基于多尺度分形参数变化的目标检测方法研究  

Target Detection Based on Multi scale Fractal Parameter Change

作　　者：李捷[1] 张天序[1] 

机构地区：[1]华中理工大学图象识别与人工智能研究所

出　　处：《数据采集与处理》1996年第3期218-221,共4页Journal of Data Acquisition and Processing

基　　金：国家自然科学基金;国防科工委基金

摘　　要：分析和研究了提取图象多尺度分形参数的地毯覆盖方法，并作了相应改进。阐明了Ｄ维面积的意义。根据自然场景和人造目标随尺度的不同变化性质，提出了分形参数变化度量函数。实验结果表明。The covering blanket method used to estimate multiscale fractal parameters is studied and its algorithm is also improved. The meaning of D dimensional area which hasn′t been explained in detail in some other references is clarified. Based on analyzing the discrepancies whose different objects show on fractal feature, the fractal parameter change measure function is presented. It may be well performed in detecting targets from natural scenes due to using the measure function mentioned above.

关 键 词：图象处理 目标探测 分形参数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]