强噪声条件下激光雷达强度图像降噪方法研究

Research on image denoising method of lidar intensity under strong noise conditions

针对传统的激光雷达强度图像降噪方法中,普遍存在着降噪效果不理想、峰值信噪比较低等问题,为了提高激光雷达图像峰值信噪比,从而实现激光雷达图像降噪,提出基于小波域混合模型的激光雷达强度图像降噪方法。分析激光雷达图像中噪声的特点,将其进行对数变换,利用小波变换对图像进行边缘检测,以得到激光雷达强度图像噪声所在位置为依据;利用自适应阈值对激光雷达强度图像变换后的小波系数进行分类,将双变量模型和空间自适应统计模型相结合实现对激光雷达强度图像降噪。实验结果表明,所提方法图像降噪效果较为理想、峰值信噪比较高,所提方法的峰值信噪比比另外两种方法高出10.93和16.13。

In order to improve the peak signal-to-noise ratio(PSNR)of lidar images,there are some problems in the traditional methods of intensity image denoising,such as the unsatisfactory effect and the low PSNR,so as to realize the de-noising of lidar images.A method of intensity image denoising for lidar based on mixed wavelet model is pro-posed.The characteristic of noise in laser radar image is analyzed,and the logarithmic transform is carried out,and the edge of image is detected by wavelet transform,which is based on the location of the noise in the image of laser radar intensity.The wavelet coefficients of lidar intensity image are classified by adaptive threshold,and the bivariate model and spatial adaptation are used to classify the wavelet coefficients.Combined with statistical model,the noise reduction of lidar intensity image is realized.The experimental results show that the proposed method is effective in image denois-ing,and the peak signal-to-noise ratio(PSNR)of the proposed method is higher than that of the other two methods(10.93 and 16.13).