The corpus of geographical entity relations is the basic data resource of geographical information acquisition and geographical knowledge services, and its scale directly affects the training effect of machine learning models. Fast-updated web text is constantly emerging as a new relational example, requiring the corpus to be updated in a timely manner to cover richer relational instances. Manually constructing and updating corpus are expensive. Therefore, it needs a more efficient technology of corpus construction for massive geographical entity relations. In this paper, we propose an efficient method of corpus construction for massive geographical entity relations through the automatic annotation technique. First of all, based on encyclopedia resources, referring geographical entity classification standard and semantic relation, spatial relation classification standard to establish an annotation scheme of geographical relation, which considers both the linguistic habits of natural language and the annotation normalization. Secondly, we combine the fully-matching with the approximate matching to improve the coverage rate of object entity finding. Thirdly, we define the rules of sentence scoring by using the optimal sequence diagram method, as well as quantitatively evaluate the results of mapping the seed triples to the sentences. Finally, a series of experiments based on the Chinese BaiduBaike are carried out, which is used to verify the effectiveness of the improved automatic annotation. The results show that, the average success rate of the automatic annotation is 67.83%, and the average accuracy of the annotated relations by our method is 76.36%. Comparing with the manually annotated corpus of the spatial relations, the proposed method constructed a large-scale corpus of geographical entity relations more efficiently, which provides a feasible scheme for expending geographical entity relations corpus automatic. Experimental results on self-built corpus by LSTM (Long Short Term Memory) network shows that the accuracy of geographical relation extracting from web texts is 73.2%, and the accuracy of relative corpora is 75.2%, which proofs that the corpus of geographical entity relations is available. At the same time, this method takes into account the semantic relationship and spatial relationship between geographical entities, and it can be used for open relation extraction task. Besides, the relation types are not limited, which can be applied to open relation extraction.
With the rapid development of mobile Internet and the wide application of location-based service technology in various industries, the public's demand for the application of place information is growing rapidly. The gazetteer query, which can provide the support for place names knowledge, is an important basic link in the location information service. At present, because of the significant increase of the data volume of the place names, the query performance of gazetteers is facing a severe challenge. Most of the existing gazetteers directly use general retrieval methods, ignoring the characteristics of the characters and the description rules of the place names themselves. In order to solve these problems, a Chinese gazetteer query method (CGQM) is proposed based on the character features of place names. The CGQM uses the character features of the names with the same character characteristics, character's number and character's position, and query the gazetteer according to the main line of "candidate place name query, place name filtering, place name similarity ranking". Firstly, the single character index of the gazetteer is constructed, and based on this index, the place names containing the same characters in the gazetteer are queried to form a candidate dataset. Secondly, the place names are filtered from the candidate dataset, which has large differences in the number of characters with the search place names. The aim of this step is to enhance the accuracy of the candidate dataset and to ensure the efficiency of the later sorting process. Thirdly, the candidate place names are sorted based on the algorithm of character position similarity. Taking the national Chinese gazetteer as an example, an experiment was implemented with CGQM and a full text query method (Lucene) on 5 test datasets. The purpose of the experiment was to verify that the CGQM method could accurately and efficiently query the gazetteer. The experimental performance evaluation indexes include the operation efficiency, the precision rate, the recall rate and the F value. The results of experiment prove that CGQM can achieve much more better query performance than the Lucene based method. In the future research on gazetteer query, we will also consider many other factors, such as glyph, semantics, etc., and learn from the distributed and multithreading techniques in the retrieval system at the same time. These methods will promote the accuracy and efficiency of gazetteer query and expand the public service of place information.
