Technical Discussion on Updating Land Use Database with SPOT-5 Remote Sensing Image —— Taking Wangkui County of Heilongjiang Province as an Example

(Heilongjiang Provincial Institute of Land and Resources Survey and Planning, Harbin, 150056)

Updating the land use database is an important content of land and resources informatization construction, which provides accurate, detailed and current basic data for land use change investigation, land planning, cultivated land protection and land management of construction projects for land management departments at all levels. Connecting with the actual situation of updating the land use database in Wangkui County, Heilongjiang Province, this paper emphatically introduces some new technologies and methods for updating the land use database by using SPOT-5 remote sensing images, analyzes the problems existing in updating the land use database, and advances some methods for solving these problems.

Keywords: remote sensing; Land use renewal; Digital orthophotomap; “3S”； Wangkui county; Heilongjiang province

The application of remote sensing technology began in 1950s, and has been developed for more than 50 years. With the progress of remote sensing satellite technology, the resolution of remote sensing images has been greatly improved, and it has been widely used in land and resources investigation with many advantages such as fast speed, short time and low cost. In particular, it provides important technical support for the rapid updating of land use database by using remote sensing technology.

County land management is the foundation of the whole land management. Due to the rapid development of economy and society in recent years, the updating of basic maps of land use lags behind the changes of land use status, which can not accurately and timely reflect the current situation of land use. It has greatly hindered the needs of economic and social development. Under this background, it is urgent to update the land use database by using remote sensing technology. Using high-resolution remote sensing data sources, combined with GPS and GIS technology, the land use database can be updated, thus providing accurate and informative information on land use status for land management departments at all levels. Therefore, this paper takes Wangkui County of Heilongjiang Province as an example, taking the high-resolution satellite remote sensing data of SPOT-5 2.5 m as the information source, and superposes the 1∶ 1 10,000 digital orthophoto map with the vector data of land detailed survey. Through the combination of vector and grid, supplemented by expert knowledge and human-computer interaction, the changed information attributes after field investigation are given to vector data, and then the topological relationship is re-established.

1 Basic information of the project area

Wangkui County is a county with agriculture as its main industry in Heilongjiang Province, which is located in the middle of Heilongjiang Province and the transitional zone between Songnen Plain and the southwest edge of Xiaoxing 'an Mountains. It is located at east longitude12610 ′ 23 ″ and north latitude 46 32 ′ 07 ″. Yinke River, Nuo Min River and Hulan River in the southeast are adjacent to Suihua City and Lanxi County, Tongken River in the west is opposite to Qinggang County, and Helen City in the north. The county covers an area of 2320 square kilometers, 62 kilometers long from north to south and 60 kilometers wide from east to west.

Wangkui County is shaped like a parallelogram. Affected by the vestiges of Xiaoxing 'an Mountains, the terrain is high in the east and low in the west, with the east-west slopes dropping by about one thousandth and the north-south slopes dropping by about one thousandth. The highest altitude is 250.9m, the lowest altitude is 127.8m, the absolute height difference is 123. 1m, and the average altitude is 167m. Geomorphological types can be roughly divided into three units. The hills in the east and south are undulating, the transition zone in the middle is slightly undulating, and the west is low and flat.

2 overall technical route

Using remote sensing images to update the land use database has a high degree of automation, which can reflect the land use situation in time, truly and objectively. Therefore, using the SPOT-5 2.5 m remote sensing image to update the land use database in Wangkui County will have some new features in the overall technical route compared with the daily changing method.

2. 1 Establishment of sampling table of image land classification

Generally speaking, the different regional features of the project, such as mountains, hills, plains, deserts, etc. , reflecting different land types in remote sensing images. In addition, the choice of resolution and the time equivalence of images will also affect the representation of land types reflected in remote sensing images. Therefore, it is very important to establish a sample table of image land classification suitable for this project area, interpret it indoors and extract the change information. The land classification sample table is also the interpretation sign of the image. It is an image feature that can directly reflect and distinguish the information of ground objects in the image, and it is the main basis for identifying the attributes of ground objects, including shape, size, pattern, tone, texture, shadow, position, layout, resolution and so on. In the database updating project of Wangkui County, the French SOPT 5 remote sensing image was selected. Through field investigation, the corresponding relationship between SOPT-5 remote sensing image and ground objects is determined. With the help of relevant auxiliary information (planning map, topographic map and other related materials), various characteristics reflected on the image map are described, and a set of representative image land classification sample tables is established, which greatly facilitates indoor interpretation and extraction of change information. Figure 1 is a part of the sampling table.

