A COMPARATIVE ASSESSMENT OF REMOTELY SENSED AND CONVENTIONAL FIELD DATASETS FOR LAND TITLE REGISTRATION IN AKURE, ONDO STATE.

Abstract

Nigeria had operated a client-initiated title registration before 2009 called the Sporadic approach of Land Title Registration; this method has been adjudged not to be effective leading to the Federal Government of Nigeria (FGN) introducing fit-for-purpose Systematic Land Title and Registration (SLTR) as against the Sporadic Land Registration (SLR) method without much research into its workability. The aim of this research is to carry out a comparative assessment of remotely sensed and field datasets for Systematic Land Title Registration (SLTR) in Ondo State, Nigeria in order to verify if geometric information acquired from remotely sensed dataset is a suitable alternative to field dataset for production of cadastral survey plans used for SLTR. The research delineated boundaries of parcel of land on remotely sensed data, carried out field measurements of boundaries; and evaluated the positional accuracy of the boundary points computed from both methods. The remotely sensed data used include World-View 4 satellite image of 0.65m resolution and a Plaiedes satellite image of 0.50m resolution covering the study area. The field dataset was acquired using a differential GPS receiver. The delineation of boundaries on both satellite images was carried out through digitizing of the images while the differential GPS in a static mode was used to determine the horizontal coordinates of the boundary points on the field. Positional accuracy was assessed for both methods, Pearson correlation index was used to analyze the similarity of area computed from both methods while T-test was used to test if there is a significant difference between coordinates of boundary points computed from both methods. Comparatively, the plots constructed from the field survey method look similar in visual analysis to the ones digitized. Digitized boundary lines were not totally aligned with the result of field survey but the boundaries digitized from the image with higher resolution (Plaiedes - 0.50m) were more aligned than that of the lower resolution(World-View 4 - 0.65m). Areas computed from the methods showed a strong positive correlation of 0.98 and 0.96 with field survey coordinates indicating that the similarity of positions of parcel of land delineated from satellite imagery is high, but increases with the resolution of the imagery used. A larger percentage of 74% and 72% of the boundary points delineated from both satellite images have their displacement outside the expected displacement and so precision of acquiring coordinates from the image is not guaranteed. The Ttest is significant for both images (P < 0.05), hence it was concluded that cadastral data acquired from remotely sensed data is significantly different from those obtained from field survey and by extension we accept the Null Hypothesis (Ho1) which states that remotely sensed data cannot be effectively used in determining cadastral boundaries.