In this paper, we proposed an improved adaptive spatial point clustering algorithm based on minimum spanning tree (MSTAA in abbreviation) to solve the problems existed in the traditional clustering algorithms. The first problem of these classical clustering algorithms is that the noise edges are determined using the global invariant. Another one is that the initial clustering parameters such as edge weight tolerance, edge variation factor, the number of clusters and initial clustering centers are determined by the users empirically. Furthermore, these algorithms can't find the noise edges at the local level. Based on these problems mentioned above, the algorithm put forward in this article aims to overcome the influence of subjective factors by defining two clipping factors. These trimming factors do not need to be determined by the users and can be automatically obtained according to the statistical features of the side length in the minimum spanning tree. The detailed realization process is as follows. In the first place, the pruning operation on the minimum spanning tree from the global level is carried out, which can eliminate the noises in the global environment. After the first round of tailoring, the initial minimum spanning tree becomes sub-tree collections. In the second place, in order to eliminate the noises at the local level, the algorithm performs the second round of pruning operation by setting the adaptive local cutting factor in the light of the side length statistics of each sub-tree. After the above two rounds of cutting, the MSTAA algorithm will get the final clustering result. In order to validate the effectiveness of the algorithm, both a simulated data and a practical application are adopted. By comparing with 4 classical clustering algorithms (k-means, DBSCAN, SEMST, HEMST), we find that the improved algorithm presented in this paper could find clusters of arbitrary shape and density in the environment where no one provides experience parameters. At the same time, not only does the MSTAA algorithm eliminate the obvious global noise points, but also it can distinguish noise points at the local environment so as to ensure a high similarity degree of cluster point set. All of the features of the MSTAA algorithm mentioned above make it possible to automatically mine hidden information behind spatial point data.
The digital elevation model (DEM) is considered as a source of vital spatial information and is widely used in many fields. The ASTER GDEM and SRTM provide almost global coverage and offer practical elevation data for geography research. However, due to the differences of remote sensing mechanism, GDEM and SRTM datasets present different accuracies on same landform units. A novel elevation data fusion approach is proposed in this paper, which eliminates the impact of landform characteristics on two DEMs and significantly improves the quality of fused DEM. This method focuses on two steps, geo-referencing and elevation fusion. An objective function of errors representing the summary of horizontal shifts between two DEMs by referring to stream link pair is first proposed, and correspondingly a multilevel grid search method is suggested to calculate the optimal horizontal offsets between DEMs. Two geo-referenced DEMs are then fused using regression models over different landform units and moreover the elevations nearby the boundaries of two units are specifically treated using a weighed non-linear regression method. This approach was tested in the area of northern Huairou using a 1: 50 000 topographic map. The statistics show that: (1) the RMSE of fused DEM decreases significantly in all landform units, and the representation of terrain is more accurate than GDEM and SRTM; (2) The difference of the elevation points between fused DEM and referenced topographic map also illustrates a normal distribution, which is obviously different from the asymmetric multi-peak distributions of two raw DEMs, indicating that the topographic effects have been effectively eliminated; (3) The accuracy of fused DEM is superior to that of GDEM and SRTM under different slope ranges, meanwhile, the influence of slope factor on the elevation accuracy of DEM is obviously reduced after fusion; (4) The RMSEs of GDEM and SRTM vary greatly with different aspects, while the RMSE of fused DEM keeps homogenous in almost all aspects, and the elevation accuracy of the fused DEM is also significantly improved in comparison with GDEM and SRTM.
Social media plays a more and more important role in the real-time disaster information distribution and dissemination. During the disaster event, social media usually generates and contains a lot of real-time disaster loss information, which is very useful for the timely disaster response and disaster loss assessment. However, the social media data has many shortcomings, such as high fragmentation of the information, sparsity of the text features, and the lack of annotated corpus and so on, which makes the traditional supervised learning method difficult to be effectively used for disaster information extraction. This paper proposed a fast disaster loss identification and classification method to extract the disaster information from social media data by extending the context features and matching feature words. By this method, we firstly extracted the keywords from a small amount of sample micro-blog text of different disaster loss categories based on Chinese grammar rules and constructed the pairs of feature words collocation. Then, we used the word vector model and the existing lexicon to supplement and expand these pairs of feature words collocation. And the external corpus was introduced to optimize the semantic collocation relationship between feature words according to the rules of the concurrence of Chinese words. At last, we built a classification knowledgebase for identification and classification of disaster loss information related to typhoon disasters included in micro-blog. An experiment system was developed to evaluate the method introduced in the paper. Typhoon "Meranti" landed on 15th September, 2016 was selected as a case study. Results show that this method has a significant effect (each comprehensive evaluation index of different categories is greater than 0.74) on identifying and classifying different categories of disaster loss information from social media. We mapped the spatio-temporal distribution of typhoon influence based on the classification results of disaster loss from social media. The experiment shows that the classification output data and maps could be used for the disaster loss evaluation and mitigation.
