摄影测量中一种基于HEIV模型的可纠错编码点法

在摄影测量中，经常发生椭圆点检测错误从而影响三维场景的匹配问题，甚至导致三维重建任务的失败。针对该问题，在基于HEIV(Heteroscedastic Errors-in-Variables)椭圆拟合方法的基础上，设计了一种可纠错的编码点，将信息论中的信道可纠错编码方法用于编码点的设计，文中对编码点的解算及提取进行了分析。实验表明，该方法大大提高了编码点的抗噪性，在编码出现一位错误时可以实现自修正的功能，纠错率为100%，较好的解决了实际应用中编码点识别错误的问题。

A Method About Error Correcting Code Based on HEIV Model in the Photogrammetry

In the measurement of photography, the error often happens when detecting the elliptic point which affects image matching ,and even causes the failure of 3D reconstruction.To solve this problem, the coding point which can correct the error is being designed based on Heteroscedastic Errors-in-Variables model and the identification and extraction of the coding points are analyzed in this paper.Experimental result show that this method can improve the robust noise immunity of the coding point,it also can correct the error automatically when has an error which resolve the error of distinguish coding point in practical application.