Pedestrian flow density and flow prediction system and method based on convolutional neural network

The invention relates to the field of intelligent traffic, and particularly discloses a pedestrian flow density and flow prediction system and method based on a convolutional neural network, and the system comprises a coordinate transformation module, a sampling module, a track embedding module, an...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHANG XIAO, NONG TINGTING, WANG TAO, SHI MENG, ZHANG WENKE, TAN JINGYU, LAN WENFEI
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention relates to the field of intelligent traffic, and particularly discloses a pedestrian flow density and flow prediction system and method based on a convolutional neural network, and the system comprises a coordinate transformation module, a sampling module, a track embedding module, an encoding module, a decoding module, a track output module and a density and flow calculation module. Firstly, data are converted and sampled, then an input pedestrian trajectory is embedded into a vector with richer information based on a multi-layer perceptron, then an encoder performs encoding and feature extraction on the trajectory vector by adopting a convolutional neural network structure, and a decoder performs decoding through convolution operation after receiving the feature information. Then, a trajectory output module converts the feature tensor into trajectory coordinates in a manner opposite to that of the trajectory embedding module and outputs the trajectory coordinates, and finally, a density flow c