基于相邻帧补偿的高速运动目标图像稳像算法及仿真  

Algorithm and Simulation of Image Stabilization for High Speed Moving Target Images Based on Adjacent Frames Compensation

作　　者：姬莉霞[1] 李学相[1] 

机构地区：[1]郑州大学软件技术学院,郑州450002

出　　处：《计算机科学》2014年第7期310-312,317,共4页Computer Science

基　　金：河南省自然科学基金(132300410190)资助

摘　　要：在高速运动目标图像视频采集过程中,高速运动、风力作用等因素将导致视频图像抖动。为提高高速运动目标图像视觉系统采集性能,改善图像采集质量,提出一种基于相邻帧补偿的高速运动目标图像稳像算法。结合自适应中值滤波方法和灰度化直方图均衡方法对图像进行预处理,用尺度不变特征变换(SIFT)算法提取视频图像中的特征点,利用仿射模型求解运动参数,采用Kalman滤波对视频图像中的正常扫描进行滤波,最后用相邻帧补偿方法将图像的前一帧作为参考帧对当前帧进行参数补偿,实现高速运动目标的视频图像电子稳像处理。仿真实验表明,新算法能在保留图像中的特征的同时去除图像中含有的抖动,非常适合高速运动视频图像的电子稳像处理,精度提高,计算量明显减少。In the video image acquisition process for high speed moving object, the video and image shake due to high speed,wind and other factors. In order to improve the high speed moving target image acquisition performance of vision system, and improve the image quality, an improved algorithm of image stabilization for high speed moving target images based on adjacent frames compensation was proposed. We preprocessed the image by combining the gray histogram e- qualization method and adaptive filter, extracted the feature points in the video image through the scale invariant feature transform （SIFT） algorithm. And the affine model was used to solve the motion parameters. Kalman filter was used to scan the video image. Finally we used the adjacent frame compensation methods of a previous frame image as the refer- ence frame of parameter compensation to the current frame,and the electronic image stabilization processing was real- ized. Simulation results show that the new algorithm can retain the features in the image while removing images shake and jitter,and is very suitable for electronic stabilization processing for high speed video image. The precision is im- proved and the computing consumption is reduced significantly.

关 键 词：高速运动目标 图像 稳像算法 帧补偿 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]