红外图像序列中运动弱小目标检测的方法研究

研究在红外图像序列中检测运动弱小目标的方法。首先通过高通或形态学滤波进行图像预处理，进一步用似然比检测分割出侯选目标，考虑到环境干扰造成的目标在某一帧暂时消失的情况，提出了利用目标运动特征通过选择合适的邻域判决条件并结合图像流分析提取运动弱小目标的一种方法。模拟实验表明该方法能够准确高效地检测出运动弱小目标。

Small Moving Target Detection in Infrared Image Sequences

An approach to detecting small moving targets in image sequences is described in this paper. First pretreat image through high pass or morphologic filtering, then progress likelihood ratio test to segment possible targets from the image. Considering the case that target disappear transitorily because of the environmental disturbance, a method of extracting small moving targets is brought forward, which uses the continuity and correlation of motion state of the target and through synthetically operating neighboring region verdict and image flow analysis to detect small moving targets. Simulation result indicates that the algorithm can detect small moving targets with high efficiency.