Multi-Target Correlation Estimation Technology Based on Line of Sight Matching

In order to solve the problems such as scale, shape change difference, and target masking in different detection fields of view when multiple vehicles perceive multiple targets cooperatively, this paper proposed a multi-target correlation estimation technology based on sensor attitude angle and target line of sight. Firstly, the line of sight angle can be converted to a reference coordinate based on the camera attitude angle when the target body line of sight angle is acquired. Afterwards, the direction vector of the target at different detection line of sight can be calculated based on the direction cosine of the target in the reference coordinate. Because the directional vectors of the same target from different detection view have the characteristic of being in the same plane and having a unique intersection point, this paper employs this characteristic to determine the consistency of the target. At the same time, this article will also design a target consistency matching strategy under multiple detection view to reduce the impact of sensor attitude angle measurement errors and target angle measurement errors on matching accuracy. Finally, the experiment shows that the method proposed in this paper can effectively achieve multi-target consistency recognition under multiple detection view.