融合超像素分割技术的虚拟成像细节重建研究

Research on Virtual Image Detail Reconstruction Based on Super Pixel Segmentation

虚拟成像是在三维空间中投射立体影像的技术,受采集设备、环境以及网络传输等因素的影响,图像易出现分辨率降低问题,导致虚拟成像细节重建效果不理想。为解决上述问题,提出基于VR和小波降噪的复杂虚拟成像重建方法。依据光学原理确定最小识别距离,结合摄像机采集目标物体多视点图像。利用简单线性迭代聚类(simple linear iterative clustering, SLIC)超像素分割技术优化小波降噪法,完成图像的去噪。基于此,通过VR技术获取全息三维影像重建点坐标和深度距离,并将其投射至透明介质上,实现复杂虚拟成像重建。实验结果表明,研究方法的虚拟成像超像素分割效果更优,平均峰值信噪比为25.5dB,平均结构相似度为0.83,具有较高的应用可靠性。

Virtual imaging is a technology of projecting 3D images in three-dimensional space.However,it is affected by some factors such as acquisition equipment,environment and network transmission,and the resolution of images is easy to be reduced.As a result,this article presented a complex virtual imaging reconstruction method based on VR and wavelet denoising.According to the optical principle,we determined the minimum recognition distance and collected the multi-view images of the target object by the camera.And then,we used the super-pixel segmentation technology based on simple linear iterative clustering(SLIC)to optimize the wavelet denoising method,thus completing the image denoising.On this basis,we used VR technology to calculate the coordinates and depth of the reconstruction points of the holographic image and projected them onto the transparent medium,thus realizing complex virtual imaging reconstruction.Experimental results show that the proposed method has better super-pixel segmentation effect and high application reliability.The average peak signal-to-noise ratio is 25.5dB,and the average structural similarity is 0.83.