基于邻近插值法还原BayerRGB的FPGA实现  

Implementation of restoring Bayer RGB on FPGA through Nearest Neighbor Interpolation

作　　者：王健[1] 应骏[1] 曾维军[1] 曾爽[1] 

机构地区：[1]上海师范大学信息与机电工程学院,上海200234

出　　处：《上海师范大学学报（自然科学版）》2016年第4期411-416,共6页Journal of Shanghai Normal University(Natural Sciences)

摘　　要：图像传感器在采集每一个像素时仅采集它的R、G、B其中一种颜色分量的强度信息.在彩色数字图像系统中,需对传感器图像进行实时处理和颜色插值.邻近插值算法的时间复杂度和空间复杂度都较小,可以通过流处理实现,这使得它容易在FPGA上实现且节省硬件资源.在进行硬件实现之前首先通过MATLAB对邻近插值法的还原效果进行验证.硬件设计中使用移位寄存器连接FIFO的架构,并使用标准Verilog代码编写实现,最终用Modelsim仿真验证设计.Image sensors only sample the intensity information of one color component of RGB when each pixel is collected. In color digital image system, it needs real-time processing and color interpolation for the sensor image. Both time complexity and space complexity of Nearest Neighbor Interpolation （ NNI） are smaller and can be achieved by stream processing, which makes it easy to be implemented in the FPGA hardware and is hardware resourcesaving. First,the NNFs restoration effect of neighbor interpolation method is verified by MATLAB before making hardware implementation. In the hardware design, shift register is used to connect FIFO＇ framework, standard Verilog code is used for realization and finally Modelsim is used to simulate and verify design.

关 键 词：邻近插值法 流处理 BayerRGB FPGA 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]