Method for calculating traces of vehicle in multi-camera scene

Abstract

The invention provides a method for calculating traces of a vehicle in a multi-camera scene, comprising the steps of identifying vehicle monitoring videos collected by the cameras by use of a convolutional neural network to acquire identification information of a target vehicle, wherein the identification information includes a model, a brand, an actual running direction and an inclination angle of the target vehicle, and a position of the target vehicle is labeled by a matrix frame; tracking the target vehicle according to an overlapping area of the matrix frame and a tracker storage region to acquire motion information of the target vehicle, wherein the motion information includes an occurrence position, a departure position, a running direction and a running trace of the target vehicle; calibrating a topological structure of a camera network according to the motion information of the target vehicle, mounting positions of the cameras and a path connected relation; and associating the traces of the target vehicle in the camera network through Bayesian estimation. The method improves the accuracy rate of calculation of the traces of the vehicle in the multi-camera scene.

Description

technical field [0001] The invention relates to the technical field of vehicle trajectory calculation, in particular to a vehicle trajectory calculation method in a multi-camera scene. Background technique [0002] In the current intelligent transportation system, the identification and tracking of vehicles has always been a core link, and it is also a difficult point. There are many studies on vehicle tracking under a single camera. [0003] The existing vehicle trajectory calculations are generally limited to using the attributes of the vehicle itself, such as the turning angle, to estimate the vehicle's driving trajectory, and the accuracy of the calculation results is not high enough. Contents of the invention [0004] The present invention provides a vehicle trajectory calculation method in a multi-camera scene to solve the above technical problems. [0005] A method for calculating a trajectory of a vehicle in a multi-camera scene according to the present invention,...

AI Technical Summary

Problems solved by technology

[0003] The existing vehicle trajectory calculations are generally limited to using the attributes of the vehicle itself, such as the turning angle, to estimate the vehicle’s driving trajectory, and the accuracy of the calculation results is not high enough.

Images