Figure 1 image land classification sampling table

2.2 "3S" integrated technology in the comprehensive application of land use database update

"3 S" integration technology refers to the technology that rs (remote sensing), GPS (global positioning system) and GIS (geographic information system) are gradually integrated in the process of parallel development, and the three are organically combined. It constitutes a technical system that integrates information acquisition, information processing and information application. It is a space technology application system that makes full use of their respective technical characteristics and gradually becomes a new discipline with strong practicality and application.

On the one hand, the SOPT No.5 remote sensing image (geometric correction registration) is made by GPS measurement, and obvious feature points are selected on the SOPT No.5 remote sensing image, and the image image is geometrically corrected by high-precision GPS measurement in the field, so as to make1:10000 remote sensing digital orthophoto map; On the other hand, the land use renewal supported by GIS includes data vectorization editing, superposition and topological relationship establishment. In the process of land use renewal in Wangkui County, the domestic GIS software, namely MAPGIS6.5 software developed by Wuhan Zhongdi Company, was used to update the land use database.

2.3 Vector data and raster data are combined, and various information extraction methods are used together.

Vector data mainly includes land use status database; The raster data mainly includes 1 ∶ 10000 remote sensing digital orthophoto map and land use status map after scanning correction. Through the comparative analysis of vector data and raster data, update vectorization and extract change information. Information extraction methods include manual visual interpretation, computer automatic extraction and the combination of manual visual interpretation and computer automatic extraction. In this update, the land use database is mainly updated by manual visual interpretation and human-computer interaction.

2.4 Remote sensing data combined with basic map of land use

Make full use of land use survey, urban cadastral survey, administrative division definition, batch land use results of construction land, land consolidation and reclamation results and daily change data to assist in determining land use types and changes, reduce the workload of field investigation and shorten the update time of land use database. But also ensures the accuracy of land use types.

2.5 Indoor Interpretation and Field Verification Route

Remote sensing observation only reflects the spectral characteristics of ground objects, that is, the representation of land cover, while land use map reflects the situation of land use, and the change of land cover does not completely represent the change of land use. In addition, the change attribute information extracted by manual visual interpretation needs field investigation, such as the width and ownership of sporadic and linear features. Therefore, it is necessary to determine the land use type through the combination of indoor and outdoor work to ensure the thoroughness of this update and realize the comprehensive update of the land use database.

Three working methods

The working method is mainly divided into four steps: ① spot-52.5 m remote sensing image processing flow, including making 1 ∶ 10000 remote sensing digital orthophoto map; (2) Indoor processing, including information extraction and map making before field work; (3) Field investigation process, including verifying the change information, investigating the width and ownership boundary of sporadic and linear features, etc. ④ The updating and warehousing process of land use database. See Figure 2 for the work flow chart.

3. 1 making1:10000 remote sensing digital orthophoto map.

Remote sensing digital orthophoto map (RSDOM) is an image data set generated by geometric correction and corresponding image processing of panchromatic and multispectral images with certain resolution obtained by satellite remote sensing, which has both geometric characteristics of maps and image characteristics. The data sources of this remote sensing image are French SPOT-5 2.5 m panchromatic image and 10 m multispectral satellite data, and the data resolution fully meets the requirements of database update. The width of the scene is 60km2×60km2, and the cloud content is zero. Taking the whole scene as a unit, the 2.5m panchromatic image is fused with 10m multispectral data by using satellite image processing software ERDAS. Because the terrain of Wangkui County is relatively flat, the fusion method of weighted multiplication has the advantages of retaining the texture information of high-resolution images and the color information of multi-spectral images, and making satellite remote sensing digital orthophoto maps. Then use the image processing software photomapper to make 1 ∶ 10000 digital orthophoto map. The technical process of image processing is shown in Figure 3.

