Abstract With the emergence of global environmental change issues， Land Use/Cover Change （LUCC） issues have received increasing attention. Therefore， the dynamic monitoring of LUCC has also become very important. In this paper， preliminary exploration was made to the research progress on the dynamic monitoring technologies for LUCC as well as their advantages and disadvantages， and prediction was made to the development trend of future monitoring technology.
　　Key words Land use/cover change; Dynamic monitoring; Environment; Remote sensing
　　 Nowadays， with the rapid increase in population and the advancement of science and technology， human beings are changing the global environment on which they depend for survival at an unprecedented speed and scale. This has led to a series of global environmental change issues， such as increased greenhouse gases in the atmosphere， ozone layer destruction， land desertification， biodiversity loss， and forests reduction， which pose a great threat to the survival and development of human society. Therefore， it has become an important subjects for the countries all over the world to understand the causes of global changes， predict their future trends and possible consequences， and formulate corresponding countermeasures. In the global environmental changes caused by human activities， Land Use/Cover Change （LUCC） and industrial production processes play the decisive role[1]. According to statistics， the amount of carbon dioxide emitted to the atmosphere caused by land cover changes is comparable to that of fossil fuels used in industrial development over the past 150 years. Since land cover is the source and sink of many material flows and energy flows that support the Earths biosphere and geosphere， all natural forces and human activities will cause LUCC. Therefore， land cover changes mainly caused by human land use activities are bound to have a major impact on the  climate， hydrology， biogeochemical cycles and biodiversity of the earth system[2]. In this context， the "Land Use/Land Cover Change" LUCC core research project was initiated in 1995 under the joint sponsor of the International GeospatialBiosphere Program （IGBP） organized by the International Science Alliance and the International Human Dimension Program on Global Environmental Change （IHDP） organized by the International Social Science Alliance， which has become a hot topic in academic research in various countries. Therefore， how to dynamically monitor LUCC has become a focus of research and exploration. This paper started with the connotation and relationship of land use and land cover， and gave an overview and brief evaluation of the LUCC dynamic monitoring methods used in the past and present. 　　Status of LUCC Research and Monitoring Methods at Home and Abroad
　　Connotation and relationship of land use and land cover
　　Land use and land cover are 2 important concepts that are closely related and essentially different[3]. Land cover is a new concept that emerges with the application of remote sensing technology. It refers to natural objects and artificial buildings covering the ground， and it reflects the natural condition of the Earths surface. Land use refers to the way humans use and utilize the natural attributes of land， including the purpose and intention of human use of land， and it is a human activity. Therefore， things like agriculture， logging， grazing and urban development are land uses， while forests， grasslands， roads and buildings， as well as soils， glaciers， and water surfaces belongs to different categories of land cover. However， in many cases land use and land cover refer to the same objects， and thus these two concepts are easily confused. For example， for the same plot of grassland or cropland， when it is only considered as vegetation without considering its purpose and use， it is land cover; if considering its use for grazing or grain production， it is the corresponding land use type. Landuse change usually leads to changes in land cover conditions， which mainly fall into 2 types[3]： modification and conversion. Modification refers to changes in the internal conditions of the same type of land cover， such as thinning forests or fertilizing farmland. On the other hand， conversion refers to the conversion of one type of cover to another， such as turning forests into farmland or grassland. In addition， maintenance， which means to keep land cover in a certain state， is also a form of human activity affecting land cover.
