基于特殊RGB排列的彩色LED阵列计算鬼成像

Computational Ghost Imaging Based on Chromatic LED Array with Special RGB Arrangement

图像的低质量--直是制约计算鬼成像(CGI)技术应用的重要问题,因此,针对CGI系统的彩色发光二极管(LED)阵列设计了-种特殊的RGB(Red,Green,Blue)排列方式。首先,自主开发了一个64×64的彩色LED阵列为CGI系统提供结构照明;然后,将LED阵列上的芯片按设计的篮织排列方式进行排列;最后,考虑到彩色LED阵列的物理结构,提出了一种针对采样图像的插值算法。实验结果表明,相比笛卡儿采样方式,篮织采样方式的均方根误差(RMSE)降低了4.6%,在存储彩色图像高频信息方面的性能更好。相比双线性插值算法和双三次插值算法,本算法的平均RMSE分别降低了2.0%和6.4%。

Low quality of images is a critical problem that inhibits the application of computational ghost imaging(CGI)technology.Therefore,a special RGB(Red,Green,Blue)arrangement is designed for the chromatic light-emitting diode(LED)array of the CGI system in this study.First,a 64×64 chromatic LED array is independently developed to provide structural lighting for the CGI system.Then,the chips on the LED array are arranged according to a designed basket-weave arrangement.Finaly,considering the physical structure of the chromatic LED array,an interpolation aIgorithm is proposed for sampled images.Experiment results show that the root mean square error(RMSE)of the basket-weave sampling is reduced by 4.6%,compared with Cartesian sampling and it has better performance in storing high-frequency 1nformation of chromatic images.In addition,compared with the bilinear and bicubic interpolation algorithms,the average RMSE of the algorithm is reduced by 2.0%and 6.4%,respectively.