While wide-area 3D Gaussian Splatting (3DGS) shows promise for scene restoration, current implementations face several challenges. There are few examples of its use, and existing 3DGS tools often have limitations when used on mobile devices, including a restricted restoration range, large data volumes, and lower accuracy. To address these issues, this study employed RealityCapture to align digital SLR camera images and point cloud data acquired from a terrestrial laser scanner. By leveraging these processed results within Postshot, a tool capable of large-scale, GUI-based 3DGS processing, our team successfully and accurately restored the vast Metasequoia Plaza at the Setagaya Campus of Tokyo University of Agriculture.

3D Gaussian Splatting (3DGS), a technique capable of generating highly photorealistic 3D representations, is beginning to influence the fields of surveying and civil engineering. However, most prior studies and practical applications have been limited to small-scale structures, and examples of urban-scale implementation remain scarce. Moreover, applying 3DGS to large-scale scenes presents several technical challenges that must be addressed. This paper presents a rare case of urban-scale 3DGS implementation using aerial imagery and examines the primary challenges encountered, along with proposed solutions.

This paper examines the combined use of Scaniverse and the handheld 3D scanner “3DSL-Vega" for recording cultural assets. Developed as a smartphone application, Scaniverse was utilized on an iPhone 12 to capture data, shared via Google Drive. Key successes include the 3D modeling of inscriptions and structures on cultural assets such as lanterns and gravestones. However, limitations remain in capturing finer details. While Scaniverse's Gaussian Splatting mode proved effective for broad-range data acquisition, it showed constraints in obtaining detailed coordinate information. The findings emphasize the effectiveness of combining Scaniverse with a handheld 3D scanner to achieve more precise and comprehensive records of cultural assets.

This paper reports on an attempt to verify the practicality of 3D Gaussian-Splatting (3DGS) technology for plant piping design work. 3D data obtained by Scaniverse is finally imported into 3DCAD, and its applicability to the conventional workflow is verified. The possibility of applying this technology to piping design work in the future is suggested.

In recent years, object detection methods using deep learning have been utilized for player position analysis in sports information science, contributing to performance improvement. However, these methods require a lot of effort in creating training datasets. The annotation system developed by the authors previously improved the efficiency of datasets generation, but full automation was not achieved. Therefore, the present research proposes a semi-automatic generation method for training datasets using image processing technology and AI. And then, we implemented a method to refine the training dataset while building the unique detection model, taking into account the background difference and detection results from the AI model. Ultimately, it was confirmed that as the unique model is refined, it will also be able to semi-automatically generate datasets for learning, contributing to the improvement of the annotation system.

In a mobile mapping system equipped with a phase-shift laser scanner, the point cloud data of road signs appear behind their actual positions. This phenomenon refers to ‘Nonius jump’ by the manufacturer while it is attributed to the retroreflective materials. However, effective research has not been found regarding preventions for the “Nonius Jump” phenomenon. On the other hand, detection of distance measurement errors is performed visually, which requires significant labor. In this study, we propose a method of detecting road signs and evaluating distance measurement errors based on the standard deviation of the relative distance between road sign and the mobile mapping system. The results confirm that the standard deviation of the relative distance can be used to detect road signs with distance measurement errors.