一种新型提高光斑图像质心精度的去噪方法*

提出了一种新的结合自适应中值滤波和阈值的去噪方法。与其他中值滤波方法不同，这里采用的自适应中值滤波的滑动窗口大小根据窗口内灰度的均值与最佳阈值下目标部分灰度均值和背景部分灰度均值的差别自适应确定。实验结果表明，与仅用中值滤波或仅用简单阈值对图像去噪的方法相比，利用该去噪方法，图像的信噪比提高了10.42 dB，光斑图像质心的精度最大提高了80%左右。

New denosing method for improving speckle image centroid accuracy

This paper proposed a new noise reduction method combined both adaptive median filtering algorithm and threshold method. The difference between this median filtering method and other method was that, the size of the filter window was selected adaptively by the difference between the average gray in the slide window and the average gray of the background and object under the optimized threshold of the image. The experimental results show that, compared to the method only based on median filtering or simple threshold method, the signal noise ratio (SNR) of the image is improved by 10.42 dB, and the accuracy of the centroid is improved 80% by using the method proposed.