红外双波段图像实时融合系统

Infrared dual-band real-time image fusion system

选用高性能定点数字信号处理器DM642作融合系统核心处理器,S3C2410作辅助控制处理器,利用DM642的高速图像处理性能及它的视频编/解码单元无缝连接的灵活可配置视频端口,结合S3C2410强大的控制能力,研制了探测中波红外和长波红外的双波段图像实时融合系统。将一种基于离散小波变换的图像融合算法应用于红外双波段图像的融合,针对实时嵌入式系统的特点,对算法进行了优化,并将其移植在以DSP+ARM架构的嵌入式平台上。实验结果表明,该融合算法经仿真优化后,只需39.6ms即可完成两幅大小为320pixel×240pixel的红外图像的实时融合,满足25frame/s的工程需求,且融合后的图像能突出中波和长波红外各自的特点。

By taking a high performance fixed-point DSP DM642 as a kernel processor,and a S3C2410 as a assistant control processor,an IR dual-band real-time image fusion system worked at Mid-wave Infrared （MWIR） and Long-wave Infrared （LWIR） is designed by using the high speed operation of DM642 and its configurable video ports combined with the powerful control capability of the S3C2410.An image fusion algorithm based on Discrete Wavelet Transform （DWT） is applied to the IR dual-band image fusion.Then,the algorithm is optimized to run on a DSP＋ARM embedded hardware platform based on the features of real-time embedded system.The experiments demonstrate that it is only 39.6 ms to be needed to complete the real-time fusion of two IR images with the resolution of 320 pixel×240 pixel,which meets the engineering requirements of 25 frame/s.It also demonstrates that obtained images can show the characteristics of MWIR and LWIR,respectively.