Practice and application of three-dimensional modeling algorithm based on multi-source data fusion in geographic information systems

The practice and application of geographic information systems have a positive promoting effect on the development of digital cities. It can provides high-quality geographic information services, which has profound research significance. Therefore, the study improves the Mask R-CNN algorithm and combines the improved algorithm with 3D modeling to construct a new 3D modeling model for multi-source data fusion. Finally, the 3D modeling algorithm is integrated with the geographic information system, and a novel 3D modeling geographic information system model based on multi-source data fusion is proposed. In the street image segmentation task, the improved algorithm achieved an accuracy of 98.97% and a loss value of 1.08, demonstrating good stability and robustness. The proposed model can not only accurately detect and segment target images of streets, but also finely segment the boundary areas between objects. In practical applications, the research model has achieved the level I modeling accuracy standard in the 3D modeling tasks of multiple typical urban blocks, with the error controlled within a reasonable range. It has good engineering practicability and promotion value. In summary, the research proposes a new three-dimensional modeling geographic information system model based on multi-source data fusion, which has good target image recognition and segmentation capabilities, providing useful suggestions and guidance for constructing digital cities.