Data Exchange between NRIS and Proposed NSDI-Spatial Data Exchange Format

Data Exchange between NRIS and Proposed NSDI-Spatial Data Exchange Format

Bharat Vaishnav
bgv@ipdpg.gov.in

P. D. Yadav
pdyadav@jpdpg.gov.in

Head, Informatics Applications Division
Space Application Centre (ISRO), Ahemdabad

Abstract
Realisation of National Spatial Data Infrastructure (NSDI) is the current national focus. NSDI is proposed as a single window mechanism for providing access to the spatial data being generated and managed by various agencies in the country. It is visualised as a network of databases consisting of the domain specific databases created and managed by various agencies like Survey of India, Geological Survey of India etc. and a central database having metadata to be managed by NSDI secretariat. Towards this, a national task force has been set up by Department of Science and Technology which is looking at all the related issues including evolution of spatial database standards at various domain servers as well as the metadata standards for providing single window access to the domain databases.

Evolution of mechanisms for exchange of spatial data sets is an integral part of such a system. Towards this, efforts have been made to define the National Spatial Data Exchange (NSDE) formats. These will have to be supplemented with automated procedures for inter-conversion of data from domain to NSDE format and vice-versa.

This paper describes the GIS based automated procedures for two-way data exchange between National (Natural) Resources Information System (NRIS) to proposed NSDE format along with the results obtained by applying these conversion processes for a study area. These have been presented to a sub-group on NSDI standards working under the guidance of NSDI task force and can be considered as guideline for development of such converters between various domain server databases and the NSDE.

The NRIS database design standards consists of various aspects including:

The node database contents along with the naming conventions, associated coding schemes and the structure of data base organisation.
The spatial database design standards including the spatial framework and the spatial and thematic accuracy levels enforced at the time of data input.

The proposed NSDE format is a set of ASCII files, independent of platform or GIS software, which is structured to incorporate various types of data elements like vector data on topography as well as various types of thematic data along with associated attributes data, Image data, DEM data, Coded raster GIS data etc.

Introduction
Geo-Spatial data, along with associated attribute data form the basis of information system useful for natural resources management, development of natural resources, planning of amenities and infrastructure, carrying out disaster management activities and management of environmental crises. Geo-Spatial information is being generated by various agencies like Survey of India (SOI), Geological Survey of India (GSI), National Atlas and Thematic Mapping Organisation (NATMO), Forest Survey of India (FSI), Department Of Space (DOS), National Bureau of Soil Surveys (NBSS), National Informatics Center (NIC) etc. Geographical Information Systems (GIS) is a modern computer based tool for management, analysis and presentation of Geo-Spatial data and extracting information needed for planning and decision making in various disciplines. These data are managed in different formats depending upon the GIS package used and at present there is no national level standard format for representing digital Geo-Spatial and associated attribute data so that exchange of data from one organisation to the other can take place easily and smoothly. Due to non-availability of standard format for Geo-Spatial data exchange, lot of effort requires to be put in converting the data in desired native format of the GIS package being used and this increases additional overheads in database creation in terms of time and manpower.

In order to overcome the above mentioned difficulties, National Spatial Data Infrastructure (NSDI) is proposed as a single window mechanism for providing platform independent and GIS package independent access to spatial data being generated and managed by various agencies. NSDI is visualised as a network of databases consisting of domain specific databases created and managed by various agencies like SOI, GSI etc. and a central database having metadata to be managed by NSDI secretariat. Towards this, a National Task Force is set up by Department of Science and Technology which is looking at all the related issues like evolution of spatial database standards at various domain servers as well as the metadata standards for providing single window access to the domain databases.

Evolution of mechanisms for exchange of spatial data sets is an integral part of such a system. Towards this, efforts are made to evolve the National Spatial Data Exchange (NSDE) format. These will have to be supplemented with automated procedures for inter-conversion of data from domain to NSDE format and vice-versa.

This paper describes the GIS based automated procedures for two-way data exchange between National (Natural) Resources Information System (NRIS) to proposed NSDE format along with the results obtained by applying these conversion processes for a study area. These have been presented to a sub-group on NSDI Standards working under the guidance of NSDI Task Force and can be considered as a guideline for development of such converters between various domain server databases and the NSDE. In the beginning, NRIS database standards and proposed NSDE format are briefly described.

NRIS Database Standards
The NRIS seeks to provide an integrated database for the use of remotely sensed data and collateral information in the framework of a spatial information system for management of natural resources and amenities planning at district, state, project and regional levels.

The NRIS database design standards consists of various aspects including node database contents along with naming conventions associated coding scheme and structure of the database organisation.