The spatial distribution of population at fine-scale has increasingly become research hotspot and a difficulty issue in the field of population geography. It has practical application value and scientific significance for relevant researches, such as disaster assessment, resource allocation and construction of smart cities. The population is concentrated in the urban area. Revealing the population distribution difference in this area is the core content of spatializing population data at the fine scale. In this paper, the urban area of Xuanzhou District was selected as the research area. The population distribution vector data at residential building scale was established by proposing a spatialization method based on urban public facility elements. The method classified residential building patches. And it treated residential building patches as population distribution locations in geographical space with community boundary and community-level demographic data as the control unit. A multiple regression model of patch area and population was constructed. The spatialization method used in this study can reveal the detailed information about the population distribution in urban area. Results show that: ① The population distribution data, obtained by adopting urban public facility elements, is proved to be high accurate and reliable. The number of patches with estimated population in a reasonable range is 35.4% of 779 residential building patches. And the proportion of patches with relative errors of ±20% in population estimation is 61.2%. Moreover, the Chengdong community and Sijia community served as accuracy verification units, the absolute relative error of population estimation in these communities is less than 9%; ② Urban public facility elements, especially primary and secondary schools and kindergartens, vegetable markets and fruit shops, are important factors for accurate estimation of population within a residential building. Their estimation accuracy of number of people is high ifor multi-storied building, but lower for moderate high-rise building.
Suspect mobility prediction enables proactive experiences for location-aware crime investigations and offers essential intelligence to the crime initiative prevention. Recent practical studies and Rational Choice Theory suggest that the crime suspect mobility is predictable. The previous approaches for suspect location prediction focused on the forecasting the spatial likelihood of anchor point (i.e. the residential or future offending place) for a suspect who committed a series of crimes. However, the monitoring data are usually poor in availability for tracking suspects. Thus, the existing methodologies failed to capture the complex social location transition patterns for suspects and lacked the capacity to address the mobility data scarcity issue. Therefore, it is intractable to reflect suspects mobility patterns from sparse monitoring data, which reduces the effectiveness of case analysis and crime risk prediction. To address this challenge, we presented a novel Crime Records enhanced Location Prediction (CReLP) model. By merging the historical crime cases information by a collaborative filtering process, the CReLP model the estimate the visiting intensity at any arbitrary spatiotemporal node for and individual suspect. Particularly, we first obtained basic spatial and temporal units by partitioning the target areas into 100×100 2D grids and segmenting the daytime into 24 time bins. Second, we built a 3D tensor to model the social mobilities of all suspects with each entry in it representing the visiting intensity at each location and each time bin for each suspect. Meanwhile, this approach employed two matrices to express general movement trends among all suspects. Third, we created a suspect-correlation matrix relying on the spatial and temporal proximities of their historical crime events, as well as the similarities between their personal properties. At last, the missing entries in the 3D tensor were filled through the joint decomposition across all tensors and matrices mentioned above. This way were able to uncover the spatial distribution pattern for each suspect at any time. We evaluated the CReLP model using a real-world suspect mobility dataset collected from 241 suspects over 6 months with about 19 thousand location records. The results showed that our model outperformed three baseline approaches by 32% to 63% at RMSE (Root Mean Square Error) and 14% to 26% in TMSD (Top-k Minimum Search Distance), respectively. Finally, a suspect's visiting spatial distributions of the ground truth and predicting results between 8 to 12 a.m. were illustrated to show the performance of our proposed model.
