一种改进的非锐化掩模深度图像增强算法

An improved unsharp masking method for depth map enhancement

摘要:

针对深度图像平滑过程中会模糊细节的缺点,提出了自适应的非锐化掩模深度图像增强算法.首先将深度图对应的彩色图像作为联合双边滤波的引导图,利用彩色图像相关特征修复了深度图像的缺失和毛糙,然后将双边滤波后的深度图像与高斯滤波后的深度图像作差,提取出不含噪声的高频部分,克服了经典非锐化掩模算法放大高频噪声的缺点,最后根据边缘以及细节的模糊程度,自适应地调节叠加到图像上的高频部分.实验结果表明,设计的算法有效地增强了深度图像细节,抑制了平坦区域的噪声,并填补了边缘缺失,较好地改善了深度图像的质量.

Abstract:

An adaptive unsharp masking filter is proposed to enhance the details blurred in the depth map filtered. Firstly, the color image is used as guided image in joint bilateral filter to recover the lost region and reduce the noise. Then a mask is extracted by subtracting a low-pass filtered depth map from the bilateral filtered depth map, which effectively overcomes the shortcoming of traditional method that amplifies high-frequency noise. Finally, according to degree of blur, the spatial importance is adaptively added to the depth map filtered. Experimental results show that the proposed method performs better in enhancing the details, reducing noise and recovering the lost region of depth map.