RETRACTED ARTICLE: Spatiotemporal Geospatial Modeling for 3D City Based on Edge Computing Using Statistical Framework Linear Injection Model Based on Multi-data Source Fusion

Abstract

In recent years, data on individual spatiotemporal levels have been increasingly available, especially due to the utilization of modern GPS tracking devices. Cloud pollution significantly restricts the potential use for geoscience applications of optical remote sensing images. This paper presents a spatial and time-based melting process for remote imaging using a model of linear injection and information for neighborhood with edge computing assistance. The detailed data have proven helpful for inferring a more precise degree of transmission than conventional methods. Nevertheless, a lack of statistically sound frames remains for modeling mechanistically for the underlying transmission process and this development is particularly important to provide an individual general predictive epidemic framework. In this paper, a New Statistical Framework Linear Injection Model Based on Multi-Data Source Fusion method is proposed to determine the transmission of cloud-free image and to enhance the noise robustness. To improve its noise strength further, the local neighborhood information is introduced for refining initial fused stage image and to obtain more accurate prediction results, derived simultaneously from resolution fine image and fused result. To improve the spectral cohesion of recovered and remaining areas, correction residual strategy based on the Poisson equation has been analyzed. The tests demonstrated that the methods proposed for cases with substantial land cover modifications can work very well; they have less biases and more reliable performance compared with several cutting-edge approaches. To sum up, our solution is a major technical improvement to the current cloud removal system and offers the possibility of handling scenes with substantial land coverage changes.