基于鱼眼相机的自运动参数异步估计

Ego-Motion Asynchronous Estimation Based on Fish-Eye Camera

自运动参数估计是辅助驾驶、机器人导航等领域的核心问题之一.本文提出了一种适用于鱼眼相机的自运动参数异步估计方法.该方法通过特征分类及虚拟面投影,对旋转运动参数和平移运动参数分步估计.解决了以往算法中,旋转参数和平移参数同时估计时,二者相互影响的问题,提高了估计精度.本文首先利用平台运动特性简化相机运动模型,并根据不同距离不同位置的景物对运动参数估计的作用,对背景特征进行分类.分析并推导了各类特征的运动规律.然后根据运动规律,利用远处背景及一般背景特征估计旋转运动参数,利用地面特征估计平移运动参数.实验结果表明,本文方法不易受光照和干扰点影响,同一些经典方法相比,本文方法更具准确性和鲁棒性.

Ego-motion parameter estimation is one of the key problems in driver assistance and robot navigation, etc. An asynchronous estimation method which is applicable to fish-eye camera is proposed. The method estimates rotation parameters and tramlation parameters respectively based on feature classification and virtual plane projection. It solves the problem in the previous algorithms that rotation parameters and translation parameters influence each other when estimating at the same time, and improves the estimation accuracy. Firstly, camera motion model is simplified with platform motion characteristics, and background features from different distances and positions are classified according to their roles in motion estimation. The motion laws of features in each class are analyzed and induced. And then according to the laws, the distant background features and general background features are used to estimate the rotation parameters, and ground features are used to estimate the lxanslation parameters. The experimental results show that the proposed method is less influenced by illumination or outiiers, and more accurate and robust than some Iraditional methods.