粒子群算法在BRDF模型参数优化中的应用  

Application of particle swarm optimization algorithms to parameter optimization of BRDF model

作　　者：曹运华[1] 吴振森[1] 齐利华[2] 张耿[1] 张涵璐[1] 

机构地区：[1]西安电子科技大学理学院,陕西西安710071 [2]西安通信学院,陕西西安710106

出　　处：《电波科学学报》2008年第4期765-768,802,共5页Chinese Journal of Radio Science

基　　金：国家自然科学基金资助项目(60771038)

摘　　要：根据粗糙样片1.06微米激光双向反射分布函数(BRDF)的实验数据,利用粒子群算法结合BRDF五参数模型,对不同样片表面的BRDF进行了统计建模,获得了对应样片BRDF模型的参数值与偏差。获得的BRDF优化统计模型与没有参与优化建模的实验数据也吻合良好,验证了建模结果的合理性与正确性。将粒子群算法与遗传算法进行了比较,表明无论是优化计算时间还是拟合的精度,前者都明显优于后者。Based on the experimental data of the Bidirectional Reflectance Distribution Function （BRDF） of two target samples of different roughness illuminated by a laser beam of wavelength 1.06 microns, the five-parameter model of BRDF of the two samples is established, in which the corresponding optimized parameters are determined with the Particle Swarm Optimization （PSO） algorithm. The obtained BRDF model can fit both the experimental data used in the optimization procedure and those not, showing the reliability and validity of BRDF modeling. A comparison of the PSO algorithm and the Genetic Algorithm （GA） is also made in the modeling of BRDF, which shows that the PSO algorithm consumes less time and reaches better modeling accuracy.

关 键 词：双向反射分布函数 粒子群算法 遗传算法 粗糙样片 

分 类 号：O439[机械工程—光学工程]