リモートセンシングとGISのジャーナル

概要

Analysis and Correction of Vector Data Generated from Multi-Temporal Satellite Imagery of Google Earth by Means of Mathematical Formulas

Mrityunjay Kar, Sunil Kumar Aggarwal, Ralte Lalnun Sanga and James Singh Thoudam

In this era of twenty first century, Google Earth (GE) delivered tangible advantage to the users. The reliability of people searching the geographical location of the earth on GE service increases and to use its geospatial information for different mapping purposes due to its high spatial resolution of raster data and other ancillary information. The study aimed to generate three separate vector layers from three different time-series satellite data of GE on the same geographical location of the earth. Since the vector layers were superimposed in a particular frame, it was observed that layers were found to be not congruent, as the multi-temporal satellite imagery of GE were shifted from one another. In order to examine the shifted error and to rectify the geometric distortion of vector data, the mathematical formulas were used in the study. Initially, Haversine formula was used to measure the shifted distance between the corresponding points of vector layers. After calculating the distance values of two corresponding points, Lagrange form of Interpolation Polynomial formula was applied to minimize the distance value of vector layers. However, this formula did not provide a satisfying result to reduce the average distance value of vector data. Finally, Affine transformation formula was fit to reduce the distance value and to rectify the geometric distortion of vector layers in comparison with Lagrange form of Interpolation Polynomial formula. Therefore, in order to obtain the correct vector data, the geometric correction of data was required for any ‘Change Detection’ study on multi-temporal satellite imagery of GE.