Research On Detection Of Small Moving Targets At Airports Based On Statistical Optical Field

This article discusses the potential threat posed by small moving objects such as birds and drones invading the airport runway area,as well as the current situation of the airport’s limited bird control equipment and expensive monitoring radar.The article proposes a solution based on statistical light field for detecting small moving targets at the airport,parallel-axis stereo cameras for distance measurement,and the development of a real-time monitoring system hardware and software.This approach enables real-time trajectory detection of small moving targets at the airport,the construction of optical bird detection equipment,and the establishment of a long-distance bird situation normalization visual monitoring system.Firstly,an analysis is conducted on the imaging characteristics of small moving targets in a long-distance and wide-field airport environment.The fluctuation sequence signals of the optical field for small moving targets are studied,and the fluctuation characteristics of the optical field are used to enhance the signal and process noise for low signal-to-noise ratio small moving targets.Local Gaussian mixture models are also used for foreground-background segmentation of small moving targets,resulting in the trajectory of small moving targets.Next,a parallel-axis stereo camera is selected for long-distance self-calibration and hardware architecture construction.By analyzing the imaging principles,coordinate transformation matrix calculation principles,and binocular matching principles of the stereo camera,and combining it with a parallel light tube,electronic theodolite,and GPS rangefinder for long-distance self-calibration parameter verification,and analyzing the deviation of the ranging model,a high-precision optical bird detection stereo camera that can be used for longdistance and wide-field environments is obtained.Finally,the small moving target detection algorithm and stereo vision hardware are developed in conjunction to build an airport global visualization monitoring interface,a small moving target matching visualization interface,and a small moving target historical trajectory point output interface,which outputs a real-time monitoring system for small moving targets at the airport.The conclusion is as follows:(1)A micro-motion target detection algorithm was constructed based on statistical light field enhancement of micro-motion target feature signals and local Gaussian mixture model foreground-background segmentation.The algorithm exhibited strong environmental adaptability and stability in different lighting environments,and demonstrated high precision and robustness in detecting various types of micro-motion targets at long distances.(2)Using a self-calibrated binocular camera system composed of a parallel light tube,an electronic theodolite,and GPS distance measurement,the distance measurement error in the calibration experiment was maintained below 6%.This system can effectively meet the shooting needs of large field-of-view and long-distance airport environments,and provides good hardware support for airport monitoring systems.(3)The airport micro-motion target real-time monitoring system,which integrates the micro-motion target trajectory detection algorithm and the binocular camera hardware development,can detect micro-motion targets within a range of 600 meters with imaging pixel values greater than 1,with a distance detection accuracy error of less than 20%,and a real-time response monitoring rate output capability of more than 90%.The system has been normalized monitored at Guanghan Airport and Luoyang Airport,and can efficiently and stably output airport long-distance micro-motion target position information and overall monitoring information,effectively reducing airport management costs and improving airport safety monitoring capabilities.