81.066°, indicating the northeast-southwest direction is the main direction of the output of construction land, the main direction and the secondary direction of the construction land output benefit has dispersion.
Coastline is the boundary between land and ocean, and the coastline position determination is the important content of the coastal zone, island and reef surveying. Coastline is generally divided into the island coastline and the mainland coastline. Under the background of global warming and the influence of the natural environment and human's exploitation, the coastline has been in a state of changing. Grasping the type, location, changing process and the future trend of the coastline accurately has great significance for guiding the coastal aquaculture, coastal zone development, navigation and transportation. Therefore, accurately and quickly extracting coastline and real-time monitoring its changes has the vital significance. Remote sensing technique has the features of observing large area synchronously, timely, and economically, which makes it an excellent choice for coastline classification and extraction. Now, remote sensing methods used in coastline extraction mainly include optical remote sensing, microwave remote sensing and laser radar technology. Various methods have been presented by researchers all over the world in recent years. However, some methods focus on waterline extraction instead of the defined coastline extraction. So this paper give summarize of waterline extraction and coastline extraction separately. Beyond that, the noise-reduction methods applied in coastline extraction and the solution of the inconsistency of level data in tidal are also mentioned in this paper. Overall, the paper reviews the recent research progresses on coastline extraction by all kinds of ways at home and abroad through analyzing their advantages, disadvantages and adaptability, and introducing their applications in many fields. Finally, feasible suggestion of the future research is forecasted based on its existent insufficiency.
Human annotation is a massive labor cost for the training sample selection process when applying any kind of supervised learning algorithm for change detection based on high-resolution remotely sensed satellite images. It is limited and unreasonable to use just one single sort of classifier generated from a supervised algorithm to extract change information of variety from the time-series images both in completeness and accuracy, let alone the inevitable salt-and-pepper noise and tiny patches falsely detected which turn out to be ubiquitous in and out of geographical entities. To tackle with problems mentioned above, a change detection approach based on a new automatic training sample annotation strategy and an improved Adaboost ensemble learning algorithm was proposed. At first, the unsupervised change detection algorithm CVA was applied to generate a low-level change detection result as referencing labels for further annotation, then the low-level result was divided into several parts with different intervals to ensure the automatic acquisition of the evenly distributed training samples with confidence. Furthermore, decision stump, logistic regression and kNN were employed as the weak classifiers to construct a hybrid multi-classifiers ensemble system with the help of the improved Adaboost algorithm, which would effectively promote the classification accuracy and generalization capacity of weak classifiers by sufficiently mining and making use of the spatial information with potential values. Finally, the SLIC segmentation algorithm was implemented in the difference image, and the segmentation border information was combined with spatial contextual information to build up a dual-filter for spatial constraint aiming at decreasing the omission rate and the false alarm rate of the detection results. To verify the validity of the proposed method, we conducted experiments using two datasets of multispectral images collected by SPOT-5 and WorldView-2. Experimental results indicated that the proposed method would significantly lower the labor costs of training sample annotation and demonstrated superiority compared with four other methods in accuracy.
Land surface temperature is one of the important parameters of scientific research such as resource environment, climate change and terrestrial ecosystem. MODIS LST (Land Surface Temperature, LST) products are important data sources for land surface temperature related research. The land surface temperature information of MODIS LST products is lost in the cloud coverage area. Therefore, the land surface temperature estimation of cloud coverage areas has become a frontier research problem of thermal infrared remote sensing. In order to solve the problem of missing land surface temperature information in the cloud occlusion area of MODIS LST products. In this paper, the Qinling area is used as the research area and the experimental data of MOD11A2 from 2001 to 2017 is selected. In the traditional Inverse Distance Weighting (IDW), Regular Spline (SPLINE), Ordinary Kriging (OK) and Trend Surface (TREND) spatial interpolation method, the important influence factor of elevation is introduced. Through a large number of spatial interpolation experiments, the traditional spatial interpolation method is improved, and a MODIS LST spatial interpolation method based on DEM correction is formed. Analysis of spatial interpolation results indicates: （1） The spatial interpolation accuracy is from high to low: OK> SPLINE > IDW>TREND, and the accuracy of the OK, SPLINE, IDW, and TREND methods based on DEM correction is increased by about 0.38°C, 0.31°C, 0.32°C, and 0.78°C, respectively; (2) The accuracy of spatial interpolation results shows seasonal differences. The interpolation accuracy is higher in summer, July, and August, and the interpolation accuracy is the lowest in January. （3）The interpolation accuracy has a certain relationship with the cloud area. When the cloud coverage area is less than 1.1km2, the interpolation error of the DEM+OK interpolation method is less than 0.55°C, and when the cloud coverage area is less than 3.1km2, the spatial interpolation error is less than 1°C. When the cloud coverage area is less than 2.7 km2, the interpolation error of the DEM+SPLINE method is less than 0.55°C, and the interpolation error of the DEM+SPLINE method is less than 1°C when the cloud coverage area is less than 10.4 km2. When the cloud coverage is 1.1~2.7 km2, the interpolation accuracy of DEM+SPLINE interpolation method is higher than of the DEM+OK interpolation method.
