The Critical Review of using Small Format Aerial Photogrammetry (SFAP) FOR Municipal Mapping in Indonesia

The Critical Review of using Small Format Aerial Photogrammetry (SFAP) FOR Municipal Mapping in Indonesia

Soeta'at And Catur Aries R.
Dept. of Geodetic Engineering - Gadjah Mada University
Grafika No. 2 Yogyakarta 55281 - Indonesia
E mail:- c-aris@ygy.centrin.net.id

ABSTRACT
The SFAP have been used in practice as a low-cost mapping activity. Unfortunately, there is not much standardization, guidelines, or researches report yet, that explore the SFAP for the Indonesia condition. This paper will review an experience in using SFAP for mapping project at the Madiun municipality. The review will focus on some implementation problems and the concept to handle those problems in contrast with the purposes of the project. The project objective is to produce an infrastructure data spatial that created by the digitization process from the orthomosaic with 1/10,000 map scale. The explanations include the problem related to the municipal condition, data acquisition, GCP preparation, and the data processing. This kind of project is typically needed by other municipality that has a limited in budget and technical ability also. It is shows that SFAP can be working for the municipal mapping. In the future, some improvement such as camera calibration, automatic tie-point matching, automation in the feature extraction, and DTM generation should be applied to increase the ability of the SFAP.

INTRODUCTION
Due to an autonomy policy in Indonesia, the request for the mapping product from the municipal or the local regency has increased since 1999. Generally, the map use to support the regional planning policy making that need accurate, clear and comprehensive spatial information. The regulation (UU No.10/2000) state that the local regency should have a planning map with scale can be varies 1/25,000 - 1/10,000. Unfortunately, there are not enough budgets available and the technical ability also to do the conventional aerial mapping. So, the SFAP (Small Format Aerial Photography) can be a low-cost alternative to produce map, beside the HRSI (High Resolution Satellite Imagery), and IfSAR (Interferometry Synthetic Aperture Radar). Obviously, there is not much standardization, guidelines, or researches report yet, that explore the SFAP for the Indonesia condition. So, the map accuracy specification can be used as a quality control of the final map product. In the last five years, the SFAP has been used in practice to produce map and the orthomosaic up to 1/5.000 in scale (see. Anto, 2001, Dipokusumo and Aziz, 2004). This paper will explain the experience in using the SFAP for the Municipal mapping. Furthermore, the explanations include the problem related to the municipal condition, data acquisition, data processing, and the maintenance of the final product. This kind of project is typically needed by other municipality that has a limited in budget and technical ability also. Hopefully, this paper can be a consideration for using the SFAP in the future.

THE CONDITION FOR THE MUNICIPAL
Indonesia has land area about 1.965 million squares kilometer that not has been mapped in an appropriate scale for the planning map. In this year there are more than 300 district regency ("kabupaten") and municipality ("pemerintah kota"). According to Amhar (2001), there are more than 131,000 squares kilometer area that have 5,000 people per 1 squares kilometer density. These areas ideally need to be remapped every five years. Shortly, there are still open problems to handle all of those mapping volumes.

Beside the mapping volumes, the following statements describe some specific conditions for the municipal, such as:

there are only a few municipality that have enough budgets to do mapping activities
At present, the need for the map products (topography map) is supplied by Bakosurtanal (National Coordination for Survey and Mapping Agency). Obviously, in general, this map was produced in the past more than five years that up to 1/25,000 in scales. Although, this map can support for planning, the existing information are not complete enough (Dimyati, 2002). There are other resources to get the map, such as the land registration map from BPN (National Land Agency) or PBB (Land and Building Taxation) office. This map can be up to 1/1,000 in scales. Unfortunately, all of those maps not have the same datum references and coordinates system.
To agree a mapping project is not an easy one. Not all municipalities act the mapping products as an infrastructure. They say "the mapping activities are waste money, why can not just use an old map?"
The mapping product should be finished in the one fiscal year.
Generally, the urban area in the municipal has a small range area between 30 - 50 km2. So, working with a conventional aerial mapping can be a high cost.
There are not much human resources that can build a spatial information system to support the government and people activities. There is no database or infrastructure data policy yet.