　　Status of LUCC research and monitoring methods at home and abroad
　　Britain is the country which carried out the earliest research on land use in largescale at abroad. It is a country with more people and less land， where the agricultural products cannot be selfsufficient. In 1930， the Great Britain Land Utilization Survey was established in order to find out the actual value of various types of land use. It was presided over by Professor Stump， Director of the Department of Geography of the School of Economics， University of London， and conducted a national land use survey in 1931-1939. After the 1960s， due to the modernization of agriculture， the area planted with crops was gradually reduced， while the land use situation changed in many ways with the development of the city， the expansion of the suburbs， and the rise of tourism. In order to protect natural resources and the environment， and to prevent natural disasters， the United States carried out regional construction planning， and together with the reasons to reasonably levy land taxes， the government needed to understand the actual situation of land use. New technologies such as remote sensing and cartographic automation have begun to be applied， and land use surveys and mapping can be carried out at the same time quickly and efficiently with less costs. Japan started a land use survey in 1953， and it is a country with relatively good land use mapping and plotting in a large proportion all through the country. 　　In China， the research and investigation on land use started very early. In the 1960s， the Institute of Geography of the Chinese Academy of Sciences conducted a study on the zoning of land use in China. In the 1980s， the Ministry of Agriculture and the Bureau of Land Management first carried out national land surveys and detailed investigations. At the same time， the Institute of Geography of the Chinese Academy of Sciences hosted the mapping of 1∶1 000 000 national land use map and the study on land use， which provided the basis for current and future research on land use， human environment and global change. Since the 1990s， the research on land use has become more colorful in China.
　　In the monitoring of LUCC， before the 1970s， the monitoring method was based on traditional monitoring methods， that is， usingawkward instruments such as theodolite， plotter， and range finder to conduct field measurements. In the 1970s， especially since the 1980s， remote sensing was widely used in land use and land cover change monitoring and investigation. Since the 1990s， "3S" technology has played an increasingly important role in the dynamic monitoring of LUCC.
　　Development of LUCC Dynamic Monitoring Technology
　　The ultimate goal of land use and land cover research is to study the development process and future trends of the land， so dynamic monitoring is particularly important. The socalled land use/cover dynamic monitoring compares the land use/cover data of different phases （at least two phases）， and analyzes its dynamic characteristics and future development trends in terms of space and quantity. In general， the accuracy of dynamic monitoring depends on the accuracy of the classification. In the study of land use and land cover change， it is necessary not only to obtain information on land use change， but also to obtain the type of change， that is， to obtain land classification information for different monitoring periods.
　　Traditional monitoring classification methods
　　The traditional monitoring method builds the land cover change archive through the establishment of ground observation network or observation sample belt， which qualitatively and quantitatively observes the evolution process and evolution trend of land cover. For example， field survey visits are used to find the driving force of land use change; field measurement is conducted to obtain the density， height， coverage， type， growth of ground land cover; investigations are made to survey biodiversity and animal habitats. The dynamic monitoring of LUCC has longterm， comprehensive and periodic characteristics， and the survey results must have high precision and accuracy. However， traditional monitoring means and methods have many limitations. The traditional ways are mainly manual survey methods， most of which are based on field measurements， consuming a lot of manpower and material resource. They are low in efficiency and long in time consuming， therefore often failing to reach the aims to obtain LUCC data directly， fast and comprehensively. Moreover， it is hard to them to locate in real time， fast and accurately in dynamic data sampling， leading to great error between the monitoring result and the actual. With the rapid development of modern industry and agriculture， land use changes are becoming more frequent. Obviously， it is difficult to meet the needs of rapid and accurate monitoring of land resource changes by using conventional monitoring methods[4]. 　　RSremote sensing monitoring and classification technology[5-12]
　　Advantages of RS monitoring
　　Remote Sensing （RS）， the "distantperception"， is a comprehensive technology for earth observation developed in the 1960s. As an important means to obtain environmental data and dynamic monitoring （especially surface information）， remote sensing technology has many advantages. First， it can observe the earth from high altitude through earth observation satellites or aircraft， and can carry out largearea synchronous monitoring to quickly and accurately obtain the  environmental information data， which is comprehensive and comparable. Second， the use of remote sensing technology to obtain environmental information has the characteristics of large data acquisition， multiple information means， large amount of information， fast speed， short cycle and few limited condition. Third， compared with the traditional LUCC information acquisition methods， remote sensing can greatly save manpower， material resources， financial resources and time， and has high economic and social benefits. At present， remote sensing technology is undergoing a developmental change from qualitative to quantitative， from static to dynamic. In the research on LUCC， remote sensing information is not only its main source of information， but also provides technical and results accuracy guarantee for its dynamic monitoring. Remote sensing has the capability to achieve realtime imaging， realtime transmission， fast processing and periodic observation to acquire remote sensing information of the same object， and can provide various timely， accurate and comprehensive resources and environmental data from dynamic monitoring in a wide range. It is an effective means to carry out land resource survey and land use dynamic monitoring and feedback. Remote sensing images can be used to quickly， repeatedly and dynamically acquire various information in large areas， thus providing technical support and accuracy guarantee for largescale， dynamic and periodic land use/cover dynamic monitoring.
