摘要
为了能够快速而精确地实现对星图目标的质心提取,采用质心提取算法中具有较高精度的高斯曲面拟合算法,利用CORDIC算法来设计其中的对数运算模块,取代以往的查找表结构算法.通过ModelSim仿真,表明该运算模块能够以流水线的方式对数据进行连续处理,实时性较好.最后在FPGA硬件平台上进行验证,对一幅512×512像素的星图进行星点目标质心提取,试验结果显示该模块的计算误差在0.005个像素以内,定位精度要高于传统的质心法,而流水线的设计结构使其具有更快运算速度.
The Gauss Curve Fitting Methods was used here in order to get the star centroid quickly and accurately.The logarithm in Gauss Curve Fitting Methods was realized by CORDIC methods to replace the traditional way of LUT.A simulation was down by the ModelSim tools,it showed that this module can process the signal in the way of pipeline.At last we tested this module based on the FPGA device with a 512×512 star image.The results showed that its precision is higher than the traditional way of centroid locating with errors under 0.005 pixel and needs less time with the configuration of pipeline.
出处
《微电子学与计算机》
CSCD
北大核心
2012年第2期111-113,118,共4页
Microelectronics & Computer
