利用特征加权进行基于小波框架变换的红外目标检测

提出了一个基于背景纹理多分辨分析的红外目标检测方法。本方法通过对原始图像作小波框架变换提取不同分辨率下背景纹理的特征，利用特征加权技术（FWT），得到中心向量，求出每个象素点的特征向量到中心向量的距离。根据距离像统计直方图实现阈值判别。实验证明该方法取得了较好的效果。

Target Detection in Infrared Image Based on Discrete Wavelet Frames Applying Feature Weight

This paper presents a new target detection approach based on multi-resolution background texture analysis, which transforms an original image by discrete wavelet frames, and then extracts different background texture feature. Applying feature weighted technology, we can obtain the center vector. A distance image can be produced after computing distance from feature vector to center vector. In the end, target detection is realized by setting a threshold. Experiments show this method is efficient.