This paper is aimed at exploring the determinants of population growth in Chinese urban areas. With the method of exploratory spatial data analysis and the data of traditional population census between 1990 and 2010, we could delve into the spatial distribution characteristics of the population growth rate and the multivariable spatial dependency during the past twenty years in Chinese city-level. Based on a thorough interpretation of population data, we are able to discover an existing spatial dependency between different cities. Obviously, spatial relations should not be negligible, because the spatial dependency is much stronger within cities living in shorter distance. It is more reasonable to use spatial regression model for our work, therefore, we use spatial lag regression model, spatial error model and classical linear regression model with spatial filtering to explore the influences of economic factors, climate factors, sociocultural factors and topography factors on population growth rate. It is showed that the classical linear regression model with spatial filtering can simulate the urban population growth rate batter than other models in our outcomes. The findings also suggest that economy is the most pivotal factors in population growth, such as the total amount of economy reflected by density of urban nightlight index plays an important role in driving population growth. Meanwhile other factors are following as well. Climatic variation is another systematic and significant factor affecting the rates of urban population growth. Some weather-related movement appears. People are willing to leave the unpleasant places and move to the places with nice weather. For example, with the increase of July heat index, there is a more and more stronger negative impact on population growth. The research shows that Chinese population growth is a complex question. There is a comprehensive action of multi-factor in generating the model of regional population growth. It is necessary to consider the different effects of economic development and climate conditions on the population growth in the researches on corresponding modeling and formulation of policy.
As a hydrologic unit with dense population, abundant resource and intensive capital, the stability of ecosystem structure of a basin lays the foundation of regional economy's development. Influenced by ecosystem's evolution together with human being's activities, the basin environment becomes more and more sensitive and vulnerable. In case of Xijiang River in Guangxi, the thesis is to make an analysis on the cause of its ecological vulnerability and construct evaluation indexes system with eleven indices under the framework of sensitivity-pressure-recovery. Supported by GIS technology, the ecological vulnerability index（EVI）was calculated and analyzed by Principal Component Analysis and difference method, which probes into the spatial and temporal characteristics of ecological vulnerability change from 2000 to 2010. Based on the factor and the interaction detect modules of geographical detector, it aims to analyze the explanatory power of impact factors on ecological vulnerability and the driving forces of factor interaction to the changes of ecological vulnerability in the basin. The results are as following: from 2000 to 2010, the ecological vulnerability index means kept at 0.69 for years, which was at moderate vulnerability. Spatially, the central basin is more sensitive than the surround area showing a declining trend from center to suburbs. The maximum of the average ecological vulnerability composite index of the whole basin for 2000-2010, is 3.40 at Guigang City. The minimum is 2.23 at Baise City. During the past ten years, the ecological fragility of the basin showed a slight deterioration. In year 2005, central and eastern parts of the basin were influenced by high temperature, which led to the higher ecological vulnerability comprehensive index in 2005 than other years. The order of the explanatory power intensity of the six factors on the ecological vulnerability is as follows: biological abundance index (0.475)>temperature at high temperature season (0.340)>vegetation coverage (0.211)>NPP (0.183)>rainfall erosivity (0.098)>Rainfall in flood season (0.030), and the explanatory power on results will be strengthened by factors cooperative interaction.
The accessibility level of scenic spots can not only reflect the convenient degree of tourism, but also can be used to measure the development potential of regional tourism. Moreover, the research on accessibility of tourist attractions in poverty-stricken mountainous areas could provide academic reference for the implement of poverty alleviation tourism strategy. Thus, this paper focused on the typical poverty-stricken mountainous areas to explore the influences of traffic cost and complex terrain features on the accessibility of tourist attractions, and discussed the influencing factors and driving mechanisms of scenic spots accessibility, by taking Dabie Mountain Area of Hubei province as an example. Firstly, this paper utilized connectivity index and accessibility index to study the tourism network structure of the study area. Then with the application of raster cost weighted distance method, this paper attempted to measure the accessibility of scenic spots and overall accessibility at county level from the perspectives of complicated terrain and transportations in different grades comprehensively. Results indicated that: (1) From the perspective of terrain factors, the accessibility of scenic spots located in such areas which have lower altitude and more gentle slope is relatively better, and the accessibility of humanistic tourist attractions is better than ecological tourist attractions; (2) Based on the analysis of transportation factors, the accessibility of scenic spots which located in the intersection areas of higher-grade highways is relatively better, hence the spatial distribution of accessibility has a strong traffic orientation; (3) By contrasting and analysis, the method of making comprehensive considerations for the complex terrain features and different grades highways in poverty-stricken mountain areas, can not only effectively identify the blocking influences of waters for the accessibility of tourist attractions, but also can highlight the improvement of scenic spots accessibility under the influence of natural factors through the melioration of traffic conditions. Consequently, this improved method may lead to more objective assessment of accessibility, even reflect the mutual influences between human and nature to a certain extent.