The spatial database design standard including the spatial framework and the spatial and thematic accuracy levels enforced at the time of data input.

Proposed NSDE Format
The proposed National Spatial Data Exchange (NSDE) format is a set of ASCII files, independent of platform or GIS software, which is structured to incorporate various types of data elements like vector data on topography as well as various types of thematic data along with associated attribute data, remotely sensed data, DEM data, coded raster GIS data etc.

The proposed NSDE format is extension of Digital Vector Data (DVD-3) format devised by the Survey of India to represent SOI toposheet data in digital form along with the point, line and polygon topology describing relationships among the spatial features point, line and polygon. The proposed format also has provision to include digital images acquired by satellites and Digital Elevation Model (DEM) and coded raster data. This format is further modified for accommodating attribute data associated with spatial features and other information.

Study Area Identification
In order to carry out NRIS to NSDE conversion and vice-versa, the study area belonging to toposheet number 46F/11 covering a part of Panchmahals district of Gujarat state is selected. Its boundaries are as follows:
Latitude (dd-mm-ss ) : 22-25-00 to 22-30-00
Longitude (dd-mm-ss) :73-30-00 to 73-37-30

The NRIS spatial data are of three categories : Point, Line and Polygon. The conversion exercise is carried out for each of the data category. Settlements representing point data, Roads representing line data and Landuse data representing polygon type are taken from NRIS database, created using ARC/INFO GIS software, for the study area belonging to Panchmahals District of Gujarat.

NRIS to NSDE Conversion
The NRIS to NSDE conversion process has two major steps:

Generation of intermediate files giving information about constant parameters of the district or study area (distconst.dat), variable parameters of the district or study area (dist_vari_param.dat), files containing information about theme n (DATACAT_n.dat) and files containing actual spatial and attribute data about theme n (DATAFILE_n.dat) where n is the serial number of the theme.
Generation of VOLDIR, GENINFO, DATACAT and DATAFIL files as per proposed NSDE format by making use of the files generated in step (1).

The two steps for converting NRIS database contents to NSDE are described below:

Generation of intermediate files
The process of generating intermediate files required for NRIS to NSDE conversion is accomplished by a set of Arc Macro Language (AML) based procedures. The coverages selected for conversion are provided as input and intermediate files are generated in the following manner:

Extraction of spatial data belonging to AOI: If the area of Interest (AOI) does not belong to entire district, it is required to extract spatial data belonging to AOI from the district database. This is done by clipping the coverages against the AOI coverage and using the clipped coverages for further processing.
Generation of AOI Tics: This process is required to be carried out only when tic file of the NRIS coverage is for full district and the area of interest for conversion to NSDE is part of that district. This process selects only those Tic points which belong to area of interest.
Generation of AOI Attribute Tables: This process is also required to be carried out when the area of interest for conversion to NSDE is subset of the district. The attribute information belonging to area of interest only is retained. For example, if the full district have 1000 villages, attribute files will contain information about all 1000 villages. If the area of interest have only 50 villages, this village will remove attribute information about 950 villages lying outside the area of interest and retains information relevant to 50 villages.
Preparation of Inputs: This process prepares inputs required for generating file DATAFILE_n.dat. It generates NODE information, adds labels to Polygon Attribute Table (PAT) if the coverage type is polygon, ungenerates coverage and stores ungenerated co-ordinates to intermediate file for further use.
Generation of Constant Prmameters: This process generates constant parameters for NRIS entire district node or for area of interest like name of the state, name of the district, projection related information, data source, tic points information and writes all these parameters in distconst.dat file.
Generation of DATACAT and DATAFILE: This process generates dist_vari_param.dat file containing information about number of coverages (p) selected for conversion, their serial number n, name of the coverage as per NRIS standards and number of attribute tables in the coverage. For each coverage, separate data category file (DATACAT_n.dat) file is generated giving information about coverage name, number of major categories as per NRIS standards, their code and description, number of NODE, LINE and AREA records, structure of each attribute table in terms of field description, number of data records in each table etc. Similarly for each coverage, separate data file (DATAFILE_n.dat) is generated by this process which contains data about NODE, LINE AREA and ATTRIBUTE records.

Generation of VOLDIR, GENINFO, DATACAT and DATAFIL files
The intermediate files generated in step (1) contain all the information required for NSDE conversion. Using the distconst.dat, distvari_param.dat and DATACAT_n.dat and DATAFILE_n.dat files of respective coverages, this process generates VOLDIR, GENINFO, DATACAT and DATAFILE files as per the structure and contents.

The flowchart for NRIS to the NSDE conversion is given in Figure 1.

