【摘要】 弹药跌落环境模拟系统主要用于对各种弹药(箱装或裸弹)、引信、仪器设备等进行跌落试验。针对野外复杂背景下高速相机采集的目标跌落图像噪声大、边缘复杂不易提取且图像数据量大等问题，设计了融合高斯核函数和边窗思想的滤波算法、基于感兴趣区域(region of interest, ROI)的改进Hough直线检测算法以及基于序列图像ROI动态更新的边缘直线自动提取算法，解决了去噪过程中目标边缘扩散和序列图像姿态角解算耗时长的问题。实验表明，所提出的改进算法保护了高噪声图像的边缘特征，直线检测耗时低于标准Hough变换耗时的1/10,实现了多帧图像运动目标姿态角的快速自动提取，姿态角解算平均绝对误差为0.36%。更多还原

【Abstract】 The ammunition drop environment simulation system is mainly used for drop test of various ammunition, fuses, equipment etc. Aiming at the problems of high noise, complex edges and large amount of image data in the target drop image collected by high-speed camera in the field with complex background, a filtering algorithm integrating Gaussian kernel function and side window idea, an improved Hough line detection algorithm based on ROI and an automatic edge line extraction algorithm based on dynamic ROI update of image sequence are designed. It solves the problem of target edge diffusion in the process of denoising and time-consuming attitude angle calculation of image sequence. The experimental results show that the improved algorithm protects the edge features of high noise images, and the line detection time is less than 1/10 of the standard Hough transform time. It realizes the fast automatic extraction of the attitude angle of moving objects in multiple images, and the average absolute error of attitude angle calculation is 0.36%.更多还原