CAJ | 학술논문

针对传统融合方法在光照不足、目标隐藏或目标和背景颜色接近时，容易出现目标信息丢失或减弱的现象，提出一种将基于CV模型的目标提取与NSCT相结合的方法。该方法使用动态轮廓线模型对红外目标进行搜索检测识别，将源图像序列分为目标和背景区域，利用非下采样Contourlet变换对输入图像进行多尺度、多方向稀疏分解，准确捕获图像中的高维奇异信息，并在目标和背景区域里分别采用不同的融合规则，将其与小波融合方法、拉普拉斯融合方法、NSCT方法作对比，并通过熵、平均梯度、空间频率、标准差等参数对融合后的图像进行定量分析。实验结果表明，该方法不但较好地提高了融合图像的目标探测性，而且融合结果中的目标比较清晰，亮度较高，目视效果较好，在主观视觉效果与客观评价指标上均取得了很好的融合效果。
침대전통융합방법재광조불족、목표은장혹목표화배경안색접근시，용역출현목표신식주실혹감약적현상，제출일충장기우CV모형적목표제취여NSCT상결합적방법。해방법사용동태륜곽선모형대홍외목표진행수색검측식별，장원도상서렬분위목표화배경구역，이용비하채양Contourlet변환대수입도상진행다척도、다방향희소분해，준학포획도상중적고유기이신식，병재목표화배경구역리분별채용불동적융합규칙，장기여소파융합방법、랍보랍사융합방법、NSCT방법작대비，병통과적、평균제도、공간빈솔、표준차등삼수대융합후적도상진행정량분석。실험결과표명，해방법불단교호지제고료융합도상적목표탐측성，이차융합결과중적목표비교청석，량도교고，목시효과교호，재주관시각효과여객관평개지표상균취득료흔호적융합효과。
In order to solve the problem that the target information is easily lost or impaired when the target is hided or the color of target is close to the background or lack of light, a new method that combined NSCT and target extraction based on CV model is proposed. The method uses the dynamic contour model to search, test and identify the infrared target, then divides the source image sequence into target and background region, and uses non-subsampled Contourlet transform to sparse decomposition in multi-scale and multi-directions for getting high-dimensional singular information accurately. Different fusion rules are used in target and background region respectively, and this method is compared with wavelet fusion method, Laplace fusion method, and NSCT fusion method. Quantitative analysis is carried out for the fused image under parameters like entropy, average gradi-ent, spatial frequency and standard deviation. The results show that this method can not only make the detection of fusion target more easily, but also make the target look clear and brighter. A good fusion effect in subjective visual andobjective evaluation index is obtained.