Fig 1: Flowchart for NRIS to NSDE conversion

NSDE to NRIS conversion
The NSDE to NRIS conversion process consists of two steps:

Generation of intermediate files
This process generates distconst.dat, distvari_param.dat and DATACAT_n.dat and DATAFILE_n.dat files of respective coverages from VOLDIR, GENINFO, DATACAT and DATAFILE files. These intermediate files are required to generate coverages as per NRIS standards.

Generation of coverages as per NRIS Standards
Generation of coverages as per NRIS standards is accoplished by a set of AMLs in the following manner:

Reading distconst.dat file for collecting the following information about Mapsheet-Id, Year of the Mapsheet, Project Name, State name, District name, Scale , Projection Parameters, Tic-points information etc.
Creation of a temporary tic file viz. tmptic.dat, which remains same for all coverages of area of interest.
Reading the information about number of coverages, their names, serial number etc. from dist_vari_param.dat file.
For each coverage in the dist_vari_param.dat file,
Information about number of major categories and their description, number of NODE, LINE, AREA records, number of attribute tables associated with the coverage etc. is obtained from the DATACAT_n.dat file where n is the serial number of the coverage.
Featutre type of the coverage (point, line or polygon) is determined.
Depending upon feature type of the coverage, the coverage is generated by reading NODE, LINE and AREA records from DATAFILE_n.dat file and appropriate point, line or polygon topology is built for the coverage.
Structure of each attribute table (total number of fields, type of each field, width of each field, number of decimal places, whether the field is a key field or not etc.) associated with the coverage is read from the DATACAT_n.dat file and accordingly attribute table is generated.
Feature codes as per NRIS standards are assigned to each point, line or polygon feature by reading information available in the DATAFILE_n.dat file.
For each attribute table (including LUT) associated with the coverage, information about number of data records in the attribute table is read from the DATACAT_n.dat file and the table is populated with those many ATTRIBUTE records from available in the DATACAT_n.dat file.
Projection of the coverage is defined by making use of the projection parameters like projection name, units, spheroid, centre latitude, centre longitude etc.
TIC file is generated for the coverage.

The flowchart for proposed NSDE format to the NRIS Conversion is given in Figure 2.

Fig 2: Flowchart for NSDB to NRIS conversion

Results of NRIS to NSDE conversion
The results of NRIS to NSDE conversion contain VOLDIR file, GENINFO file, DATACAT file and DATAFILE file as per the proposed NSDE format. The results contain information about three themes namely LANDUSESMALL (polygon type), ROADSSMALL(line type) and SETTLEPSMALL(point type) themes. Each file contains several records having length of 72 bytes. The contents of each type of file is described below briefly:

File VOLDIR: This file contains information about Map Sheet Number of Area of Interest and number of NRIS themes whose information is contained along with their names, serial number etc.

File GENINFO: This file contains information about Map Sheet Number of Area of Interest, Type and Format of the data, Name of the State, Name of the District, Year of map publication, Projection parameters, Name of the project and data generating agency, Map Sheet Numbers of surrounding areas, Information on Tic-Marks and H/W and S/W environment used for conversion.

File DATACAT: This file contains information about Map Sheet Number of Area of Interest; Number of major categories, their codes and description; Number of Attribute Tables associated with the theme, number of NODE, LINE and AREA records; Name and structure of each attribute table etc. The information in this file is useful in knowing and retrieving data contents available in DATAFILE. For convenience, DATACAT file contents of three themes is shown separately.

The LANDUSESMALL contents show that there is one attribute file (LUSESMALL.LUT) associated with the theme, there are four major categories of landuse namely Agriculture, Forest, Wastelands and Water_bodies, there are 12 records in the attribute file, 396 NODE records, 519 LINE records, 171 AREA records and there are 6 fields in the attribute file namely LU-CODE, DISCR-L1, DISCR-L1, DISCR-L1, DISCR-L1 and SYMBOL. Here LU-CODE is the key-field.The type of each filed is also given.

The SETTLEPSMALL contents show that there are nine attribute files(SETTLEPSMALL.LUT,DEMO_91SMALL.DBF, EDUFAC_91SMALL.DBF,VDI_ 91SMALL.DBF etc.) associated with the theme, there is one major category of settlement type namely Village, there are 4 records in the first attribute file, zero NODE records, 53 LINE records, zero AREA records and there are 4 fields in the attribute file namely SCODE, LOCATION, V-TYPE, and SYMBOL. The type of each filed is also given. The structure of the other eight DBF attribute files are also given in terms of total number of attribute records in each file, their field names, type of each field etc. Here SCODE is the key-field.