Shortly, there are not enough budgets available and the technical ability also to do the conventional mapping activities. Furthermore, the production time should be as fast as possible.

Some of the possibility to maintain an update map products for the municipal, such as:

Using a new or low-cost technology in mapping. The SFAP can be a good alternative. Starting this mapping activities on which the people and the government needed.
Promote the understanding that good mapping products are a key to have a good infrastructure data spatial.
Using an existing map product as a base map, so the activities emphasize on the map up dating.
Digitize all the map products to build an infrastructure data or an open source spatial information systems. In long term, this will reduce the cost for the maintenance.

Furthermore, this paper will focused on the SFAP as a low-cost technology in the aerial mapping.

SFAP FOR MUNICIPAL MAPPING
SFAP (Small Format Aerial Photography) is not a new technology (see. Warner, et. al., 1996). At present, the reusing of SFAP is supported by the development in equipment of the digital photogrammetry. SFAP is a low cost, do-it-yourself technique to obtain actual data that can be used for a wide range of applications. For example to detect recent changes caused by disasters, like mudflows, flooding and earthquakes, but also to analyze urban changes, land degradation or land use changes over a longer period, when recent images are not yet available (Nagelhout and Hofstee, 2004). Unfortunately, there are no guideline or technical specifications for applying the SFAP in Indonesia. In 2001, the department of geodetic engineering - Gadjah Mada University had experience to prepare the photo flight specification for SFAP. In practice the SFAP have been used for some researches (see. Bitelli, et., al., 2002, Aber, et. Al., 2002, Misra, 2003, . Anto, 2001, Dipokusumo and Aziz, 2004, Nagelhout and Hofstee, 2004).

The developments in SFAP technology are emphasizing to keep the cost as low as possible and the production time as fast as possible without loosing the geometric accuracy. The experience for mapping in the Madiun Municipality can be good example to critic the use of SFAP technique. The purpose of the mapping activities is to produce an infrastructure data spatial that created from the digitization process from the orthomosaic with 1/10,000 map scale. The following statements are some concluding remarks about the use of SFAP, such as:

Madiun municipality is locating near (< 50km) the Iswahyudi (an Indonesian air force base). It is not easy to get a security clearance. So, the best choice is to use a light aircraft such as Trike that can not disturb the air combat activities. Trike can be taking off and landing at the soccer field. This problem is typical for the municipal that not have an airbase nearly.
For the photo flight activities, it is efficient to use two 35mm cameras (up to 72 exposures) with GPS navigation type as a pilot navigation. It is not easy to keep the exposure position in 30m range from the exposure planning. This caused by using the Trike plane. So, sometimes the overlap area is less than 50%. To compensate this condition, the camcorder can be used as supplement equipment (see Figure 1).

Figure 1. The camcorder as supplement equipment

The bottle neck process in the data processing is the tie-point matching. This due to a large frames pairs that must be processed. For the Madiun Municipality with area about 39km2, it is needed more that 720 photo frames with 1/15.000 in scales.
Generally, the mosaicing from pure rectified frames individually can look seam in the object shape. This is due to distortion that dominantly comes from lens distortion and relief displacement effect. To minimize this, it is needed more GCP (Ground Control Point) and DTM or it can do a camera calibration procedures.
It is hard to derive a good quality DTM from the SFAP. This is due to the small base-high ratio and the missing minimum overlap between frames. The best ways to derive DTM is from a stop and go GPS survey that combined by the old map.
The efficient way to make a controlled image mosaic is to rectify several frames (every 5-8 frames) as a reference frames first (see Figure 2). Furthermore, doing the image to image registration for the others frames with triangulated method. As a consequence, these methods need more GCP.

Figure 2. Sample distribution of the controlled frame
Doing an aerial triangulation process is not an efficient way. It is because a huge frames and loosing the overlap (< 50%) between frames.
Another bottle neck process is the on screen digitizing to interpret the orthomosaic image becomes a thematic map. The automation in this process can be done with image understanding algorithm such as feature extraction. But, this still need an intents communication to the users (government staff). The best way to do this is to train the local staff to handle the manual digitizing process, because there are many sectors with different interest or theme.

