摘要
提出了一种用于CMOS图像传感器的新型自适应降噪方法,并进行了逻辑电路实现。该方法通过运动检测和边缘检测技术进行噪声功率统计和分析,选择性的利用中值滤波、均值滤波以及运动自适应滤波方法,对传感器输出图像进行了数字降噪处理。实验表明本文方法降低了高斯噪声和椒盐噪声的影响,有效提高了图像质量和峰值信噪比。其结构易于电路实现,适用于CMOS图像传感器内部的实时降噪处理。
A novel adaptive noise reduction method for CMOS image sensor is proposed and implemented by digital logic circuit. Motion detection and edge detection is adopted for noise power statistic and noise analysis. To decrease the output digital image noise, the median filtering, average filtering and motion adaptive filtering, are selectively presented. Experiment results show that Gaussian noise and salt-pepper noise are reduced. The image quality and peak signal-noise ratio is impressively improved. This method is friendly for hardware implementation and real-time de-noising in CMOS image sensor.
出处
《传感技术学报》
CAS
CSCD
北大核心
2008年第9期1561-1565,共5页
Chinese Journal of Sensors and Actuators
基金
国家自然科学基金资助项目(60576025)
关键词
CMOS图像传感器
自适应降噪
噪声功率统计
峰值信噪比
CMOS Image Sensor(CIS)
adaptive noise reduction
noise power statistic
Peak Signal-Noise Ratio (PSNR)