Objectives  Compared with the outdoor space, the indoor three-dimensional space structure is complex, and indoor crowd gathering is more likely to lead to safety accidents.As the development of indoor positioning technology, it is possible to collect indoor trajectory, which also provides an important data source for the identification of indoor crowd gathering area.

  Methods  We propose a novel method called IndoorSRC (indoor simplification reconstruction cluster) to detect indoor crowd gathering areas. Firstly, a new indoor user trajectory simplification algorithm, indoor spatial-temporal agglomerative nesting (Indoor-STAGNES), is designed to identify indoor user stay points and simplify indoor user's trajectory. Then, an indoor trajectory reconstruction method based on Kalman filter is constructed to realize the alignment and resampling of indoor user trajectories. Finally, a new indoor space-time density clustering algorithm: Indoor spatial-temporal ordering points to identify the clustering structure (Indoor-STOPTICS) is proposed to find indoor three-dimensional space-time crowd gathering area.

  Results  The real shopping mall indoor trajectory data are used for experimental analysis.The experimental results show that: (1)The gathering areas in the mall are mostly concentrated in the noon time and are mainly located in the dining area.(2) Compared with the traditional outdoor identification method, the error of recognition can be reduced by 23.7% in the case of IndoorSRC with little difference in running time.

  Conclusions  IndoorSRC can provide an effective supplement to the indoor crowd gathering area identification, and provide technical support for early warning and emergency rescue of indoor safety accidents.