ArcGIS World Imagery 기반 고해상도 위성영상 데이터 센서모델링

High-resolution satellite imagery is widely used as a core spatial information resource in various fields such as national land monitoring, urban planning, and terrain analysis; therefore, an accurate sensor model is essential. Although RPC (Rational Polynomial Coefficients) provided by high-resolution satellites offer high versatility, the initial RPC often contain positioning errors ranging from several meters to tens of meters due to attitude errors, time synchronization issues, and other factors during image acquisition. Consequently, the acquisition of GCP (Ground Control Points) for post-processing bias compensation of RPC is indispensable. However, field-based GCP collection is time-consuming and costly, and its application is further constrained in areas with limited accessibility. World Imagery is a globally mosaic dataset that integrates and continuously updates high-resolution orthorectified images from multiple sources, demonstrating high potential for use in sensor modeling of both domestic and international satellite data. Accordingly, this study proposes an automatic GCP generation and RPC bias compensation method using World Imagery as reference imagery for satellite images. Experiments were conducted using KOMPSAT-3A data over the Daejeon area. Prior to applying World Imagery for satellite image bias correction, its positional accuracy was evaluated using aerial orthophotos, confirming a root mean square error of approximately 3.2 pixels(approximately 1.6m). Subsequently, when used as GCP for KOMPSAT-3A imagery, the proposed approach enabled sensor modeling with a precision less than 2 pixels(approximately 1.5m), indicating potential for application in overseas and inaccessible regions.