3.2 Indoor treatment

Using domestic GIS software MAPGIS6.5, the engineering data is established, and the database is updated by combining vector data with raster data, comparing remote sensing images with land use vector data. Information that can't be judged in the office is specially marked when making the field map, and printed at the same time, and the land use change is verified through field investigation.

Fig. 2 Workflow diagram of database update

Vectorization update collection includes: ① The update collection of graphic data is carried out in layers, mainly including contour lines, water systems, roads, administrative boundaries, ownership boundaries, land boundary, other linear elements with width, sporadic elements, text notes, etc. According to the needs of database building software, the main levels are subdivided to realize the obvious distinction of different graphic elements, which are used for database building respectively. (2) The updating and collection of attribute data should be strictly based on the table attributes provided on the spot, and most attributes should be distinguished and automatically converted through classification coding, layer definition and entity definition, and then their ownership should be determined through the relative position relationship of geometric figures; (3) Edge stitching of map data, which combines different parts of the same graphic object divided by adjacent map sheets into a complete object according to the principles stipulated in the specification. Edge connection processing includes graph edge connection and attribute edge connection, and attribute edge connection is to ensure the consistency of attributes of edge elements in adjacent sheets.

3.3 Field investigation

On-the-spot investigation is the key and foundation to obtain the information of land status. Without solid and meticulous field investigation, even the latest remote sensing image data can not be accurately interpreted. In addition, the boundary investigation of ownership of rivers and villages involved in the project must also be carried out on the spot. The field survey includes administrative boundary survey, ownership boundary survey, sporadic feature survey, linear feature width survey, land type survey, residential area survey, land use unit survey and river name survey. According to the Technical Regulations for Land Use Renewal Investigation and related technical regulations, and the National Three-level Classification Standard for Land Classification (applicable during the transition period), all field investigation information is recorded in the Record Book for Land Use Renewal Investigation.

Fig. 3 flow chart of image processing technology

3.4 Land use database update

The database software used in this update is Aidi Land Use Management Information System. Therefore, firstly, modify the data dictionary, then modify the data of the partitioned index map and the administrative index map, and finally put the files needed for accurate storage into storage, including sporadic feature point files, linear feature line files, land type map files and outline modification annotation point files, and then make statistical summary of the county area, output partitioned land use status maps in a standard format, and print various land summary tables.

4 existing problems and solutions

This project uses SPOT-5 remote sensing image to update the land use database, giving full play to the characteristics of high speed, high precision and wide range of remote sensing image data acquisition, which not only shortens the survey time, but also ensures the survey accuracy and saves the survey cost. However, some problems have also been found in the process of this updated survey, which need further study and solution.

4. 1 image processing and accuracy evaluation

In SPOT-5 remote sensing image processing, the image quality requires rich layers, legibility, uniform color and moderate contrast. Although the image processing of this project meets the work requirements, the image quality needs to be improved. In the long run, if the data source is American fast bird, its resolution is 0.6 1m, which is more than 4 times higher than the 2.5m of SPOT-5, which requires higher image processing quality. It not only directly affects the extraction of change information, but also affects the accuracy of visual interpretation. Therefore, we should accumulate experience and improve the quality of image processing.

Accuracy evaluation is the evaluation standard of digital orthophoto map production quality. Accuracy evaluation refers to the difference between the measured coordinates of field characteristic control points and the image coordinates of the same name points in the image. With the help of GPS, the coordinates can be measured at the same point in the image and compared with the image coordinates, which is an important link in the early stage of updating the land use database with remote sensing images. Only when it meets the national standards can remote sensing images be used for updating.

4.2 Migration processing of vectorized linear objects and screen vector acquisition processing

Through the combination of vector and raster data, the land use database and 1 ∶ 1000 digital orthophoto map are nested and updated, resulting in the problem that the vector linear target deviates from the linear target in the image with the same name. According to the national standard and the technical scheme of land use renewal in Heilongjiang Province, when the database file is nested with the image, when the deviation between the vector line and the linear element with the same name on the image is greater than 1 mm (equivalent to 4 pixels of the SPOT-5 2.5 m image), it is regarded as an offset, and the offset linear element should be corrected. If the width and class code of offset linear features have not changed, they can be corrected by translating and copying in online tools. If the attribute changes, the change information will be verified through field investigation.

