FPGA中实现Bayer模板双线性插值算法的改进

针对单板CMOS图像传感器采集的Bayer模板的数据,结合FPGA硬件结构FIFO_DFF构成3×3插值模板,提出一种改进的双线性插值算法。插值过程中利用当前要还原的分量同通道的水平、垂直以及两个对角线方向上的已知分量进行边缘检测,用梯度小的方向的分量进行插值还原缺失的分量。实验结果表明,在FPGA中使用该算法插值得到的图像色彩准确,增强了图像的边缘和细节,图像清晰度高。

Improvement of bilinear interpolation algorithm for Bayer pattern in FPGA

Using a 3×3 interpolation mold that is constructed by hardware resource of FIFO_DFF within FPGA to improve upon bilinear interpolation algorithm for image data from single CMOS sensor in Bayer pattern.In the interpolation process,horizontal,vertical and diagonal directions of the known color components in same channel are used to conduct edge detection,and the components in a small gradient direction are taken to compute the missing color components.The experimental results indicate that improved algorithm in FPGA is able to offer an image with accurate color and enhance the image edges,details,and has high definition as well.