探讨了红外成像系统实时信号处理算法、硬件结构等关键技术。提出了一种新的自适应两点非均匀性校正(ATPC)算法,该算法以快门为目标场景来实时更新校正参数,能有效克服红外焦平面阵列(IRFPA)响应随时间漂移导致的两点非均匀性校正算法失效问题。采用查找表结构来实现平台直方图均衡(PE)算法,并对PE算法进行了优化,降低了运算量和存储空间需求。硬件系统采用两片同步动态随机存储器(SDRAM)的乒乓缓存结构,在单片现场可编程门阵列(FPGA)上实现了IRFPA的ATPC和PE算法的并行操作。实验结果表明,对320×240的IRFPA,在50MHz系统时钟下,帧频为60Hz时系统工作良好,处理后的红外图像质量有了明显改善,系统结构简单便于小型化,能够满足实时动态检测及追踪需求。
Key technologies such as real-time signal processing algorithms and hardware structure intended for driving a 320 × 240 infrared focal plane arrays (IRFPA) are presented here. Owing to the response of the IRFPA is instable with time, an adaptive two point correction (ATPC) algorithm was introduced in which the coefficients were updated accoding to a shutter image as the target. In addition, the real-time processing for plateau equalization (PE) is difficult although the infrared image can be enhanced effectively. In this article, PE algorithm was optimized and simplified by a loop-up table in which the size was limited by the maximum and minimum values in the frame. Therefore, the calculating numbers and requirement of memory were reduced greatly. A real-time infrared system structure based on field programmable gate array (FPGA) was introduced, in which two synchronous random access memortes (SDRAM) were used as frame buffer. In this article, the structure of the system and the software realization of the ATPC and PE were described in details. Meanwhile, the infrared images before and after the process of ATPC method were shown here as well as images with different Plateau values. These results suggested that, for a 320 × 240 IRFPA, the system works well in the 60 Hz frame rate and 50 MHz system clock. Furthermore, the infrared images were improved significantly and the requirement for real signal processing was achieved consequently.