On this basis, the screen vector linear objects are migrated through images. At the same time, it also involves the problem of screen enlargement. If the magnification is too large when collecting the screen, the image will be distorted, and the edge contour of the linear object can not be seen clearly, which will affect the offset processing position of the linear object. But when the magnification of the screen is too small, the linear object migration will be incomplete, too rough and inaccurate. After many experiments during the project, it is agreed that the vectorization amplification standard of reference remote sensing images is 0.9 times. In this way, not only the image edges are clear and undistorted, but also the integrity of linear object migration processing is guaranteed.

4.3 The workload of database updating is heavy.

Compared with the traditional updating methods, the advantages of updating the land use database with SPOT-5 remote sensing images are very obvious. However, due to the heavy workload of preliminary preparation and field investigation, manual visual interpretation was mainly used to extract the change information in the early stage. Moreover, the land use database was the result of detailed investigation in the early 1990 s, so a lot of change information was extracted. There are nearly 20,000 changed plots extracted by this renewal project. Moreover, the extracted change information should be verified one by one through field work, and every change information should be recorded in the Record Book of Land Use Update Survey. This workload is also large, and human error is inevitable. Therefore, while using remote sensing images to update the land use database, we should study and explore on this basis, and find more effective methods to extract change information on the computer, reduce the workload in the field and improve the accuracy of the data.

4.4 Improve the data quality inspection mechanism and further standardize the database update.

Data quality inspection runs through the land use database updating project, which is an extremely important link in database construction, and its workload accounts for more than 60% of the whole database construction work. Data quality check is an important tool to ensure the quality of database. Quality inspection includes computer logic inspection and manual inspection, mainly for data structure and accuracy inspection. In this project, a three-level quality inspection mechanism is established, including group self-inspection, first-level inspection and second-level inspection, to protect the project quality.

At the same time of project implementation, the relevant national standards and requirements and relevant specifications promulgated by the provincial level are fully referenced. But in the course of the project, some more detailed work still has no rules to follow. Therefore, it is necessary to further standardize the standards and requirements for updating the land use database and realize data sharing.

4.5 It is necessary to cultivate a high-quality land renewal investigation team.

In the early stage of the project, it is necessary to strengthen the training of the team that updates the land use database. Because the quality of the update survey determines the success or failure of the update project; Moreover, the land renewal survey involves many disciplines such as land management, geoscience, pedology, agronomy, surveying and mapping, informatics and so on. Investigators must have the theoretical knowledge and application skills of the above subjects to ensure the quality of land renewal survey.

Compared with the traditional methods, updating the land use database with SOPT-5 remote sensing image is time-saving, rapid, labor-saving, capital-saving and accurate, which ensures the quality of the updated survey results and provides accurate and detailed data for the revision of the new round of overall land use planning, rural collective land registration and certification, basic farmland protection, land development and reclamation, and construction land approval. Moreover, the application of remote sensing technology will promote the automation of land change survey and the informatization of land resources, greatly improve the accuracy and timeliness of land use change survey data, and fundamentally solve the problem of updating the land use status map.

refer to

Chen Jianping. Investigation on the Update of Land Use Status Based on "3S" Technology. Zhejiang Second Institute of Surveying and Mapping, 2005.

Zhou Fei. Preliminary study on the workflow and existing problems of database construction of land use renewal survey in Guangxi. Guangxi Surveying and Mapping, 2005 (2)

Liu Fengxian, Chen Yinghui Road. Discussion on updating method of county land use status database. Beijing: Geological Publishing House, 2005.

Ren ling. Application of "3S" integrated technology in dynamic monitoring of land use. Planning Department of Land and Resources Bureau of Anyang City, Henan Province, 2005

Study materials of the training course on new techniques and methods of land use survey. Ministry of Land and Resources Cadre Education and Training Center, 2005 (9)

Study on land use change based on RS and GIS. Science and Technology Management of Land and Resources, 2005 (2)

Zhao Aihua, Yang Fenghai, et al. Application of 3S technology in land ecological environment investigation. Beijing: Geological Publishing House, 2005.

Technical specification for land use renewal survey. Ministry of Land and Resources, 2004 (3)

Liu shunxi Land resources survey. Beijing: Contemporary China Press, 2006.