摘要
为了提高雷达图像的融合质量,创新性地将非下采样Contourlet变换(NSCT)与脉冲耦合神经网络(PCNN)相结合,运用到可见光和红外雷达图像的融合中。先对待融合的两幅源图像进行NSCT分解,利用得到的低频子带系数去触发PCNN的神经元,最后进行NSCT重构,得到所需要的新图像。结果表明此方法较传统的融合方法,提高了信息量和清晰度,获得了较好的识别率。此方法得到的图像更有利于对流云形成时的预测。
In order to improve the quality of the radar image fusion, the method to combine Nonsubsample Contourlet Translation （NSCT） with Pulse Coupled Neural Network （PCNN） for visible and infrared radar image fusion is innovationally adopted. The process of this method is.. the two source images under fusion are decomposed by NSCT, then the low-frequency subband coefficient is utilized to trigger on PCNN neurons, and finally the image is reconstructed with NSCT to obtain a new required image. The results indicate that, compared with traditional fusion methods, this method has higher information ca- pacity, clarity and identification rate. The conclution is that the image got by this method is more advantageous to predict when the convective cloud is generated.
出处
《现代电子技术》
2012年第12期82-83,86,共3页
Modern Electronics Technique
基金
中国气象科学院灾害天气国家重点实验室开放课题(2008LASW-A02)
