【作者】 常青；

【Author】 Chang Qing East China University of Science and Technology,Shanghai 200237

【机构】 华东理工大学信息科学与工程学院；

【摘要】 红外目标图像的特殊性导致单单依赖基于灰度统计信息的特征提取方法区分背景和目标很难奏效。本文使用分形方法对复杂背景目标识别算法进行研究。针对传统方法计算量大、运算时间长的特点,采用基于灰度剖面降维的形态学分形维数计算方法,减少了计算量。针对传统分形方法某些情况下,由于滑窗选取不当导致分形维数计算偏差较大的局限性,采用计算多方向分形维数的方法,较为有效地标记出被检测目标。实验结果表明了本算法的有效性。更多还原

【Abstract】 The IR image from the complex scene is not as simple as the point-source target image. So the traditional feature extraction method that based on gray statistical to distinguish the target and the scene is hard to make effect. The fractal geometry theory presents the new direction. To the limitation of the traditional algorithm that has lower calculation speed, a descend-dimension algorithm is put forward to meet the request. The result of experiments has demonstrated effectiveness of this method. A new fractal-based infrared image feature extraction method is presented and the detailed theoretic analysis and implement procedure of this algorithm is submitted and the method is tested in some experiments too.更多还原