项目名称: 基于图像光流的全天候机场防侵入关键技术研究
项目编号: No.U1233125
项目类型: 联合基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 陈震
作者单位: 南昌航空大学
项目金额: 38万元
中文摘要: 机场侵入是当前全球民航业亟须面对的重大安全问题,是民航安全工作的一项重要内容。 本项目主要研究基于图像光流的全天候机场防侵入关键技术,包括:研究多传感器外部标定方法,使用三维立方体作标定靶,利用标定靶上点与直线的对应关系,建立约束方程,解决多传感器的快速外部标定问题;研究可见光图像和红外图像的融合算法,采用改进非采样Contourlet变换的图像融合方法,解决准确的多传感器图像融合问题;研究基于融合图像光流的目标检测和分割方法,通过设计光流算法、区域分析和提取边界得到目标物体上的点云,检测和分割出机场跑道上的运动目标物体;搭建全天候的机场防侵入多传感器视觉系统,给出一种具有自主知识产权的全天候机场防侵入解决方案。 本项目的研究将推动我国机场防侵入技术的发展,提高机场管理的智能化水平,减少由机场侵入带来的经济损失,从而对科学技术发展和社会经济发展起到推动作用。
中文关键词: 外部标定;图像配准;图像融合;光流;目标检测
英文摘要: The airport invasion is a major security problem in the global civil aviation, which is also an important content of the safety work in the civil aviation. The main research of this project lies in the key technologies of the all-weather airport anti-invasion based on optical flow of the multi-sensor images, which includes: research on the extrinsic calibration method of multi-sensors, settlement on the fast extrinsic calibration problem with the constraint equations which obtained by using the 3D cube to be the calibration target; Study on the image fusion of the visible image and the infrared image, using improved non-sampled Contourlet transformation to solve the the accurately multi-sensor image fusion problem; Research on the target detection and segmentation based on the fusion image optical flow, to get all the point cloud on the objection by designing the optical flow strategy, region analysis and boundary extraction method; Construct the multi-sensor visual system for the all-weather airport anti-invasion, and prensent a independent intellectual property right of the all-weather airport anti-invasion technology. This project would be helpful to promote the development of the native anti- invasion technology and the intelligent management level of the airport, and reduce the economic losses c
英文关键词: extrinsic calibration;image registration;image fusion;optical flow;target detection