基于功率谱AR模型的3D激光虚拟图像重建研究

当前常用的无样本图像重建方法具有分辨率低的弊端,有样本图像重建方法得到的3D激光虚拟图像重建结果有很大误差。为此,提出一种基于功率谱AR模型的3D激光虚拟图像重建方法。建立功率谱AR模型,将Burg技术和自相关法结合在一起,得到合理的AR模型参数。在存在平移、旋转参数的情况下,在功率谱AR模型基础上对图像进行配准处理,先完成旋转步长,然后进行平移估计与步长,以提高旋转配准精度与平移配准精度。将3D激光虚拟图像信息看作希尔伯特空间的元素,依据配准结果,在重建的3D激光虚拟图像重建视觉效果达到要求或连续若干次迭代结果改变情况区域稳定的情况下,结束迭代,实现3D激光虚拟图像重建。经验证,所提方法重建精度最低为97%,重建误差最低为2%,而且重建花费时间较短,证明了本文方法在图像重建方面的优越性,可应用于图像重建的相关工作中。

3D laser virtual image reconstruction based on the AR model of the power spectrum

At present,the standard method of image reconstruction without sample has the disadvantage of low resolution.The result of 3D laser virtual image reconstruction with sample image reconstruction method has a great error.Therefore,this paper proposes a 3D laser virtual image reconstruction method based on AR model of the power spec-trum.The Burg technology and autocorrelation method are combined to get the reasonable AR model parameters by establishing the power spectrum's AR model.In.the case of translation and rotation parameters,based on the power spectrum's AR model,the image registration is processed.The rotation step is completed first,then the translation estimation and step size are carried out to improve the accuracy of rotation registration and translation registration.The 3D laser virtual image information is regarded as the element of Hilbert space.According to the registration results,when the reconstructed 3D laser virtual image reconstruction visual effect meets the requirements or the region is stable after several successive iterations,the iteration ends to realize 3D laser virtual image reconstruction.After verification,the proposed method‘s reconstruction accuracy is 97%,and the reconstruction error is 2%.Moreover,the reconstruction time is relatively short,which proves the superiority of the proposed method in image reconstruction,and can be applied to image reconstruction related work.