基于FPGA的红外图像非均匀校正实现方法

发布时间：2018-05-26 16:17

本文选题：非均匀性校正 + 神经网络算法　；参考：《激光与红外》2016年08期

【摘要】：针对某国产探测器成像特点,对传统两点校正和神经网络非均匀性校正算法进行了改进和定点化处理。对算法实现时的存储和数据流需求进行分析后,利用存储控制器对DDR2高速读写的优势,在以FPGA为核心的红外成像装置预处理平台上实现了校正系数的在线标定和自适应迭代。在系数更新时,引入运动判断环节,以防止神经网络校正算法带来的目标退化和鬼影现象。成像系统仅采用一片FPGA芯片,使得系统小型化成为可能,充足的资源余量使其具有功能可扩充性。实验证明该实现方法明显改善了红外成像装置的非均匀性,在抑制时间漂移上也取得了满意的效果。
[Abstract]:In view of the imaging characteristics of a domestic detector, the traditional two points correction and neural network non-uniformity correction algorithm are improved and fixed point processing. After analyzing the storage and data flow requirements of the algorithm, the advantages of the memory controller for DDR2 high-speed reading and writing are made, and the pre processing platform of the infrared imaging device with FPGA as the core is used. The on-line calibration and adaptive iteration of the correction coefficient are realized. When the coefficient is updated, the motion judgment link is introduced to prevent the target degradation and ghost phenomenon caused by the neural network correction algorithm. The imaging system only uses a chip of FPGA, making the system miniaturized possible, the full resource remainder makes it functional extendibility. It is verified that the implementation method significantly improves the non-uniformity of the infrared imaging device and achieves satisfactory results in restraining time drift.
【作者单位】：中国空空导弹研究院;
【分类号】：TP391.41;TN219

【相似文献】