The ROADSSMALL contents show that there is one attribute file (ROADSSMALL.LUT) associated with the theme, there are three major categories of roads namely Metalled_Black_Topped, Unmetalled_Water_Bound, and Others, there are 3 records in the attribute file, 13 NODE records, 11 LINE records, zero AREA records and there are 4 fields in the attribute file namely RD-CODE, TYPE, SUB-TYPE and SYMBOL. Here RD-CODE is the key-field. The type of each filed is also given.

File DATAFILE: This file contains information about Map Sheet Number of Area of Interest and contents of data values of NODE, LINE, AREA and ATTRIBUTE records.

The contents of DATAFILE for LANDUSESMALL theme show that there are 396 NODE records containing Node-Id, X-coordinate and Y-coordinate for each node. There are 519 LINE records showing Line-Id, Starting and Ending Node-Ids, Number of intermediate points in the line and the X and Y Co-ordinates of intermediate points, if any. There are 171 AREA records containing Polygon-Id, Centroid Co-ordinates, Number of lines(arcs) forming the polygon and their Line-Ids and LU-CODE of the polygon. There are 12 ATTRIBUTE records containing information about landuse codes and their description as per LEVEL-1, LEVEL-2, LEVEL-3 and LEVEL-4. The last value in each attribute record contains SYMBOL number.

The contents of DATAFILE for SETTLEPSMALL theme show that there are 53 NODE records containing Node-Id, X-coordinate and Y-coordinate for each node. There are 53 LINE records showing Line-Id, Starting and Ending Node-Ids, Number of intermediate points (which is always zero for this point theme) and SCODE for each line. The contents of 9 ATTRIBUTE files follow for 53 villages. This includes SCODE, Name of the village, Habitation type, Demographic information etc.

The contents of DATAFILE for ROADSSMALL theme show that there are 206 NODE records containing Node-Id, X-coordinate and Y-coordinate for each node. There are 163 LINE records showing Line-Id, Starting and Ending Node-Ids, Number of intermediate points in the line and the X and Y Co-ordinates of intermediate points, if any. The RD-CODE for each line is also given. There are 3 ATTRIBUTE records containing information about road codes and their description in terms of TYPE, SUB-TYPE and SYMBOL. Here RD-CODE is the key-field.

Results of NSDE to NRIS Conversion
The process of NSDE to NRIS conversion described earlier was applied to re-generate these three NRIS coverages (with different names) whose information is stored in VOLDIR, GENINFO, DATACAT and DATAFILE files. The coverages were successfully generated including attribute information associated with them. From the projection information and Tic points information given in GENINFO file, the coverages were generated with projection and Tic points as per NRIS Standards. The geometry of the coverages was checked by displaying original coverages and the converted coverages one over the other. There was no mis-match in the geometry. The key-fields of the coverages were also properly generated in the graphic as well as attribute files. The topology of the coverages in terms of number of nodes, number of arcs(lines), number of polygons(areas), etc. was also perfectly created.

Conclusion
Procedures for NRIS to proposed NSDE conversion and vice-versa has been developed and tested for a study area. This procedure will provide means of representing geo-data and associated attribute information in a format that is open and not proprietary to any particular GIS package. Thus a Standard Exchange format has been evolved and procedures are developed to convert the data to Standard Exchange format and re-generate the data from Standard Exchange format. This will help in representing Geo-data and associated attribute information in a project-independent, GIS package independent neutral format so that data among different projects and organisations can be shared easily. This exercise will serve as guideline for development of such converters between various domain server databases and the NSDE.

Acknowledgements
The authors are thankful to Shri A. K. S. Gopalan, Director SAC and Shri A. R. Dasgupta, Dy. Director, SITAA for giving opportunity to work for developing these procedures.

Refererences

National Geo-Spatial Data Infrastructure, Lt. Col. (Dr.) R Siva Kumar, Member Secretary, NGDI Task Force. Article published in ISG Newsletter, page 20-23, March/June 2001.
NRIS Node Design and Standards, SAC/SIIPA/NRIS-SIP/SD-02/98, July 1998.
Standard Exchange Format for Digital Vector Data, SOI Document.
Dasgupta A. R. and Goel R. K., 2002, National (Natural) Resources Information System - Design Concepts, MAPINDIA-2002, New Delhi.
Goel R.K, 2002, Suggested Approach For Evolving NSDI Standards at Domain Servers�, MAPINDIA-2002, New Delhi.
Shah S. A., Yadav P. D. and Goel R. K., 2002, National Spatial Data Exchange Format under NSDI, S. A. Shah, P. D. Yadav and B. G. Vaishnav, MAPINDIA-2002, New Delhi.