Qinling-Daba Mountains, as the geographical dividing line in central China, is the main water source area of the middle line project of transferring water from south to north, and also a sensitive area of the regional response to the global climate change. The spatio-temporal variation of vegetation, as a main component of terrestrial ecology system, represents the comprehensive response to climate change and human activities. Based on the MODIS data during 2001-2014, this study estimated the soil moisture using temperature vegetation dryness index, dynamically monitored the distribution characters and spatiotemporal variations of vegetation cover and soil moisture, analyzed its change trend, forecasted its future trend, and explored the interrelationship between vegetation cover and soil moisture in Qinling-Daba Mountains. The main conclusions are as follows: (1) Vegetation showed an increasing trend in Qinling-Daba Mountains during 2001-2014 years, as well as soil moisture, which respectively represented the best in 2010 and 2011, then both showed declining trends. (2) Vegetation distribution had a feature of “higher overall and lower in the middle”, while soil moisture with “lower in the north and higher in the south”, which showed a positive spatial correlation. (3) Vegetation had an obvious improving trend, and the significantly improved areas are dispersionally distributed, with no obvious concentration areas, and degraded areas were mainly concentrated in the north of Weihe River and a few areas along the eastern edge. Soil moisture increased significantly, and the increased areas distributed almost over the entire area except the northwest and northeast edge, and the reduced areas were small and not noticeable. The reverse variation of vegetation cover was stronger than continuous change, but soil moisture was just the opposite. In predicted future, in 44.36% of the study area, dispersed in entire area, the vegetation cover will undergo a process from increase to reduction, while soil moisture will continue to grow in 43.98% of the study area, including Qinling mountain in Shaanxi Province, Daba mountain in Sichuan province, and some parts of Jialing River basin and Han River. (4) The correlation between vegetation cover and soil moisture is positive. Nearly 69.71% of the study area showed an increasing tendency in both vegetation cover and soil moisture content, which was distributed in almost all parts of the study area except for the edges.
The rampant pseudo base stations have become a major public hazard. They undermine the normal telecommunications order, endanger public safety, seriously infringe the property rights of the masses, and violate citizen privacy. How to dig out the spatio-temporal patterns of the pseudo base stations’ activities from massive spam messages, design effective prevention and control programs, and fight against the crime from the source, has become the focus of government agencies and researchers. The traditional methods for identifying pseudo base stations through the user terminal, however, face great challenges in terms of accuracy, comprehensiveness, and analytical ability, which no longer meet the requirements of identifying small-scale and mobile pseudo base stations. Utilizing data on the spam messages from February 23rd, 2017 to April 26th, 2017 in Beijing, this paper analyzes the spatio-temporal distribution of pseudo base stations through non-negative matrix factorization. We also constructed a classification model through TF-IDF (Term Frequency-Inverse Document Frequency) which compares types from different classifiers (k-Nearest Neighbors / K-Support Vector Machine /Random Forest/ Single-Layer Neural Network) and selects the most accurate random forest classification method. Combined with the land use data, we analyzed the spatio-temporal distribution of pseudo base stations that send different types of spam messages. The results of non-negative matrix factorization and spam message classification were analyzed in detail. The results show that most of the spam messages in Beijing are sent along the road network and in the central city. The number of spam messages during the day is much more than that during the evening. As time goes by in the day, the distribution of spam messages along the road network gradually shrinks inward. The pseudo base stations that send different types of spam messages differ in the spatio-temporal distribution, but all of them favor the traffic facilities and residential area within the Fourth Ring. The non-negative matrix factorization, which provides reliable results that match with traditional spam message classification, has shown simplicity in performing the analysis and interpretability in the form and result of the decomposition. It can help understand the spatio-temporal patterns of different types of spam messages and provide evident-based suggestions for government agencies to fight against the pseudo base stations effectively. By targeting the source of the spam messages, it is also beneficial for governments to combat the illegal behaviors based on pseudo base stations.