　　RS dynamic monitoring method
　　Pixelbypixel comparison method and postclassification comparison method
　　The most commonly used land use remote sensing monitoring methods can be divided into 2 types： pixelbypixel comparison method and postclassification comparison method. The pixelbypixel comparison method first compares the spectral features of the same phase image in different years of the same region to determine the location of land use change. On this basis， the classification method is used to determine the land use change information. This method can generally detect the changed pixels in a more sensitive way， but it can not obtain the specific type information of land use change at the same time. The postclassification comparison method first classifies the images of different phases of the whole supervised area. Then， the classified results of the images of the same position are compared to determine the location and type of land use change. This method can obtain a detailed land use transformation matrix， but this method is obviously affected by the error caused by separate classification， and will inevitably amplify the degree of change. 　　Principal component analysis
　　In view of the unsatisfactory situation of the above 2 methods， some researchers have proposed the principal component analysis method after multiple time phases （multiphase） remote sensing image superposition. This method is to classify the superimposed images instead of separately classifying the images of the respective phases， thereby greatly reducing the amplification of the change degree. The current technical method is to select satellite images of 2 phases as the main data source， followed by geometric correction， geometric registration and data fusion， and classify the images through computer automatic extraction， visual interpretation and humancomputer interaction interpretation.
　　（1） Direct extraction of change information
　　The direct extraction of change information is a pointtopoint direct operation of remote sensing images of 2 phases， and the information of land use change is obtained through the discovery and classification processing of the changed features. There are mainly 4 methods. First， image difference method. It is the method in which the spectral gray value of a certain band of one phase subtracts the spectral gray value of the corresponding pixel of the other phase， and the earliest application is the band image difference method. It is difficult to extract dynamic information in the singleband difference image， and color synthesis of the MSS7， MSS5， and MSS4 difference images can synthesize the dynamic information of each band and highlight the vegetation change information. Second， image ratio method. This is to perform the division operation to the spectral gray value of the samenamed pixel in the 2phasemultiphase data. The ratio method can partially eliminate the influence of shadows， highlight the contrast between certain features， and have a certain image enhancement effect. On the one hand， the ratio image can be directly interpreted， so as to extract the thematic information. On the other hand， a little logical transformation can be used to directly detect the environmental elements that are obviously changed. Third， vegetation index method. It is a comprehensive utilization of the strong absorption of vegetation in the red light part and the strong reflection characteristics in the nearinfrared part to extract vegetation dynamic information. ratio vegetation index， normalized vegetation index， and vertical vegetation index， which are in particular widely used in the dynamic monitoring of forest resources. Fourth， multiphase composite classification. The 2phase or multiphase remote sensing data is combined to extract the change information through remote sensing classification. It is difficult to determine the training area in the supervision process of this method. 　　Agricultural Biotechnology2019
　　（2） Visual interpretation method
　　Visual interpretation is based on the image features and spatial features （shape， size， shadow， texture， pattern， location and layout） of the sample， combined with a variety of nonremote sensing information （land use status survey data）， to carry out sortingout comprehensive analysis and logical reasoning proceeded from one point to another， from the outside to the inside using the laws correlated with biogeography with comparative analysis， thereby determining the interpretation marks of various localities， so as to obtain the remote sensing classification images of the various phases after drawing the boundaries of the localities on the remote sensing image， which are then compared. There are two main forms of application of the visual interpretation method： one is to combine the aerial image， the satellite image or multiple remote sensing image data to carry out artificial interpretation， followed by manual editing the land use map or digital processing to obtain the land use classification information. This form has been widely used since the beginning of using remote