AN ACCURACY ASSESSMENT OF REMOTELY SENSED AND CONVENTIONAL FIELD DATASETS FOR LAND TITLE REGISTRATION

Abstract

In Nigeria, the transition from the sporadic Land Title Registration (SLR) to the Systematic Land Title Registration (SLTR) is aimed to enhance effectiveness and ability to create wealth with land. It is often said that “land without registration is a dead asset”. This study compares field and remotely sensed datasets for SLTR in Kwara State, Nigeria. Orthophotos from UAV and remote sensing were used to delineate land boundaries, compared against field measurements using differential GPS and Total Station. Positional accuracy was assessed for both methods, Pearson correlation index, Standard deviation and T-test were used to analyze and test if there is a significant difference between coordinates of boundary points computed from the three methods. The coordinates of the points generated through digitizing of UAV image and Google earth image gave positive correlation coefficients 0.95 and 0.98 respectively. The google earth data had 62% of its displacements outside the expected displacement while Drone data has only 5%. Analysis showed that while digitized boundaries from UAV images closely matched field data, Google Earth digitization differed more. Statistical tests indicated significant differences between coordinates from all three methods, highlighting challenges and strengths in cadastral survey accuracy from remote sensing data