机载雷达图像目标识别模型仿真研究

发布时间：2018-04-02 17:32

本文选题：机载雷达　切入点：塔机目标　出处：《计算机仿真》2014年12期

【摘要】：机载视觉雷达对目标准确标识别,关系到航空领域的安全。机载视觉雷达的目标识别多是高空作业,识别的目标图像信息易受到飞机倾斜角度、高空外部噪声干扰、机身异常抖动、被测物体的抖动和采样速度过低等因素的干扰,使得识别目标区域模糊,可识别特征发生严重衰减。传统机载视觉雷达的目标识别方法中,在运动状态下对高空目标的图像衰减特征分割一直很困难,分割过程会出现过分割和欠分割的问题,导致目标识别结果不理想。提出采用机载视觉雷达倾斜状态下的目标识别方法,得到图像中对地目标的运行速度,通过雷达视场距离的标定成像几何原理,将地面目标三维场景投射到二维象平面中,并采用数据链驱动无缝集成模式来运算识别地面目标的实际数量,获取准确的塔机目标识别检测结果。实验结果说明,所提机载视觉雷达倾斜状态下的目标识别模型获取的地面目标识别更加准确,并且具有较高的检测效率和精度。
[Abstract]:The accurate identification of target by airborne visual radar is related to the safety of aeronautical field.The target recognition of airborne visual radar is mostly aerial operation, and the target image information is easily disturbed by the factors such as aircraft tilt angle, high altitude external noise interference, fuselage abnormal jitter, jitter of measured object and low sampling speed, etc.It makes the target area fuzzy and the recognition feature attenuate seriously.In the traditional target recognition method of airborne vision radar, it is always difficult to segment the image attenuation feature of high-altitude target in moving state. The segmentation process will appear the problem of over-segmentation and under-segmentation, which leads to the unsatisfactory result of target recognition.In this paper, a method of target recognition in the tilt state of airborne visual radar is proposed to get the moving speed of the ground target in the image. The 3D scene of the ground target is projected into the two-dimensional image plane by the geometric principle of the calibration imaging of the radar field of view.The data link driven seamless integration mode is used to calculate the actual number of ground targets and obtain accurate detection results of tower crane targets.The experimental results show that the target recognition model obtained by the proposed airborne visual radar is more accurate and has higher detection efficiency and accuracy.
【作者单位】：华东交通大学信息工程学院;华东交通大学江西省先进控制与优化重点实验室;
【基金】：江西省科技计划项目(20122BBE500048) 国家自然科学基金(51174091,61164013) 江西省科技落地计划项目(赣财教[2011]243号) 江西省自然科学基金项目(20114BAB211017) 教育部人文社科基金项目(12YJCZH099)
【分类号】：TN957.52

【参考文献】