Impact of the spatial distribution of ground control points on the image registration of remotely sensed images based on a Voronoi modelling
摘要
The exponential growth of remote sensing satellite launches worldwide is the result of breakthroughs in space technology. A high-resolution camera produces high-resolution satellite images, and the sensors consistently generate vast amounts of satellite data. Satellite representation is an essential method for gathering geospatial data due to its high efficiency and extensive data inclusion, without being constrained by spatial parameters. The acquisition of control points using this data is a crucial step in the geometric correction and registration of satellite images. In this study, a Voronoi diagram (also known as a Thiessen polygon) was used to analyze the impact of the quantity and spatial distribution of ground control points on the accuracy of image registration and correction. To examine the effects of various control point distribution patterns on image registration and correction, a simulation study employing remote-sensing images was conducted. The adjusted findings were calculated using residual separation and root mean square error (RMSE). This demonstrates that the RMSE decreases the most owing to the center distribution. Furthermore, the control points scattered throughout the middle of the image were less distorted than those distributed along the borders and corners. Considering the overall deformation rate of the entire structure, a centralized, uniform distribution of the ground control points yield better results.