Shortly, there are still some open problems to improve the quality of the SFAP. Those improvement should be emphasized on the automation process and deriving DTM.

PROBLEM with GCP
Working with SFAP with huge frames, these need more GCP (Hz, Elev.) to keep a good geometric quality. In urban area just like the Madiun municipality, doing a GPS survey is not a hard one. The problem is how to keep the cost as low as possible for doing a GCP survey. This can be done if we can use a low cost instrument such as a navigation type GPS receiver. Another ways to get the GCP is to use an existing GCP or taking from an existing old map. The best geometry configuration for the GCP is an equally space point that separated every 500m.

According to the map specification, the horizontal position accuracy for the 1/10,000 map scale is about 3m (0.3 mm x 5000) and the elevation accuracy is 2.5m (0.5 x interval contour line). Let's try to evaluate the three methods for getting GPC, they are:

Taking an existing GCP that usually present in urban area. This can be used as a pre marking point that need to be signalized (see Figure 3). This kind of point can be the control points from Land registration Office (BPN) or Infrastruktur department (Kimpraswil).

Figure 3. Examples of signalized an old GCP
Taking the GCP from an old map with 1/1,000 scale. Generally, this is a land registration map that had produced not older than 7 years. This map should be digitize or scanned first and than registered. This map can show some well defined object such as a parcel junction, edge of building, etc. Assumed the error reading from the map is 2.5m (2.5mm x 1,000).
Doing a GPS survey with minimum two GPS navigation type receivers. To improve the accuracy is not use an absolute position, but computed with differential method. The experiences show that using the differential calculation can reduce the position error varies between 1-2m. This accuracy can be accepted to produce map scale 1/10,000.

Shortly, in the urban area, it is possible to get GCP with a low cost operation. This is because we can use an existing GPS or Map, and doing the GPS survey with inexpensive equipment (navigation type receivers). The position accuracy of the GPC still be acceptable for producing the 1/10,000 map scale.

CONCLUSION
In practice, the SFAP technology can be use to produce a map with accepted accuracy for the municipal mapping. This technique can be a low-cost alternative, fast in producing and also meet the need. Some improvement such as automation in tie-point matching, a simply camera calibration procedures, feature extraction, and the DTM generation are still an open problem for future.

REFERENCES

Aber, J. S., Aber, S. W., Pavri, F., 2002. Unmanned small format aerial photography from kites for acquiring large-scale, high resolution, multiview-angle imagery. ISPRS Archives, Vol. XXXIV, Part 1.
Amhar, F., 2001, Transfer Teknologi, Supervisi, dan Aplikasi Pemetaan Digital, Prosiding Forum Geoinformasi dan Pemetaan, Cibinong, Indonesia
Anto, G.,H., 2001, Updating Peta Skala 1:25.000 dengan Metode Pemotretan Udara Format Kecil (SFAP) dan Videogrammetry Dalam Rangka Menunjang Otonomi Daerah, Prosiding Forum Geoinformasi dan Pemetaan, Cibinong, Indonesia
Bitelli, G., Girelli, V. A., Tini, M. A., Vittuari, L., 2002, Low-Height Aerial Imagery And Digital Photogrammetrical Processing For Archaeological Mapping
Dimyati, M., 2001, Beberapa Catatan Tentang Pemanfaatan Peta Rupa Bumi Indonesia Skala 1:25.000 dalam Penataan Ruang, Prosiding Forum Geoinformasi dan Pemetaan, Cibinong, Indonesia.
Dipokusumo, B. S. and Aziz, T. L., 2004, The Use of Small Format Aerial Photography in Disaster Management (Case of Cililin Area Landslide in West Jawa), 3rd FIG Regional Conference Jakarta, Indonesia, October 3-7, 2004
Misra, P., 2004, SFAP for highway feasibility study, available from http://www.GISdevelopment.net, accessed, 2004.
Nagelhout, A. and Hofstee, P., 2004, Creating a mosaic using small format aerial photographs, http://www.itc.nl/ilwis/applications/application26.asp
Warner, W.S., Graham, R.W. and Read, R.E. 1996. Small format aerial photography. American Society for Photogrammetry and Remote Sensing, Bethesda, Maryland, 348 p.

Page 1 of 1