An Integrated Rural Mapping using
Remote Sensing and GPS Techniques
Types of GPS Surveying
The GPS, from its inception, has been a surveying and locating instrument. It has found application in different fields, and even in surveying, for different kind of surveying. A few of the applications of GPS in the field of surveying are described in the forthcoming paragraphs.
Geodetic surveying: Dual frequency GPS receivers, in differential mode, have been in use for quite a long time towards geodetic surveying for construction, highway surveying, mine surveying, crustal deformation studies, to name a few. This type of surveying requires sub-cm level accuracy, and in some cases, even of the order of mm level accuracies. Such an accuracy is achieved through use of dual frequency receivers (capable of receiving both L1 and L2 signals), carrier phase post-processing of GPS observables (achieved through post-processing scientific software like Bernese, GAMIT, GYPSY, and vendor specific software). Special care is taken in this case for establishment of base station for the survey, and atmospheric modeling during the post-processing. These care lead to better ambiguity resolution for better accuracies. Advancements in the data processing techniques in the recent past have increased the accuracy to a considerable limit.
GIS Mapping: Single frequency code measurement GPSs and availability of carrier phase measurement facility on these single frequency GPS receivers has led to the entry of GPS for spatial data collection for the GIS projects. These receivers are of medium accuracies, of the order of sub-meter to few meter accuracy levels, and are being used in almost all GIS project, these days. The use of GPS has virtually removed the conventional surveying methods being used in the GIS projects. This has become possible due to the decrease in the price of GPS receivers and the removal of S/A. Another factor due to which these receivers are being widely used is the availability of coast guard beacons all along the coast. With the availability of differential corrections, through these beacons, the need of an additional GPS receiver has almost been eliminated and real time corrections can be applied to the data collection process for better accuracies. Another advancement in this field is the availability of post processing software for the single frequency receivers, which were not available, some time back. Now, the cheap and robust GPS like GARMIN can also be used as a base station as well as rover for data collection through RHINO (http://www.uspositioning.com) and the position can be resolved with an accuracy of few cm. These have eliminated the use of costly GPS receivers and data processing software.
Mobile Mapping: The miniaturization of GPS receivers and computing devices has evolved the technique of mobile mapping. In this technique, any mobile computing device, like palm-top, PDAs and laptops, running a mobile GIS software, is connected to a mobile GPS like Pretec card GPS with or without external antenna. The mobile GIS software can be any like ArcPAD from ESRI (http://www.esri.com), Onsite from Autodesk Inc. (http://www.autodesk.com), HGIS from Stalpal Inc. (http://www.starpal.com), to name a few. This software work with most popular GPS receivers which support the data transfer through NMEA protocol version 2.0 or later and Trimble TSIP protocol through RS232 port. In addition to these commercially available softwares, few free software are also available like Gramchitra, developed by Media Las Asia at CSDMS (http://www.csdms.org). The software Gramchitra, is a LINUX based mobile GIS software available on a open platform and performs most of the functions of popular mobile GIS software through an innovative algorithm. The base map of the area can be loaded at the background, over which the GPS and other spatial data could be collected through an easy to use user interface. The data is stored in its own proprietary format and can be exported to the popular formats like ArcView .shp and MapInfo .tab.
GPS for Rural Mapping: A case study
At the GIS Institute, CSDMS, the utility of low cost mapping using GPS and its integration with satellite imagery, without sacrificing the accuracy, was explored. The work was initiated in a village, named Riwajpur, in Faridabad District of Harayana State, India. The village has about 120 households and with 600 population. The settlement area of the village extends for about 1.5sq km (Figure) . The existing map of the village was prepared by land survey department in 1953 and there was no updation in the map till date. Hence to update the map with low cost technology, in the first stage, GARMIN make e-trex GPS receiver and Casio make Protrek watch-cum-GPS receiver was used. To get the skeleton map of the village, the streets of the village was mapped, using both GPSs in continuous mode and compared. The track map of the village Riwajpur is shown in figure. The important places, in the village, were also marked on the map.. The IRS-PAN image was georeferenced using the co-ordinates obtained using the same GPS and digitally processed. This street map was then overlaid on the 5.8m resolution satellite image. The roads, as visible on the satellite image, were then interpreted. The figure shows the map generated from the satellite image and also image interpretation was carried out to identify other features. Further, this map was corrected, after ground verification and final map was generated. Figure shows the final corrected and updated map. Ground verification in the measurements shows that there is a 5mts error in the horizontal. Hence further processing with little higher accuracy GPS was used for further research.
The process was later refined with Pretec GPS and PDA running ArcPAD. The IKONOS 1mt PAN was merged with 4mt MSS data to get the better spatial and spectral resolution using different image processing techniques. Subsequently the image was georefenced using Pretec GPS reading and was used as the background image for the mapping. The track and the house corners were marked on the image, in the field itself. This expedited the survey process and map making. There was no need of any later ground verification and post-processing of the data. The attribute and the map could be generated simultaneously, with reasonably good accuracy. Figure shows the map created using ArcPAD for the same village.
Discussion
The above-discussed procedure is good for mapping in the rural areas but there are few inherent problem areas and issues, which need to be understood before undertaking such a work.
Problems with Remote sensing procedure
Due to various policy restrictions, it is very difficult to get any high-resolution image/aerial photographs of rural areas and also aerial photo is a not available for all rural areas. The time taken by the data supplying agencies is very high for fresh acquisition or supplying the achieved data. The recent quotation for IKONOS Geoproducts is cost about $30 per square kilometer (www.spaceimaging.com) and minimum order should be 49sq.km. The data have to be processed with the specific image processing software for geo correction and post processing. Even though the cost of the IRS-PAN data is very less compare to IKONOS the information on house level cannot be mapped. This is an important parameter while generating any database. The Geoproduct of IKONOS 1mt resolution data has the horizontal accuracy of ± 50mts. Hence to prepare high accuracy map it requires precision products which is much costlier. The geometric distortion is also high with these images.
Problems with GPS surveying procedure
The GPS used in the study was a single frequency GPS whose accuracy is 30m in stand-alone mode. However, it can be increased with better surveying techniques and post-processing methods. Also the GPS sometimes, does not work in the densely populated areas. Another factor, which need to be kept in mind, is the accuracy for rural mapping. Since the area is very small and the map generated is of the order of 1:2000 or less, the base map used need to be georeferenced using the GPS, being used for mapping. Another factor is the datum. Since GPS gives the data on WGS84 and the maps and images available are on modified Everest ellipsoid, the transformation parameter need to be known, before the GPS data is transformed to appropriate datum. Since the transformation parameter is classified, it is not available to the general users, and it becomes difficult for general users to convert to the local datum, i.e., modified Everest.
In spite of these problems, fairly good maps were created integrating the data of GPS and satellite images, which could be used for developmental planning of the villages.
Conclusion
The comparative research study on various mapping technique carried out in the Riwazpur village reveals that mobile mapping through Pretek GPS is faster and more accurate. The mapping of small lanes in the village settlement areas is an advantage with this methodology since such details cannot be mapped even in the high-resolution data. Similarly the accuracy, cost and time for preparation of map using other technologies is often much higher and more time consuming and hence an attempt was made to use low cost mapping techniques, namely mobile mapping to prepare the rural settlement maps. This was further integrated with GIS for evolving strategies for better governance. However, the present method needs to be further refined for better accuracy in the maps and for integrating maps with the toposheets.
|
