论文部分内容阅读
红外焦平面器件的非均匀性产生机理复杂,难以准确拟合探测元响应曲线。提出了一种基于相关干扰抵消的非均匀性校正算法,以预先采集到的一帧黑体面源图像做为自适应干扰对消器的参考输入图像,自适应滤波器由参考输入图像迭代计算出待校正红外图像的空间噪声的最佳估计,实现从空间噪声中提取真实图像信号。自适应滤波算法采用变步长最小均方误差算法,减少了算法的运算量,提高了算法的收敛速度。理论分析以及针对实际红外图像的仿真结果表明,提出的算法校正效果好,收敛速度快,更易于工程实现。
The mechanism of the inhomogeneity of the infrared focal plane device is complicated and it is difficult to accurately fit the detector response curve. A non-uniformity correction algorithm based on correlated interference cancellation is proposed. A pre-captured image of a blackbody surface is used as a reference input image of an adaptive interference canceller. The adaptive filter is iteratively calculated from the reference input image To be corrected the best estimate of the spatial noise of the infrared image, to extract the real image signal from the spatial noise. The adaptive filtering algorithm uses a variable step size minimum mean square error algorithm, which reduces the computational complexity of the algorithm and improves the convergence speed of the algorithm. Theoretical analysis and simulation results for the actual infrared images show that the proposed algorithm has good correction effect and fast convergence speed, and can be easily realized by engineering.