摘要
20世纪80年代发展起来的高光谱成像技术能充分探测并利用地物的光谱信息,有着较强的伪装识别能力,而传统的伪装评价方法主要基于伪装和背景的空间特征,不适用于基于光谱特征的高光谱伪装探测技术。提出了一种面向高光谱探测的伪装效果评价方法。该方法以伪装目标和背景的光谱特征及光谱导数特征的综合相似度为评价指标,综合相似度值越大则说明伪装效果越好。同时通过实验评价了三种伪装材料的伪装效果,其中绿色棉质布料与背景综合相似度为0.976 6,伪装效果最佳。最后在高光谱图像分类技术基础上提出了错分率的概念,验证了评价方法的客观性。所提出的伪装评价方法对于面向高光谱探测的新型伪装材料的设计有一定的指导意义。
Camouflage is an important factor for winning the modern warfare by preserving fighting capacity. Hyperspectral imaging technology is an efficient method for camouflage detection for its ability of detecting and utilizing targets′ spectral information. Traditional camouflage effect evaluation method mainly based on spatial characteristics, which cannot work when hyperspectral imaging technology is used. So a new camouflage effect evaluation method was proposed for hyperspectral detection, which was based on the comprehensive similarity of camouflage and background combined characters of spectrum and its derivative. Target with higher comprehensive similarity meant better camouflage effect. Experiment was conducted to assess the camouflage effects of three camouflage materials by the new method. The camouflage effect of green cloth outperforms others with the comprehensive similarity 0.976 6. And error rate based on hyperspectral classification technology was proposed to verify the objectivity of the new method. The new camouflage effect evaluation method has a certain guiding significance for the design of the new camouflage material.
出处
《红外与激光工程》
EI
CSCD
北大核心
2014年第10期3228-3232,共5页
Infrared and Laser Engineering
基金
国家自然科学基金(10376044)
国家863计划
关键词
伪装效果评价
高光谱成像
光谱特征
高光谱分类
错分率
camouflage effect evaluation hyperspectral imaging spectral characters hyperspectral classification error rate