基于改进SGM算法的双目立体匹配

Binocular Stereo Matching Based on Improved SGM Algorithm

SGM算法在立体匹配中对遮挡区域及视差不连续区域匹配能力较弱,容易产生较多空洞。针对此问题,提出一种SGM算法结合泛洪填充和中值平滑的视差优化算法,该算法是根据空洞边缘的视差值对缺失的视差值进行填充,即可实现对视差图中缺失视差的修复,并把所提算法在Middlebury立体匹配数据集上利用均方根误差(RMSE)、峰值信噪比(PSNR)、结构相似性(MSSIM)三个评价指标进行定量分析。实验结果表明对SGM算法的改进和优化能够有效地填充视差图中的空洞,减少图像的噪声,且视差图的匹配精度比SGM平均提高9.00%。

The SGM algorithm is weak in stereo matched to the occluded area and parallax discontinuity area.Tend to produce more voids,In response to this question.An SGM algorithm combines flood fills and median smooths for parallax optimization is proposed.The algorithm is to fill the missing parallax values based on the parallax values at the edge of the cavity.without the need to raise a larger computational volume.the repair of missing parallax in the parallax map can be achieved.The proposed algorithm is also quantitatively analyzed on the Middlebury stereo matching dataset using three evaluation metrics:root mean square error(RMSE),peak signal-to-noise ratio(PSNR),and structural similarity(MSSIM).The experimental results show that,The improvement and optimization of SGM algorithm can effectively fill the voids in the parallax map.reduce the noise of the image,And parallax map matching accuracy is on average 9%better than SGM.