基于双特征融合的动态图像分析算法

为解决汽车碰撞实验过程进行测量和记录数据困难的问题，提出一种能从汽车碰撞动态图像中检测、识别和追踪标志目标的算法。该算法采用二值图像同或相关法分离出目标和背景；在找出感兴趣区域后提取相邻帧目标的坐标和纹理进行特征匹配；给出了对匹配量化值进行加权平均融合的策略，并由等错误率最小准则确定融合的最佳权系数；在融合量化值定义为相似度的基础上，通过决策阈值对相邻帧目标间的最大相似度组合进行识别；从而进一步提出了用同构映射原则来判断相邻帧目标的最佳配对。实验结果表明，该算法对相邻帧目标的配对准确率比传统单特征法提高5%，能更有效的对目标进行追踪。

Analysis Algorithm of Dynamic Image Based on Double Feature

In order to solve the difficult problem in car crash experiment to measure and record data,a dynamic image from a car crash detection,target identification and tracking algorithm is proposed.The algorithm uses binary images related to the target and background separation in the region of interest identified targets adjacent frames after the extraction of texture coordinates and feature matching;gives quantitative values of the matching strategy of integration of the weighted average by the criterion of mini...