摘要
研究飞行器红外图像复原优化校正问题,为了解决气动退化图像复原处理中的病态问题,提出了一种气动退化图像自适应空域正则化方法。首先,构建了包括图像模糊程度恢复正则项、边缘保护正则项和噪声平滑正则项的多重约束的空域正则化模型,然后,根据退化图像局部特征使用先验策略选取了局部正则化参数,并对正则项设置了步长函数以及非负等多重约束;最后,结合Steffensen加速迭代法和牛顿法实现了最小泛函的加速求解。多帧气动退化图像复原实验结果表明,复原后的图像质量得到极大程度的改善,纹理细节得到有效恢复,满足了气动退化图像复原处理的需要。
To solve the ill-posed problem in the restoration of aero-optics degraded image, an adaptive spatial regularization restoration algorithm of aero-optics degraded image was put forward. Firstly, spatial regularization mod- el consisting of regular terms of recovery of blur, regular terms of edge preservation constraint and regular terms of noise smoothing constraint was put forward. Then, the transcendental strategy was used to choose local regular param- eter according to local variance of degraded image. The function of step and multiple constraints were used to restrict regular terms. Finally, minimax function was solved by combining steffensen accelerated iterative method with Newton method. Large numbers of experiment results show that image quality of restoration is improved and texture detail is resumed in a great extent, all above satisfies the need of restoration disposal of aero-optics degraded image.
出处
《计算机仿真》
CSCD
北大核心
2013年第11期218-223,共6页
Computer Simulation
基金
国家自然科学基金项目(61175120)
关键词
气动退化图像
空域正则化
正则项
正则化参数
加速迭代法
牛顿法
Aero-degraded image
Spatial domain regularization
Regular term
Regular parameter
Accelerated it-erative method
Newton method
