基于两点参数及自适应窗口的实时盲元检测和补偿技术

Real-time Blind-pixel Detection and Compensation Technology Based on Two-point Parameters and Self-adaptive Window

在红外焦平面(IRFPA)成像系统中,盲元检测和补偿是红外成像过程中必不可少的一步。针对FPGA,提出一种基于两点参数的简化型3盲元检测算法,并介绍了自适应窗口盲元补偿算法。该算法通过分析两点参数矩阵的分布情况,以3为基准裁定盲元;再在自适应窗口中补偿盲元,优化了补偿结果。具体介绍了硬件功能结构和算法关键数据流的设计。仿真和实际场景实验结果表明该算法对盲元能做到精确检测和有效补偿,且速度快和资源省,能在硬件上完整地实现。

For an infrared imaging system, blind-pixel detection and compensation is an inevitable step in the infrared imaging process. A new method of blind-pixel detection which is based on two-point parameters and 3σ theory and self-adaptive window compensation method is introduced. The algorithm detects blind-pixels through the distribution of two-point matrixes and 3σ theory; The blind-pixel compensation is operated in the self-adaptive window and the compensation value is optimized. The design of hardware function structure and the key data stream of this algorithm are introduced. The results of simulation and real scene experiments prove that this method can detect and compensate blind-pixels accurately, effectively, fast and saving resources, and it can be transplanted to FPGA easily.