The majority of the traditional camouflage target detection methods rely on the human eye to make a simple magnification observation using optical instruments, but the detection result obtained is usually inferior. The hyperspectral instrument can acquire spectral images of a certain band range and then obtain a continuous spectral curve of the object. In general, the spectral curves of different objects are different. The camouflage effect of targets in different backgrounds is analyzed using hyperspectral images obtained in the visible and near-infrared bands. The difference between the camouflage clothing and the background spectral curve is evaluated based on the spectral amplitude and variation tendency of the spectral shape. From the comprehensive evaluation of the experiment, it is observed that camouflage clothing in the real grass camouflage effect is better than plastic fake turf. It meets the requirements and provides good camouflage for military action. However, the experiment also shows that the variation trend of spectral curves of different objects is different in the visible light band. The spectral reflectance of the objects varies greatly in the near-infrared band, but the spectral change rate is small. In camouflage target detection, if we only perform the evaluation from one aspect, we may obtain erroneous results. Therefore, in camouflage detection, it is necessary to comprehensively consider the detection target from various viewpoints.%传统的伪装目标探测识别大部分是依靠人眼通过光学仪器进行简单的放大观测,但是通常探测效果不佳.高光谱仪能采集一定波长范围的光谱图像,从而得到物体连续的光谱曲线,一般来说,不同物体间的光谱曲线存在差异.利用在可见光及近红外波段下获取的高光谱图像,分析伪装目标在不同背景下的伪装效果.通过光谱幅值及光谱形状变化趋势,评价伪装目标与背景光谱曲线的差异.根据实验结果综合评价,伪装目标在真实草地中的伪装效果相对塑料假草皮来说更好,符合实际需要,能为军事行动提供良好的伪装.同时,实验也表明,在可见光波段下,不同物体光谱曲线变化趋势不同,在近红外波段下,物体间的光谱反射灰度值差异较大,但是光谱变化率差异较小.在伪装目标探测时,若仅从某一方面来评价,可能得到错误的结果.所以在探测识别时,需要从各方面综合考虑.
展开▼
