CAJ | 학술논문

通过研究可见光视觉图像与红外图像序列的相关特性及融合处理中存在的问题, 提出一种基于视觉图像预增强处理的曲波与小波双变换融合机制来有效保留边缘信息和图像细节信息。实验结果表明所提出的方法提高了融合图像的视觉感知质量, 为遮挡和伪装目标的检测和定位提供了更为有效的空间环境, 并且融合增强后的图像具有更高的信噪比和更少的人工痕迹, 与当前最新的其他方法相比是一种具有更好融合质量和更高效率的图像/视频融合方法。
통과연구가견광시각도상여홍외도상서렬적상관특성급융합처리중존재적문제, 제출일충기우시각도상예증강처리적곡파여소파쌍변환융합궤제래유효보류변연신식화도상세절신식。실험결과표명소제출적방법제고료융합도상적시각감지질량, 위차당화위장목표적검측화정위제공료경위유효적공간배경, 병차융합증강후적도상구유경고적신조비화경소적인공흔적, 여당전최신적기타방법상비시일충구유경호융합질량화경고효솔적도상/시빈융합방법。
This paper proposed a novel method by studying the characteristics of visual, infrared image/video and the problems introduced from the fusion processing. The proposed method enhanced the illumination of the visual image firstly, and then used curvelet and wavelet dual-transform to save the edge information and preserve the details of the visual and the infrared images during the fusion enhancement. Experimental results show that the proposed method enhances human perception for image/video and improves camouflaged target's detection and location in complex environment. It is proved to be a high quality and high efficiency image/video fusion method with the characters of higher PSNR and less artifact compared with the other state-of-the-art methods.