摘要
红外成像系统由于探测器加工工艺的限制,很难通过减小像元尺寸或增加阵元数量的方式实现高分辨率成像。压缩感知理论提供了一种新的提高成像分辨率的方法,在光学系统的焦平面处放置编码掩膜,使得红外探测器得到的图像是被观测场景的压缩采样,再通过稀疏优化算法重构出原始图像。决定图像分辨率的不是探测器的像素尺寸,而是编码掩膜的孔径大小。在此框架下,设计了合适的光学编码掩膜子阵,利用多路技术实现了对同一场景的多次压缩采样,采用了线性Bregman迭代思想进行重构算法的设计,解决了二维成像大规模重构算法的求解速度和精度问题。数值仿真表明,该方法在保证图像重构质量的前提下可显著提高红外成像的分辨率。
For the restriction of infrared sensor＇s processing technology, it is difficult to realize high resolution imaging by reducing pixel-pitch or increasing the number of array elements. Compressed Sensing （CS） provides a new approach for improving imaging resolution. By designing a coded aperture mask on the focal plane in the optical system, the infrared image sensor obtains the compressed samples of the observed scene, and then the image reconstruction can be conducted using an optimal algorithm from these samples. The image resolution was determined through the size of the sub-array on the coded aperture mask other than the pixel-pitch. Based on the framework of CS, a suitable optical coded aperture mask was designed, and the same scene could be sampled repeatedly during a single exposure by using multiplexing technology. Linear Bregman iteration was employed as the reconstruction algorithm, which ensures the accuracy and efficiency of large-scale two-dimensional image reconstruction. Numerical simulation shows that the proposed method is feasible. It improves infrared imaging resolution significantly and ensures the quality of recovered images.
出处
《红外与激光工程》
EI
CSCD
北大核心
2011年第11期2065-2070,共6页
Infrared and Laser Engineering
基金
武器装备研究项目
