Currently, immersive visual analytics is increasingly being applied in scientific research across various fields, becoming one of the important technologies to assist in scientific work. However, traditional immersive visualization and interaction methods have not fully considered the association between data and physical entities, especially how to quickly identify and track physical entities, with challenges including scarcity of computational resources and dynamic changes in perspectives in complex scenes. In addition, current registration and tracking algorithms such as SIFT, SURF, and K-Nearest Neighbors, which tend to have problems like feature point redundancy, low matching efficiency, and large search spaces.

To overcome these challenges, researchers from the Advanced Interactive Technologies and Applications Development Department of CNIC innovatively proposed a high-precision feature extraction and 3D registration tracking method based on the Grid-ORB algorithm. This method successfully achieves precise perception of physical entities on resource-constrained devices, significantly enhancing scientists' cognitive abilities regarding physical entities in immersive environments. At the same time, the team designed and implemented a Kriging method with a drift term, effectively filling in gaps in numerical physical space data, allowing scientists to more intuitively observe real-world physical values and their trend fluctuations, greatly improving the accuracy and operability of data analysis.

This work was titled "Immersive Situational Analysis Method Based on Generalized Augmented Grid Statistic," and published in Applied Soft Computing (SCI District 1, Chinese Academy of Sciences). The first author is Zhang Yue, a master's student at CNIC, and the corresponding author is Researcher Tian Dong at CNIC.

Figure 1: Algorithm workflow

Figure 2: Using virtual and real fusion for physical experiment situated analysis display diagram

Related Achievements:：

Zhang Y, Shan G, Zhang J Z, et al. Immersive Situational Analysis Method Based on Generalized Augmented Grid Statistic[J]. Applied Soft Computing, 2024, 162: 111651.

Paper URL：

https://doi.org/10.1016/j.asoc.2024.111651