条纹非均匀性是线扫红外焦平面阵列和非制冷凝视型红外焦平面阵列成像系统中一种特殊的固定图案噪声。提出了一种基于实际场景和非均匀性分离的模型方法,能够在单帧内实现条纹非均匀性的校正。通过局部窗口模版遍历相邻列的误差函数,得到其非均匀性校正参数估计,并依据相邻两帧相关性的继承性,完成条纹非均匀性的校正。通过对模拟条纹非均匀性和实际图像的实验和理论分析,结果表明,文中算法能够显著提高条纹非均匀性校正效果。
Stripe nonuniformity is a very typical fixed pattern noise in line infrared focal plane arrays (IR-FPA) and uncooled staring IR-FPA. In this paper a model was put forward based on the separation between stripe nonuniformity and real scene, and the correction could be achieved in a single frame. The estimated nonuniformity correction parameter was obtained by traversing the error function of adjacent column pixels in local template window. Through the succession of two adjacent columns" correlation, the stripe nonuniformity correction was achieved. By analyzing the experiment results of the simulated imagery and the real infrared image sequence, it is obvious that the proposed method can significantly enhance the performance of stripe nonuniformity correction.