Intelligent ship-bridge collision avoidance algorithm research based on a modified Gaussian Mixture Model

Abstract

Faced with ever more serious security situation in bridge area, A Moving ship target detection and tracking method based on the complex background of the bridge area is presented. The background model based on a modified Gaussian Mixture Model. The algorithm upgrades the background model in real time, Foreground area is extracted by adaptive threshold method by marking each moving region in the binary image, the geometrical features parameters such as width, height, centroid position and speed can be extracted, which lay the foundation for the analysis and track of moving. Ship motion situation calculation module based on visual geometry model, and calculates the movement state information of target ship's position in space coordinate, Ship's size, speed, heading etc. Collision risk prediction and collision avoidance maneuvering decision-making module combined ship real-time motion state information, and Bridge area ship's motion, introducing the collision risk to rank term, and early warning and collision avoidance maneuvering decision-making. Through calculation of ship maneuvering and collision avoidance related distance parameter of bridge area, Introducing collision risk to rank terminology, Forming rationalization collision avoidance decisions, Thus reach the ship-bridge collision avoidance warning and precontrolling effect. Experimental results show that the proposed method is robust, real-time and accurate.