基于相机阵列的红外光场成像

发布时间：2018-06-19 03:37

本文选题：成像系统 + 红外图像　；参考：《光学学报》2017年09期

【摘要】：光场成像作为一种应用广泛的计算成像方法,其研究仍局限于电磁波谱中的可见光波段。因此,提出基于相机阵列的红外光场成像方法。首先利用单台被动红外相机移动到同一平面的不同位置分别对同一静态目标场景成像,获取原始红外图像序列,经平移视差校正后生成红外光场数据库;然后基于光场渲染理论,获得重采样图像。与只记录二维位置信息的传统红外成像比较,红外光场成像能够记录包含位置和方向的四维信息。实验结果表明,该法能够融合多幅低信噪比的图像获得一幅信噪比提高的重采样图像,通过孔径叠加平均可实现穿透遮挡物成像,通过选取不同深度值进行重采样可实现红外场景不同深度的数字重聚焦,使红外成像在军事和民用领域更有应用价值。
[Abstract]:As a widely used computational imaging method, optical field imaging is still limited to the visible band of the electromagnetic spectrum. Therefore, the infrared field imaging method based on the camera array is proposed. First, a single passive infrared camera is used to move to the same plane to image the same static target scene, and the original red is obtained. The external image sequence is corrected by translation parallax correction to generate the infrared field database, and then the resampling image is obtained based on the light field rendering theory. Compared with the traditional infrared imaging that only records two-dimensional location information, infrared field imaging can record four dimensional information containing the position and direction. The experimental results show that the method can fuse multiple low signal noise. A resampling image with improved signal to noise ratio is obtained by comparing the image to the image. The image can be achieved through the average aperture superposition, and the digital refocusing of different depth of the infrared scene can be realized by the selection of different depth values. The infrared imaging is more valuable in military and civil fields.
【作者单位】：郑州大学物理工程学院;
【基金】：国家自然科学基金(61505178,11504333) 河南省高等学校重点科研项目(15A140014,17A140002) 郑州大学优秀青年教师发展基金(1521317007)
【分类号】：TN219

【相似文献】