面向低照度环境的三维弹性离轴分布感知技术研究

Integral imaging has drawn a great attention to researchers because it provides the observing images with full parallax and continuous viewing points. Compared with holography, integral imaging is able to work with incoherent light, relatively small amount of data, and simple imaging architectures. However, the precise position for the sensor in integral imaging has been a challenge for scenarios of practical interest. Therefore, a new flexible imaging system with off-axially distributed sensing will be studied in this project, where horizontal and longitudinal perspective information is obtained by flexibly translating and imaging along the off-optical axis. The proposed system will address the issue in terms of complicated architectures for current integral imaging systems, and meanwhile it remains shifting and scaling features as integral image does. Furthermore, for applications under low illumination, the probability reconstruction algorithm will be considered based on prior distribution and a Bayesian approach. The studies may significantly extend practical applications of integral imaging in deep ocean detection, satellite remote sensing, biomedical science, etc.

集成成像作为一种无视觉疲劳的全视差、连续视点的裸眼显示方式，现已成为真三维立体成像及显示领域的研究热点。与全息三维技术相比，集成成像技术以简单的成像结构、较小的数据计算量及无需相干光源和苛刻的成像显示环境等方面具有明显的优势。然而，现有的集成成像存在成像结构复杂的问题，而且精密的传感器个体空间分布给成像带来极大挑战。为此，本项目将深入研究弹性离轴分布感知系统，其自由弹性结构既解决了精密复杂的二维成像结构问题，既保留了横向及纵向分辨率的同时又融合了偏移和缩放特征。此外，为解决在低照度环境下派生出的三维场景可视化重构问题，进一步提出为抑制噪声引入图像先验分布的光子计数可视化重构算法。本项目的研究能够有效地深化发展集成成像在深海探测、卫星遥感、生物医学等领域的实际应用。

集成成像技术作为一种三维成像和可视化领域的新技术，在三维重建、目标探测及生物医学微成像等领域有着广泛的应用前景。本课题实现了一种具有偏移和放缩特征的弹性离轴分布感知系统，解决了目前基于相机阵列的集成成像系统中结构复杂、成像效率低等技术难题。在低照度环境下的弹性离轴分布系统设计和三维重构算法，显著性目标检测和遮挡去除等实际应用方面取得了一系列研究成果。在本项目的资助下，课题组共发表学术论文21篇，其中SCI论文7篇，国际会议论文14篇；获批1项国家自然科学基金；获得2项辽宁省自然科学学术成果奖，1项全国智能移动终端应用创新大赛一等奖；授权3项国内发明专利；培养了硕士生5名。本项目的研究不仅能够有效的推动集成成像在实际应用领域上的发展，还将为我国发展下一代三维可视化、目标探测及生物医学领域的三维立体微成像等领域所涉及的技术难题提供技术支撑，具有重要的理论价值和实用前景。