无人机采场图像二维经验小波变换降噪研究

Research on 2D empirical wavelet transform noise reduction of UAV stope images

针对传统图像降噪方法难以适用于露天矿无人机(UAV)图像降噪的问题,该文提出了一种基于经验小波变换(EWT)算法思想的露天矿UAV图像快速降噪方法。采用Littlewood-Paley小波算子检测提取UAV图像噪声特征,以提高阈值函数匹配精度;借助改进的自适应H阈值函数对UAV图像噪声进行计算,提高UAV图像边缘细节度;并通过二维经验小波变换逆重构,从而提高UAV图像纹理精度。实验结果表明,本文算法能够实现露天矿UAV图像的快速降噪,保留边缘细节和纹理特征,并具有较好的降噪效果。

Aiming at the problem that traditional image denoising method was not suitable for UAV image denoising in open-pit mine,this paper proposed a fast noise reduction method for UAV image based on EWT algorithm.The noise characteristics of UAV images were extracted by Littlewood-Paley wavelet operator to improve the matching accuracy of threshold function.The improved adaptive H threshold function was used to calculate the noise of UAV image,which improved the edge detail of UAV image.The texture accuracy of UAV image was improved by inverse reconstruction of two-dimensional empirical wavelet transform.The experimental results based on an open pit mine in Henan showed that the proposed algorithm could achieve fast denoising of UAV images,preserved edge details and texture features,and had better noise reduction effects.