Abstract: In this paper，a method by intraframe processing and interframe processing for automatic target capture in FLIR image sequences is proposed.Firstly，the paper analyzes filter by reconstruction and reveals property of background reconstruction in FLIR image on complete lattice.Then，utilizing advantages of both linear wavelet image denoising and morphology wavelet keeping texture of the image，an adaptively morphological wavelet algorithm based on adaptive lifting schemes can farthest suppress image white gaussian noise and keep target contours as well.Lastly，beside adaptive thresholding and criteria processes，these can implement intraframe processing.Interframe processing is implemented by dilation-based addition along the time direction.The experiments perform on three FLIR image sequences，the results show that the method can capture dim target rapidly and accurately.

Key words: infrared image, filter by reconstruction, morphology wavelet, automatic capture

摘要： 提出一种新的基于形态重构与自适应小波算法的自动捕获红外弱小目标的方法，处理方法分为帧间与帧内处理两部分。首先论证了形态重构滤波器在完备格理论下可以实现对红外图像背景的重构，其次分析了自适应小波结合了线性小波强抑制噪声特性和形态小波好的保留目标细节特性，在红外图像能够获得更好的去噪性能，最后通过结合自适应阈值处理方法和准则处理实现了帧内处理，利用时间序列上的形态膨胀累加完成帧间处理。通过对三种不同的红外图像序列的仿真处理，实验结果表明自动捕获方法可以快速准确地捕获图像的弱小目标。

关键词: 红外图像, 重构, 形态小波, 自动捕获