Abstract:
Mobile mapping systems (MMS) capture geospatial data using vehicle-mounted sensors and cameras for generating digital maps from capturing equirectangular panoramic images with high definition. However, the collected images contain unwanted information that needs to be remove for further use. Although object removal approaches have been proposed, these methods may not be well-suited for MMS imagery due to their unique characteristics and processing requirements and require more time processing. In this study, we propose a Generative Adversarial Network (GAN)-based technique tailored to remove unwanted vehicles from MMS scenes. To handle the challenge of inpainting full panoramic scenes, we convert MMS images into spherical coordinate system to extract viewports containing target vehicles and remove target from image. Subsequently, a Pix2Pix-based inpainting network is employed to inpaint the target object in viewport. Our approach has many advantages such as 1) our method can inpaint target object not required information other than information in one image 2) our method preserves the output quality while efficiently processing large raw MMS image datasets. Considering the characteristics of MMS datasets, our approach is appropriate for MMS image datasets that require computational efficiency.