sensing in land use surveys， such as the early "3North" Shelter Forest Remote Sensing Comprehensive Survey （1986-1991） and the previous work of the "National Land Use Status Survey" hosted by the former State Land Administration. This method can make full use of the knowledge of the interpreters， has good flexibility， and it is good at extracting spatial related information， but it takes much more time and has personal differences. The second form is the humancomputer interactive visual interpretation method mainly formed with the development of computer technology and remote sensing image processing technology. It can use the remote sensing image processing software to arbitrarily enlarge and reduce the image. After performing various enhancement processing on the remote sensing image to achieve the best visual interpretation effect， the interpreter can directly and accurately plot the ground object boundaries using the mouse along the edges of the image features based on the screen interpretation marks of the various ground objects on the images. Chen et al.[13]used Corel DRAW6 software to make an exploratory study on the conditions， processes and main features of humancomputer interactive land resource remote sensing interpretation. The results prove that the humancomputer interactive method can give full play to the advantages of people and computers， and can realize interpretation and mapping in one time. Moreover， compared with traditional remote sensing land use mapping， it greatly reduces manpower waste. 　　（3） Computer automatic classification
　　The computer automatic classification method is to separately classify the remote sensing images of each time phase before comparing the multiphase remote sensing images. There are 2 types of unsupervised classification and supervised classification.
　　A. Unsupervised classification
　　The socalled "unsupervised" is to classify the remote sensing images in a natural way based on the distribution law of the spectral features of the remote sensing images. In this way， classification is only made to different categories， but the attributes of the categories are not determined during the online process. The generics is determined after the analysis on the  various spectral response curves after the classification， as well as the comparison with the field investigation time phase. In the unsupervised classification， the main algorithms are mixed distance method （ISOMIX）， circular cluster method （ISODATA） and synthetic sequential product group method. Although unsupervised classification is less affected by human factors and there is no need to have a lot of practical understanding of the surface， it is widely believed that the results of the unsupervised classification is not as satisfactory as those of the supervised classification due to the existence of phenomena such as "homogeneous heterogeneity"， "homologous spectrum" and mixed pixels， so the unsupervised classification does not apply to the precise classification of cultivated land in the mountains， but only applies to the general classification of images in which the categories have been already known. For example， Jose[11]first used the unsupervised classification to get the general category and then performed the detailed classification in the case of land cover/land use mapping. HegaratMascle et al.[11]emphasized the advantages of unsupervised classification when using multiphase ERS images and radar data to identify land cover types. Thomas et al.[11]proposed an unsupervised classification method based on multiphase ratio data， and emphasized that the method is not only simple but also highly accurate.
　　B. Supervised classification
　　Supervised classification is also known as training area classification. Its most basic feature is that before the classification， people can have some priori knowledge to the category attribute of the image ground objects in some sampling areas on the remote sensing image with the combination of field sampling investigation and visual interpretation， and then the computer can "train" the decision function according to the features of the priori knowledge to complete the classification of the whole image. The classical supervised classification methods include maximum likelihood method， parallelepiped method， Mahalanobis distance method and minimum distance method. Compared with unsupervised classification， supervised classification has certain advantages， but the classification results produced by it often have more misclassifications and missing points， resulting in lower classification accuracy. Therefore， in order to improve the accuracy of supervised classification when extracting land use information， some measures are taken before or during the classification of images. The measures taken before the image classification are mainly for the training area， because the accuracy of the supervised classification is closely related to the selection of the training area. Both Wu and Yang proposed the theory and method of training sample purification， and the experimental research showed