Research of Three Dimentional Survey System from TV Image.

街道匹配技术包含了从电视或CCD图像中提取三维信息的各种算法。本文利用彩色电视摄像机在室外拍摄的图像，对几种算法进行了比较和评价。在进行立体匹配之前，首先将图像的色彩空间基由RGB变换为HSV或HSL，然后对变换后的图像进行主成分分析。零交叉法检测第一主成分图像的边缘部分。对由边缘交叉线组成的不同层次进行边缘匹配。利用边缘匹配结果，分别采用互相关法、SSDA法和傅立叶互相关法对匹配精度进行了检验。

Streo matching technique contains various algorithms to extract the three dimensional information from TV or CCD image. In this study several algorithem have been compared and evaluated using images tgaken by color TV camera at the outside. Before executing the stereo matching proceduce, the basis for color space of image is transformed from RGB into HSV or HSL, and then the principal component analysis method is applied to the transformed images. Zero crossing method detects the edge part in the first principal compornent images. Edge matching procedure is preformed for the distinctive lavels which consist of edge crossing lines. The accuracy of matching is examined about the cases adopting each of Cross correlation method, SSDA method and Fourier cross correlation method using edge matching result.