Since Urban forests played important roles in improving air, water and land quality, absorbing and mitigating carbon dioxide and many pollutants, mitigating urban heat island and reducing storm water runoff, its monitoring is a major issue for urban planners. It is of great significance to obtain the tree species timely and precisely in urban planning and green space management. At present, urban forest tree species mapping has benefitted from advances in remote sensing techniques. Using an object-oriented method combing spectral, textural, indicial and geometric features from high-resolution WorldView-2 imagery, this paper aimed to carry out the classification of seven main tree species in Fuzhou university, including Banyan (Ficus microcarpa), Mango(Mangifera indica L.), Camphor tree (Cinnamomum camphora), Bishop wood (Bischofia polycarpa), Chinese orchid tree(Bauhinia purpurea L.), Weeping fig (Ficus benjamina L.), and Kapok tree (Bombax malabaricum DC.). A random forest method was employed to determine the feature selection in this study. When eliminating 20 percent of the total features, the in situ validation results showed that the overall accuracy reached a highest value of 74.95% with Kappa coefficient of 0.67 when using 34 features for classification, which including 15 spectral features, 6 textural features, 8 indicial features and 5 geometric features, and the feature of mean spectral was the most significant, however, the standard deviation of each band is less important. The results also revealed that the feature selection of random forest could reduce or avoid the data redundancy and Hughes phenomenon, and thus could improve the classification accuracy of same type tree species. Moreover, the four additional bands of WorldView-2 imagery, especially the yellow and red edge band, and their composite indexes showed a higher importance in classification, which also indicates that these bands have great application prospects in vegetation remote sensing, especially in tree species classification.
The rapid urban expansion has induced and aggravated the urban heat island phenomenon, which makes it a big challenge for human health and human survival environment. Research is needed to explore the impacts of urban form on the surface urban heat island. Taking 13 mega cities in China as the study area, this study mainly focuses on the relationship between urban forms and urban heat islands beside the traditional impact factors of surface urban heat islands. Using the MODIS land surface temperature products of the daytime and nighttime in summer 2015 (including June, July and August), along with the land cover, population, demographic and meteorological data of these 13 cities, the relationship between urban heat island and four factors, i.e. land covers composition, spatial configuration of land covers, population and location, were explored. Furthermore, the urban heat island intensity (UHII) index was employed to evaluate the urban heat island effect, which represents the mean LST difference between the urban region and the rural region. The results indicate that the urban heat island effect varies considerably among the 13 mega cities, showing a higher mean UHII in the daytime than that in the nighttime. The factors controlling annual mean daytime UHII are the area ratio of built-up area, the area ratio of forest, the mean patch area of built-up area, the mean patch area of forest, aggregation index of built-up area and population density. The nighttime UHII is significantly influenced by the mean patch area of built-up area, the mean patch area of forest, aggregation index of built-up area and the patch density of forest. Increasing the built-up area and the forest area will both increase UHII. Measures to mitigate the urban heat island include decreasing the built-up area or increasing green urban areas. Moreover, the urban heat island effect can be mitigated by altering the form of cities, such as, reducing the mean patch area of built-up area or reducing patch aggregation.
This paper presented a method of monthly net primary production (NPP) estimation of grassland in the Three-River Headwater Region (TRH) based on GF-1/WFV data. First, a preprocessing of radiometric calibration and atmospheric correction is applied on GF-1/WFV 1A data by ENVI software. Secondly, geocoding is processed by Rational Function Model (RFM) with GF-1/WFV RPC (Rational Polynomial Coefficient) and the orthophoto images with high georeferenced accuracy are conducted after block adjustment. The processed GF-1/WFV data is comparable in space and time. Then, cloud and cloud shadow per scene are detected using Multi-feature Combined method; NDVI is retrieved based on GF-1/WFV image and monthly NDVI is generated by Maximum Value Composite (MVC) method. The values of pixels still affected by cloud or cloud shadow cover in monthly NDVI mosaic are extrapolated using linear regression using least square method based on MODIS 13Q1 NDVI. Finally monthly NPP of grassland is calculated based on Carnegie-Ames-Stanford Approach (CASA) with monthly NDVI and other variables including monthly total precipitation, monthly averaged temperature and monthly total solar radiation. A case study was conducted in Maduo country and results showed that: (1) reliable monthly NDVI data at medium spatial resolution can be obtained based on GF-1/WFV under the support of MODIS 13Q1 product; (2) The accuracy of estimated grassland NPP based on GF-1/WFV was over 70% based on field data validation, which is better than MODIS 17A3 NPP production and the former can occupied more detailed NPP spatial variation. Monthly NPP can be successfully estimated based on GF-1/WFV under the support of MODIS 13Q1 product in TRH. However, some details need to be improved for further study: (1) more area of cloud and cloud shadow in images, lower precision of the extrapolated NDVI and the error of simulated NPP may be greater; (2) in low temperature, NPP is 0 in CASA, which overestimates the grassland NPP because underground root of grassland is still alive in TRH in winter and NPP should be negative; (3) monthly NDVI generated by MVC represents the best growth situation of vegetation in the period, not the average one, which may overestimates NPP. Besides, mapping accuracy of vegetation type will also affect the simulated NPP result precision; (4) field data collection is difficulty due to the study area is in remote area of high altitude, so the current ground data is not enough to cover all months in growth season and the uncertainty of this method remains to be further tested.