that after the training samples are purified， improvements were made in the divergence between the types， the probability density function of the sample pixels， the fitting degree of the Gaussian distribution and the accuracy of the classification results. Some scholars also proposed the automatic or semiautomatic extraction method of the training area for the limitations of the traditional manual training area extraction method， and some studied the influence of the distance between the samples of the training set on the classification accuracy. Based on the phenomenon of "homogeneous heterogeneity" and "homogeneousheterogeneity" in the classification process， many experts have proposed improvement methods. In the supervision and classification of the data of Qingdao City， Ping used separate sampling methods for the phenomenon of "homogeneous spectrum"， and thus obtained satisfactory results. Pan et al.proposed to subdivide the types to solve the problem of "homogeneous heterogeneity"when dealing with remote sensing images of the Three Gorges region， and the misclassification brought by "homogeneous heterogeneity" could be corrected by the introduction of geocontrol system， so as to improve the plotting accuracy.  Mo et al.[12]proposed and used subarea classification method when using TM data to monitor the dynamic change of land use， avoid a large number of mixed and misclassified phenomena. In addition， people are also working on the application of various auxiliary data in computer classification of remote sensing images. The sources of auxiliary data are more extensive. In addition to topographic maps， aerial sketches， soil maps， vegetation maps， geological maps and other map materials， there are also relevant groundbased measured data and statistical data. For example， Yang et al.[12]highlighted the 3 methods of matching solar incident angle data， radiation correction classification method and elevation classification method. 　　Combined classification of RS， GIS and GPS
　　The conventional remote sensing statistical classification method is mainly based on the reflection characteristics of the ground object spectrum， and is operated based on a single pixel. However， since the remote sensing data generally has the characteristics of comprehensive spectral information （that is， one pixel is sometimes the sum of the spectrum of various ground objects）. As a result， computer classification faces many ambiguous objects， resulting in reduced accuracy. To this end， people are constantly researching and trying new classification methods. Both remote sensing and Geographic Information System （GIS） study the spatial entities in nature. As an effective tool for spatial data processing and analysis， GIS can provide a good environment for remote sensing applications， enabling remote sensing images to hive higher classification accuracy with the support of GIS. GIS can store a large amount of attribute information and topographical topological information in the form of numeralization （quantification）. This information can be fully applied to the classification decision of remote sensing images. Slope， location， topography， and elevation have a great influence on the distribution of vegetation， and in the meantime， human factors are increasingly affecting land use and land cover change. Thus， the effective use of these quantitative information can improve the accuracy of classification. Paul et al.[14]studied matrix superposition analysis of SPOT classification results with GIS support， so that the classification image and land use partition information were combined， which made the accuracy increase to 78%. Paul[14]used spatial thematic information such as soil texture and topography to improve the accuracy of land use classification for TM data. When using the TM data for classification and auxiliary mapping research， Liu[13]pointed out that GISassisted classification can not only improve classification accuracy but also improve reliability. Li also proposed a new method of using GIS technology to extract shape information and improve classification accuracy， so that some confusing classifications were corrected.
　　In addition， as for traditional techniques， after the change region is discovered by remote sensing technology， the quantitative determination of the change region is only measured by the regional boundary on the remote sensing interpretation map. The imaging mechanism of the remote sensing image， the inherent error of the image classification method and other errors （such as drawing errors） make the regional boundary on the remote sensing interpretation map only a boundary with schematic and considerable ambiguity. Especially for largescale maps， the accuracy of the change region boundary derived from the satellite image is too low. Therefore， after the land use change region is basically determined by means of remote sensing， accurate measurement of the change region is very necessary. The Global Positioning System （GPS） can achieve the above goals by providing global users with allday， continuous， realtime， highprecision 