摘要
提出一种新方法,解决了红外图像中目标每个面的准确分解问题,通过分解假目标图像综合数据评价假目标红外的示假效果。利用目标可见光曲率图像边界信息的易获得性分解目标图像,获取目标可见光图像各面的边界坐标,确定对应的目标红外图像各面边界。使用仿射不变矩实现对三维多面体假目标红外示假效果的综合量化评价。车辆目标评价实验表明,该方法具有很好的有效性、可操作性和稳定性。
A new evaluating method is proposed. The method resolve the problem how to disassemble each plane of target exactly. By segmenting the target image, we can synthesizing the data for effect of decoy in infrared band. This method achieves the coordinate of each plane edge of target＇s visible image by making use of the easy-achieving character of boundary of curvature image. The each plane of the target＇s IR image could be confirmed by the coordinate. Then the affine invariant moments were used to evaluate the effect of decoy in IR band. This method was effective, exercisable and steady proved by the test of car-targets.
出处
《红外技术》
CSCD
北大核心
2010年第3期152-156,共5页
Infrared Technology
基金
国家自然科学基金资助项目(10372038)
总参科技创新资助项目
关键词
图像分割
曲率
仿射不变矩
效果评价
image segmentation, curvature, affineinvariantmoments, Evaluation
