Abstract:
In the use of the space–time permutation scan method, if we just measure the spatial straight line distance between two observation points, the distance will ignore factors such as buildings, parks, and other obstacles in the city, causing deviations from the actual distance between the two points, which will influence scan results. This paper proposes a space-time permutation scan optimization method based on temporal distance. Firstly, we can obtain the actual path between two observation points based on the influence of obstacles on the actual path in the city. Then we calculate the actual travel time between the two observation points. Finally, the travel time is used to represent the actual distance between the two points. The San Francisco fire dataset will be used to verify the optimization method in this paper. The model performance is generally improved by 2%~5%.