摘要
针对常见白平衡算法存在色温修正性不足或算法复杂度过高的问题,提出一种基于改进的灰度世界算法、改进的完美反射算法和单色算法相结合的三模互补自动白平衡算法。为完美反射算法增加色温估计约束条件以增强色温估计的精度,对灰度世界算法增加有效性判断。结合完美反射算法和灰度世界算法各自的特点,以弥补单个算法的不足。单色算法利用色温曲线排除单色场景,避免大片单色区域造成的白平衡失效。在一个CMOS图像传感器的信号处理过程中实现该算法。在不同色温环境以及多种场景下进行测试,结果表明该算法对图像偏色有较好的校正效果和适应性。
The common automatic white balance algorithm is lack of adaptability or too complicated. An automatic white balance algorithm based on improved gray world algorithm,improved perfect reflection algorithm and mono color algorithm is proposed. In order to enhance the precision of perfect reflection to increase color temperature estimation constraint conditions, the effectiveness of judgment is added on the gray world algorithm. The combination of perfect reflection and gray world algorithm makes up for the deficiency of the two algorithms effectively. Mono color algorithm using color temperature avoids white balance malfunction in the case of one color image. The algorithm is realized in the design of a CMOS image sensor Image Signal Processing （ISP）. The test result under different light source and scene shows that the presented method can available adjust image color bias and has strong adaptability.
出处
《计算机工程》
CAS
CSCD
北大核心
2015年第9期245-250,共6页
Computer Engineering
基金
上海市经信委软件和集成电路产业发展专项基金资助项目(130303)
关键词
CMOS图像传感器
图像信号处理算法
自动白平衡
完美反射
灰度世界
色温曲线
单色
CMOS image sensor
Image Signal Processing （ISP） algorithm
automatic white balance
perfect reflection
gray world
color temperature curve
mono color
