全光采集增强技术与模型研究

Due to the compact optical design, the hand-held plenoptic camera presents superior portability and low application system complexity. However, it also introduces a series of limitations in low imaging quality, narrow field of view and shallow depth of field, which strongly constrain its applications. Consequently, targeting future applications of hand-held plenoptic cameras in dynamic detection, this project researches on enhancement and modeling for plenoptic acquisition. The plenoptic system response together with fast light field correction method will be proposed by defining pupil modulation function to adapt to heterogeneous modulated light paths of plenoptic cameras. The acquired light field will be corrected to improve the imaging quality. Then, a series of multi-light-field fusing techniques will be proposed adapting to heterogeneous multi-light-field geometry to improve the field of view and to extend the depth of field efficiently. Based on the proposed plenoptic system response and enhanced light field, multi-object accurate inverse projection model together with optimized ray tracing model is built, that compensates the optical aberrations in the imaging system, to approach accurate inverse projection from the imaging space to the object space. The research tries to deliver a series of original innovations in fundamental theories, algorithms and techniques, and optimization mechanisms. They will serve for the leaping development in the areas related to real-time detection, biomedical imaging and so on by exploiting light field information, which presents obvious theoretical significance and broad application prospect.

备受关注的手持式全光相机，因其紧凑的光路设计，在带来便捷的可移动性和低应用复杂度的同时也导致其成像质量、光场视场角和景深非常受限，直接影响其推广应用。本项目面向手持式全光相机的动态观测这一光场未来应用，研究全光采集增强技术与模型：首先，建立全光系统响应和快速光场校正方法，定义光瞳调制函数，适应异构全光成像调制光路，校正采集光场，提升成像质量；进而，提出一系列系统层多光场融合技术，适应异构多光场几何关系，有效提升光场采集视场角和景深；最后，基于全光系统响应和增强的光场，建立多对象精准逆投影模型，优化光线追踪，补偿成像系统光学偏差，实现从像空间到物空间的精准投影。该研究成果将以基础理论、算法技术到优化控制方面的原始创新，服务于实时检测、医疗成像等领域基于光场信息的跨越式发展，具有重要的理论意义和广泛的应用价值。

备受关注的手持式全光相机，因其紧凑的光路设计，在带来便捷的可移动性和低应用复杂度的同时也导致其成像质量、光场视场角和景深非常受限，直接影响其推广应用。本项目面向手持式全光相机的动态观测这一光场未来应用，研究全光采集增强技术与模型：首先，建立全光系统响应和快速光场校正方法，定义光瞳调制函数，适应异构全光成像调制光路，校正采集光场，提升成像质量；进而，提出一系列系统层多光场融合技术，适应异构多光场几何关系，有效提升光场采集视场角和景深；最后，基于全光系统响应和增强的光场，建立多对象精准逆投影模型，优化光线追踪，补偿成像系统光学偏差，实现从像空间到物空间的精准投影。该研究成果将以基础理论、算法技术到优化控制方面的原始创新，服务于实时检测、医疗成像等领域基于光场信息的跨越式发展，具有重要的理论意义和广泛的应用价值。项目发表SCI检索高水平国际期刊论文7篇，国际重要学术会议论文13篇；授权中国发明专利10项，公开相关中国发明专利8项，授权美国发明专利1项；培养高质量的博士和硕士研究生11名，并荣获多项学术荣誉，圆满的完成了项目的预期目标。