摘要
针对现有宽动态合成算法因复杂度高而难以应用于实时系统的缺点,提出一种有效且复杂度低、适合硬件实现的实时宽动态图像合成方法。算法首先获取同一场景两副不同曝光量的照片,通过使用线性的亮度映射函数T,从短曝光图像对应区域的信息恢复出长曝光图像中过曝区域的信息。该算法只需利用两幅原始图像,所需数据量少,且无须恢复图像中的亮度与曝光比信息,计算复杂度很低。同时,T用线性关系表示更易于硬件实现。实验表明,算法能够实现良好的宽动态图像效果。
For the complexity of the existing wide dynamic synthesis algorithm, which is difficult to apply to the real-time systems such as video surveillance,a synthesis method is proposed which is very efficient with low complexity,while suitable for the wide dynamic image of hardware implementation. Firstly the proposed algorithm gets two images of the same scene at different exposures. The missing information of long-exposure image can be restored from the information of corresponding area in the short-exposure image with the intensity value mapping function T. The proposed algorithm uses only two original images, which means the amount of data is less,without the need to restore the radiometric response function and the exposure rate,and the computational complexity is low. Meanwhile, the linear representation of T is more conductive to the hardware implementation. The experimental results show that the effect of the synthetic wide dynamic image is satisfactory.
出处
《电视技术》
北大核心
2013年第11期47-50,共4页
Video Engineering
基金
国家自然科学基金项目(60902052)
西北工业大学研究生创业种子基金项目(z2012093)
教育部博士点基金项目(20096102120032)
关键词
WDR
多次曝光
最小二乘法
动态范围扩展
累计直方图
WDR
multiple exposure
least square method
dynamic range extension
accumulated histogram