基于单调场景运动的复杂非均匀性校正

Non-uniformity correction based on monotonous scene motion

在红外探测器中,高频非均匀性主要表现为列或者行条纹,而低频非均匀性则主要表现为片状。在实际应用中,场景是沿着一个单调方向,并且有不同的非均匀性出现。基于这样的考虑,提出了一个基于场景的非均匀性校正技术,不仅符合上述校正要求,而且也实现了小型化、低功耗的实时硬件系统。通过为配准原则开发一种新的投影估计来解决问题,同时在硬件系统上实现了这一技术。通过评价标准测试了这项技术的性能,展示了在单调运动下系统非均匀性校正的实际效果。实验结果表明:该算法具有校正精度高、收敛速度快的优点。

In the real IR detector, the high frequency non-uniformity is mostly expressed as the column or the row stripes and the low frequency non-uniformity is mostly expressed as the patchy pattern. In real application, the scene is towards one continuous monotonous direction and has different appearances of non-uniformity. Under this consideration, a scene-based non-uniformity correction technology was proposed, it not only fitted the situation mentioned above but also was realized on the small packaged, low power consumed real-time hardware system. These problems were solved by developing a new projection estimator for the registration with a criterion, and this algorithm was immigrated into a FPGA_based hardware system. This system is fully engineered for some particular usage. The performance of the proposed technology was tested by the evaluation indexes, and it demonstrates the actual effect of correcting the non-uniformity under a monotonous motion on the system. According to the practical application, this algorithm has the advantage of the strong ability to eliminate the image degradation and fast convergence speed.