Accuracy assessment of Preliminary Index Diagrams (PIDs) from High Resolution Orthoimage

Abstract

Land tenure reforms and modernization to conform to current land management and development demands remain a challenge facing developing countries today. Promotion of land titling is the starting point to achieve this target Land titling exercise relies on some form of description anellor rough sketches known as cadastral maps. In Kenya, various types of cadastral maps have been used for land administration; the most famous being Registry Index Maps (RIMs).

This is due to their ease of production by simple surveying , techniques and air survey methods. RIMs produced from unrectified photographs referred to as Preliminary Index Diagrams (PIDs) were initially intended as a temporary measure to speed up land registration in the rural areas pending preparation of more accurate documents, are still in use today. This has resulted in unreliable and lack of up to-date survey information relating to property boundaries in the majority of rural parts of Kenya.

In this context, the study investigated a new system that would facilitate quick production of reliable, accurate and up to-date cadastral maps for land administration. Earlier studies recommended the use of high spatial resolution satellite imagery such as QuickBird amongst many others. The study investigated the potential use of high spatial resolution QuickBird satellite imagery data for cadastral surveying. Three types of data were integrated to provide the database; namely QuickBird orthoimage, aerial digital orthophoto and measurements of parcels from RIMs.

The evaluation was made by statistically comparing parcel areas from the official PID Area List, orthophoto and QuickBird satellite orthoimage. Double tailed t-statistics was used to assess for any difference in areas. The results at 99% confidence interval (p = 0.01) indicated that there was no significant difference between parcel areas from orthophoto and satellite orthoimage while there was significant difference between PID and orthophoto areas. Good results were obtained for large and medium parcels with an average area difference of 0.3% and 1.0% respectively and 2.6% for small parcels.

However, with regard to the minimum requirements for a Land Registry Index Map to be of sufficient accuracy to perform its core functions of parcel identification, boundary relocation, mutation surveys and area computation, it can be reasonably concluded that PIDs from QuickBird orthoimage at a scale of 1 :5000 met these requirements. Therefore, the study has demonstrated that high spatial resolution QuickBird satellite imagery data can be used as an input for indirect land surveying methodology.