3D position， 3D speed and time data. In practical applications， GPS is mainly used to provide targets in real time and quickly， including the spatial location of various sensors and carrier platforms. In this way， GPS has become a revolutionary change in the measurement disciplines due to the high degree of positioning flexibility and the high precision that conventional measurement technology cannot match. Therefore， GPS provides technical support for accurately grasping the land use change region. Furthermore， GPS data is also a necessary and useful supplement to remote sensing information. The response of remote sensing means to cumulative changes can become more obvious after a period of time. For a city or countylevel administrative region， an obvious， largearea change region can be roughly determined by the satellite， but for smallarea or sudden changes that have a greater impact， which may not be reflected on the satellite image， there may be no need to determine using remote sensing means. At this time， the GPS receiver can be used to conveniently obtain the data of the change region in the field and update the dynamic monitoring information system database. Therefore， GPS can not only quantify the field positioning of remote sensing results， but it is also an independent data source. 　　Conclusions and Prospects
　　It has been proved that remote sensing technology has become more than a traditional monitoring method and it plays an important role in the LUCC dynamic monitoring process. There are various remote sensing monitoring methods. In order to determine the type of land change， remote sensing image classification methods are introduced for the monitoring. The improvement of remote sensing classification methods has always been an important area of the research on remote sensing technology. Although it has been pointed out that visual interpretation is still a successful classification method to date， with the development of computer technology， the new classification method will be gradually applied to largescale land cover and land use research. The combination of more scientific， quantitative and practical "3S" technology represented by RS， GIS and GPS with conventional groundmeasured data has become the mainstream way of land cover and land use dynamic monitoring. With the continuous deepening of model research， it can be predicted that the 3Sbased land use/cover dynamic monitoring technology plus with model quantification prediction method will become a rapid and effective land use dynamic monitoring means and a platform for spatiotemporal data management and analysis for the relevant administrative departments of provinces， cities and counties， thereby providing the decisionmakin basis for the rational use of regional land resources and sustainable socioeconomic development.
　　References
　　[1] SUN J， ZHAO W， ZHAO LQ. Introduction to senseremoting dynamic monitoring technology in landusing[J]. Land and Resources in Shandong Province， 2005， 21（4）： 271， 237.
　　[2] BAI WQ， BAI SQ. The Status and Roles of Land Use and Cover Change in Global Change Study[J]. Areal Research and Development， 1999， 18（4）： 13-16.
　　[3] CHENG F. Review of research on land use and land cover change[J]. Land and Resources Information， 2003， 1： 53-56.
　　[4] WANG SM， ZHAI HQ. Application of remote sensing technology to land use/cover change in China[J]. Geospatial information， 2004， 2（2）： 31-32， 38.
　　[5] LI YZ. Land use/cover change dynamic monitoring by "3S" technique[J]. Journal of Agricultural Mechanization Research， 2007， （1）： 197-199.
　　[6] ZHU XJ， WANG DL， GUO JY. Study on application of GPS technique in updating investigation of land use[J]. Journal of Taishan University， 2005， 27（6）： 37-40. 　　[7] WANG XD， CUI WH. 3S Roles in detection of land use[J]. Territory & Natural Resources Study， 1998， （3）： 27-31.
　　[8] GE QS. Land use change in China during the 20th Century[J]. Acta Geographica Sinica， 2000， 55（6）： 698-706.
　　[9] WANG DB， CHEN BH， CUI SW. A research on use of kinematic RS monitoring in land use[J]. Bulletin of Surveying and Mapping， 2004， （2）： 46-49.
　　[10] GAN FP， WANG RS， WANG YS， et al. The classification method based on remote sensing techniques for land use and cover[J]. Remote Sensing for Land & Resources， 1999， （4）： 40-45.
　　[11] HAN T. Some research advances and methods on detecting land cover change by remote sensing[J]. Arid Meteorology， 2004， 22（2）： 76-81.
　　[12] ZHANG YH， ZHAO GX. Classification methods of land use/cover based on remote sensing technologies[J]. Journal of China Agricultural Resources and Regional Planning， 2002， 23（3）： 21-25.
　　[13] CHEN NQ， DAI JF. Method research on manmachine interactive interpreting and analyzing of the land resource by using remote sensing[J]. Remote Sensing Technology and Application， 1998， 13（2）： 15-20.
　　[14] ZHANG SL， YANG BJ， WANG F， et al. GISbased remote sensing monitoring of humanmachine interactive image interpretation system[J]. Transactions of the Chinese Society of Agricultural Engineering， 1999， （2）： 185-188.
　　Editor： Na LI Proofreader： Xinxiu ZHU