抗噪高清图像清晰度评价算法的研究

Research on Anti-noise and HD Image Clarity Evaluation Algorithms

监控摄像机的图像在夜晚黑白模式下噪点较多,为解决在高噪声下清晰度评价函数效果差的问题,首先提出一种使用斜条形窗口的条形区域选择方法大幅度缩减高清图像的运算量,之后提出一种使用清晰度评价因子来修正Laplace算子检测结果的清晰度评价算法。该算法首先对图像进行均值加权滤波,在缩小图像尺寸的同时进行降噪处理,对滤波后的图像进行Laplace算子清晰度计算,获取其清晰度值,之后使用像素区间比例因子和归一化邻近差值两种计算方法对滤波后的图像进行处理得到清晰度评价因子,使用该因子对清晰度值进行修正。仿真实验表明,其在高噪声下仍能具有快速良好的评价效果。

The image of surveillance camera has high noise in black-and-white mode at night.To solve the problem of poor evaluation effect under high noise,firstly,a strip region selection method based on oblique strip window is proposed to reduce the computational complexity of HD image greatly.After that,a clarity evaluation algorithm using clarity evaluation factor to modify the results of Laplace operator detection is proposed.Firstly,the image is filtered by means of weighted filter,which reduces the size of the image and noise at the same time.After that,the filtered image is processed by using both the pixel interval scale factor and the normalized adjacent difference method to get the clarity evaluation factor.The simulation results show that it can still have a good and fast evaluation effect under high noise.