本项目研究单曝光HDR（High Dynamic Range）成像关键技术。单曝光HDR成像技术可以解决多曝光HDR成像技术计算复杂、耗时长以及无法对动态场景成像等问题，也可以解决硬件直接HDR成像成本高、体积大等问题，是具有极大发展潜力的HDR成像技术。充分发挥这种技术的潜力，为实现能够满足图像处理相关邻域应用需求的便捷、高效的HDR成像奠定理论基础是本项目的研究目标。为此，本项目将对以下四个方面进行研究：1）由单帧图像生成多帧图像的算法；2）多帧图像融合的权值函数；3）像素动态范围扩展算法；4）图像中光源高亮度区域的HDR重建算法。项目中将采用成像机理分析、数学建模、方案探索以及计算机仿真的方式对单曝光HDR成像技术展开研究。

The key technologies of single-exposure High Dynamic Range (HDR) imaging are researched in this project. The single exposure HDR imaging can solve the problems occurred in multiple exposure HDR imaging technologies, such as high computation complexity, high time consuming and useless for dynamical scenes; meanwhile, it can avoid high cost and large size in HDR imaging by hardwares directly. The single exposure HDR imaging is the most promising HDR imaging technology. To bring into full play this potential and build the basis for achieving convenient and efficent HDR imaging which meets the market consume demands will be the research goals of this project. Thus, the project will research on the following contents: 1) multiple LDR images generation from single LDR image; 2) the fusion of multiple images; 3) pixel dynamic range expansion algorithm; 4) HDR reconstruction of the pixels in highlight areas. During the project, the methods including mechanism analysis, mathematical modeling, scheme exploration, and computer simulation will be used for the research on single exposure HDR imaging.