Studying the distribution of urban functional areas and the laws of population activities on a fine scale has important implications for the government and relevant departments to rationally adjust the allocation of urban internal resources and arrange the layout of urban facilities. Taking the center area of Shenyang City in Liaoning Province as research area, based on the principle of nuclear density estimation, the distribution of functional areas in downtown Shenyang is explored based on point-of-interest (POI) data, and the temporal and spatial distribution patterns of urban population on working days and weekends are explored by interpreting the multi-temporal Baidu thermogram data. This paper analyzed the spatial structure of Shenyang city center from two perspectives: the distribution of the urban physical facilities and the law of the population activities. Finally, we used SPSS to analyze the correlation between population heat and urban physical facilities and established multiple linear regression models. The results showed that: (1) Shenyang City vitality area showed a multi-center distribution model, mostly in the commercial center, financial center or urban function complex center. (2) The spatial distribution of the hot spots in the working day was more scattered than that on the weekend. The area was larger and fluctuated greatly. However, population hot spots were mainly concentrated in the commercial center and urban complex functional area and there was a large fluctuation during the daytime on weekend. (3) The distribution of functional areas was significantly correlated with the population heat. Life service facilities, financial facilities, transports, accommodation facilities and attractions are the main factors influencing the population heat on the working day. While educational institutions, transports, public service facilities, catering facilities and financial facilities are the main influential factors on weekend.
Urban expansion is a complex time-conversion process, in which the different land use types are converted into urban land. The expansion of urban land use affects not only the development of the city itself, but also has a great impact on the ecological environment in urban areas. Taking into account the insufficient accuracy of urban built-up area extraction from a single data source alone, this paper analyzed the urban spatial expansion characteristics of Yucheng County in Henan Province based on the built-up area extracted by the decision tree classification and statistics variable extraction method from remote sensing imagery. In this experiment, the mean shift segmentation algorithm is used to segment the high spatial resolution image (GF1) firstly, then, based on the decision tree classification algorithm, image segmentation blocks were classified into three categories: cultivated land, water and construction land. Finally, the standard deviation information based on the land use type in 0.1 km × 0.1 km window is used to obtain built-up boundary. Taking Yucheng County as an example, the built-up area data is firstly extracted from high-resolution GF1 image (2017) and then, based on this data, we got the built-up area boundary of Yucheng county in 2000, 2009 and 2013 from temporal TM images, and finally analyzed urban expansion characteristic. The results show that the data of built-up area extracted from high spatial resolution imagery is more reliable by mentioned method. In this paper, the boundary of the built-up area has been further improved by the generalized supervised classification extraction method, with an accuracy of 89%. This paper showed a trade-off in the study of urban expansion to solve the problem of the low accuracy caused by only depended on low spatial resolution image and the low efficiency caused by only depended on high spatial resolution imagery to extract built-up area by using one high-resolution image and multiple low-resolution images.
