基于空间平面约束的视觉定位模型研究  

Vision-based localization model based on spatial plane constraints

作　　者：高飞[1] 葛一粟 汪韬[1] 卢书芳[1] 张元鸣[1] Gao Fei;Ge Yisu;Wang Tao;Lu Shufang;Zhang Yuanming(Zhejiang university of technology,Hangzhou 310023,China)

机构地区：[1]浙江工业大学

出　　处：《仪器仪表学报》2018年第7期183-190,共8页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(61402410);浙江省重点研发计划(2018C01064)项目资助

摘　　要：针对现有视觉定位方法中存在的立体匹配效率低下等不足,为实现平面运动目标位姿的实时视觉计算,提出基于空间平面约束的视觉定位模型。首先,通过双目视觉标定得到相机内外参数;然后,使用平面标定方法计算出空间约束平面方程;其次,根据空间平面方程给出了目标平行平面上任意点的空间坐标计算方法,进而实现运动目标位姿测算;最后,以1∶49的集装箱卡车视觉引导进行了模拟实验。实验结果表明,该模型计算得到的物理位置的标准差在略优于双目模型的情况下,具有更强的实时性。The current vision-based localization methods have the limitations of stereo matching efficiency,etc. To realize the visual accurate measurement of real-time planar moving pose,a vision-based localization model with spatial plane constraints is proposed.First,the internal and external parameters of the camera are obtained by stereo vision calibration. Secondly,the plane equation is also calculated by using a plane calibration method. Thirdly,according to the proposed vision-based localization model,the equation of target parallel plane and the point coordinates are achieved,which is further utilized to estimate the orientation of the moving object. Finally, the simulation experiment of vision-based localizing with container truck with the ratio of 1 ： 49 is implemented. Experimental results show that the proposed model is real-time and with equal or a little more measurement accuracy to/than the stereo model.

关 键 词：双目视觉 视觉定位 平面约束 三维信息获取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TH744[自动化与计算机技术—计算机科学与技术]