With the improvement of Beijing rail transportation, the subway has become an important transportation for people's daily travel. Monitoring and controlling land subsidence along metro line become an important basic work to ensure the normal operation of linear engineering. Based on 55 images of the 3m high-resolution TerraSAR-X data covering Beijing, this paper used multi-temporal InSAR analysis technology to obtain the ground settlement deformation information along the subway network from April 2010 to December 2016, and analyzed the spatial-temporal evolution of the ground settlement along the Beijing subway network. The spatial-temporal ground subsidence related to the construction of subway tunnels is usually modeled by peck formula in the space domain, which is used to calculate the ground surface settlement and determine the maximum settlement value of the settlement trough curve and the settlement trough width. Taking Ciqikou-Guangqumen station section as an example, the InSAR results were modeled by Peck formula, we estimated the spatial development characteristics of the ground subsidence trough. The results show that there are different degrees of the deformation along Beijing subway line, and the serious deformation is mainly concentrated in the eastern and northeastern regions, annual maximum subsidence rate greater than 100mm/a. Compared with other lines, the overall situation of lines 4 and 10 is relatively stable, followed by lines 14 and Yizhuang, and the uneven settlement of lines 6 and 7 is the most serious. In addition, the road sections show different deformation characteristics in different construction periods, and the settlement of the construction period is more serious than that of the operation period. The width and maximum settlement value of the settlement trough between Ciqikou and Guangqumen station (line 7) show an increasing trend from 2010 to 2016, the maximum settlement trough width reaches about 180 m.
As one of the climate change risk which profoundly affected the natural environment and human society, heat waves has aroused increasing attentions all around the world. Based on VSD （The Vulnerability Scoping Diagram), the evaluation index system of t heat waves risk is constructed. With the method of exploratory spatial data analysis, the spatial-temporal characteristic, hot spots evolution and spatial differentiation of heat waves risk in Fujian Province from 2000 to 2015 are carefully examined. The results show that: ① The index of heat waves risk is decreasing in Fujian Province, at the same time, the internal transitions among different risk levels are obvious . ② The spatial distribution of heat waves risk in Fujian Province has the “layer structure”, and the risk index varies from central to peripheral areas as characteristic of “low-high-low”. ③ The spatial agglomeration degree of heat waves risk decreases. The hot spots presents a tendency of shrinking, from “multi-core” to “dual core”, while the cold spots presents a tendency of stabilizing in the northeast. ④ The heat waves risk in Fujian Province can be divided into 5 types, including high risk area of capital-deputy provincial area, sub-high risk area of prefecture-level city district, medium risk area of river valley, sub-low risk area of coastal plain, and low risk area of eastern-inland mountainous area. This result is the prejudgment of spatial evolvement of heat waves risk of Fujian Province in the future. It also can provide references for risk management and public service facilities .
The monitoring of vegetation cover change is the basis of regional resource and environmental bearing capacity research. This paper estimates the vegetation of Yunnan Province from 2001 to 2016 by calculating the MODIS-NDVI vegetation index from 2001 to 2016, supplemented by trend analysis, and coefficient of variation. Next, the spatial and temporal variation characteristics of vegetation coverage and its distribution relationship with topographic factors are discussed in depth. Results are shown as follows: ① From 2001 to 2016, the vegetation coverage in Yunnan shows a significant increase, with a growth rate of 4.992%/10 a.② Spatially, the spatial pattern of vegetation coverage appears to be gradually decreasing from the south to the north and from the west to the east. The vegetation coverage is highest in the west and southwestern Yunnan and the lowest in the northwestern Yunnan. The stability of the vegetation coverage is characterized by increasing volatility from southwest to northeast; the increase of vegetation coverage in northeastern Yunnan was significantly better than other areas. The study region of the vegetation coverage change trend which was increased, basically stable and decreased, accounting for 49.53%, 43.76% and 6.71%, respectively.③ The area transfer matrix results of vegetation coverage in the three periods from 2001-2006, 2006-2011, and 2011-2016 all showed that the vegetation cover evolution area was larger than the degraded area, and the ratios of the two were 1.42, 1.63, and 2.0. It indicates that the vegetation coverage shows a continuous improvement trend in the study area. ④ The relationship between vegetation coverage and topographic factors in Yunnan Province shows that the average vegetation coverage increases first, then decreases, then increases, and then decreases with the increase in altitude; it increases first and then decreases with increasing slope; Changes have gradually decreased from north to south.
