适用于光电跟踪的初始目标区域估计法  

Initial target region estimation suitable to optical tracking

作　　者：刘丹平[1] 段钦义[1] 李云伟[1] 刘晓明[1] 李正周[1] 

机构地区：[1]重庆大学通信工程学院,重庆400030

出　　处：《光学精密工程》2008年第11期2286-2289,共4页Optics and Precision Engineering

基　　金：重庆市自然科学基金资助项目(No.2006BB3162)

摘　　要：为了提高光电跟踪度的定位精度，提出了用于初始目标区域估计的宏像素方法。定义了有效像素比r的概念，利用它分析了噪声和非对称性对目标初始定位精度的影响，得到了高精度强抗噪声能力的最大宏像素和确定宏像素目标的限制操作。对目标占4pixel×4pixel左右，大小为128pixel×128pixel的非对称光斑图像，选择宏像素大小为36pixel，目标宏像素扩大2pixel区域。在SNR为5dB时，r值达到1。动态图像模拟表明，该方法处理速度可达18～20frame／s。In order to improve the accuracy of optical tracker, the macro pixel method is used for estimating initial target region. A term,valid pixel ratio, is put forward and then is applied to analyze the effects of noise and asymmetry on the location for initial target region. Finally, the most macro pixels and the restricting operation way with high-precision and strong anti-noise ability are obtained. For the unsymmetry target of 4 pixel × 4 pixel in 128 pixel× 128 pixel spot image, the most macro pixel is 36 pixels and target macro pixel is enlarged by 2 pixel. The value of r is 1 when SNR is 5 dB. Dynamic picture test shows that the method has the processing speed of 18-20 frame/s.

关 键 词：光电跟踪 质心算法 宏像素 宏像素划分算法 

分 类 号：TN386.5[电子电信—物理电子学] TP391[自动化与计算机技术—计算机应用技术]