The heavy rain may induce flood disaster inundating crop lands while the long period of the continues heavy rainfall may strongly affect the growth of crops even not evolving into a flood disaster. The impact of heavy rainfall on the growth of crops is a gradual process due to the long period of time needed for soil to become saturated. Multi-satellites remote sensing observations can capture and characterize the ground conditions over large area in multiply time. To develop the potential applications of the multi-satellites remote sensing observations, this paper proposes a method of extracting dynamic information of heavy rain disaster and its impacts on the growth of crops using multi-satellites data including Terra/MODIS, Landsat and Sentinel. We implemented the application of the proposed method in the studying area around Chaohu lake, where the heavy rainfall started from the end of June and continued to August in 2016, and a heavier rain in July brought about the flood in large crop areas. The beginning period and the duration of the heavy rainfall, leading to the impacts on the growth of crops, were identified using a dynamic threshold method of multi-temporal NDVI derived from MODIS. Based on this information, the area with crop fields impacted by heavy rainfall and flooded were obtained. On the other hand, the dynamic information of flooded lands was extracted by using Landsat observing data in July and Sentinel observation data in August, respectively. These results provide more accurate area of flooded crop fields, and can be used to modify the area derived from MODIS although they have only few temporal data available. In conclusion, multi-satellites remote sensing, as one of the tools for monitoring and assessing the influences of heavy rainfall, can obtain the dynamic information of the heavy rainfall impacts on the growth of crops in addition to flooded land area and the recovery of the farmland, which provide the supporting scientific data for the assessment of the loss caused by the disaster and making the disaster relief policy.
Influenced by scale effect, different objects have different spatial and attribute scales, so there is no one set of segmentation parameters can suit all objects in the same image. However, classifying similar objects into same region, and then setting optimal segmentation parameters for different regions can improve the overall segmentation accuracy of the images effectively. In cultivated land extraction, it is critical to have clear and continuous boundary for the segmented plots. This paper presents a cultivated land extraction method combining image region division and segmentation parameters estimation. Temperature inversion was used to divide the image into different regions, of which the types of covered objects are different or the growths of crops are different. Next, regional image segmentation is performed for different regions. Since different object's inherent spatial scales are different, so different regional images' optimal segmentation parameters are also different. The optimal segmentation parameters of the image can be estimated to a certain level by analyzing the characteristics of the image space quantitatively. Compared with other segmentation parameter optimization selection methods, this method can be accurately and quickly positioned, and has a higher efficiency. Next, estimated segmentation parameters were used in the process of cultivated land segmentation by using edge restraint watershed segmentation algorithm. Edge restraint watershed segmentation algorithm uses the canny operator for the post-processing of the watershed segmentation algorithm. The boundary of the canny operator is used to constrain the consolidation process. The experimental result shows that the method proposed in this paper can set segmentation parameters for different regions quickly and accurately. Compared with other segmentation methods, this propose method has a quite clearer and more continuous block boundaries and the boundary fragmentation problem is greatly relieved.
Anthropogenic heat discharge not only constitutes the cause of urban heat island (UHI) formation, but also is an important indicator related to energy consumption. It is important to analysis the magnitude and variation of anthropogenic heat discharge in order to mitigate UHI effect and improve energy efficiency. This paper examined the spatio-temporal variation of anthropogenic heat discharge in the Xiamen Island, China using Landsat TM data and meteorological data. First, the anthropogenic heat discharge was estimated with a remote sensing-based surface energy balance model. Then, the urban functional regions derived from IKONOS data were combined with the anthropogenic heat discharge. The results indicate that the anthropogenic heat discharge in different types of urban functional regions reaches the maximum in summer and the minimum in spring. The anthropogenic heat discharge of industrial area was higher than those in the other regions for all seasons. The high anthropogenic heat discharge occurred in the old industrial bases in the west of Xiamen Island. In traffic area, high anthropogenic heat discharge was observed in the Changan Road, Jiahe Road, Chenggong Avenue, Xianyue Road, North Hubin Road-Lvling Road, South Hubin Road-East Lianqian Road. In residential area, high anthropogenic heat discharge was observed in the old town. The high anthropogenic heat discharge occurred in the large single buildings in commercial and public area. Overall, the anthropogenic heat discharge in the western part of Xiamen Island was higher than that in the east. The differences of spatial and seasonal distribution were closely related to land cover types, population and the degree of economic development. Moreover, the density and height of the buildings and materials of land cover change the amount of anthropogenic heat discharge by affecting other surface fluxes. This paper brings a more microscopic perspective by analyzing the spatio-temporal variation of anthropogenic heat discharge in different urban functional regions to study urban thermal environment and energy utilization, as well as to provide a theoretical basis for promoting urban sustainable development.
