论文部分内容阅读
针对飞秒激光烧蚀硅晶片微结构产生的等离子体光斑图像中的噪声及光斑亮度信号噪声,分别采用多尺度小波阈值方法进行去噪。首先,用CCD相机连续记录烧蚀过程中的等离子体光斑,并转化为序列图像;其次,采用多尺度小波分解对等离子体光斑目标图像进行小波系数提取,去除图像中包含的噪声干扰,再经小波逆变换实现图像的重构;再次,对滤波后的光斑图像进行亮度特征提取,获得光斑序列图像的连续亮度变化曲线;最后,对光斑序列亮度曲线再次进行小波分解、系数提取及信号重构,使不同烧蚀区间的光斑亮度平均信噪比提高了19.74%,获得了平滑的光斑亮度特征变化曲线。
Aiming at the noise in the plasma spot image and the spot brightness signal noise generated by the micro-structure of the femtosecond laser ablated silicon wafer, the multi-scale wavelet threshold method is used respectively to denoise. First, the CCD camera was used to continuously record the plasma spot in the ablation process and transform it into a sequence image. Secondly, wavelet transform was used to extract the wavelet coefficients from the target image of the plasma spot by using the multi-scale wavelet decomposition to remove the noise interference contained in the image. Wavelet inverse transform is used to reconstruct the image. Third, the brightness feature of the filtered spot image is extracted to obtain the continuous brightness curve of the spot sequence image. Finally, the wavelet decomposition, coefficient extraction and signal reconstruction , So that the average signal-to-noise ratio of the spot brightness in different ablation zones increased by 19.74%, and a smooth characteristic curve of the spot brightness was obtained.