多智体协同作业模式可应用到矿井自动化开采/救援等工作中，有效降低矿井地质灾害的影响，达到减灾/救灾的目的。自主相对导航作为矿井多智体（即多个矿井智能机器人）协同工作的关键技术亟待解决。.本项目以数学形态学与对偶四元数为理论基础，研究矿井多智体视觉自主相对导航算法。主要研究：（1）结合矿井多智体/被操作目标（如矿工、矿车等）的外形特征及其工作环境，进行基于数学形态学的矿井多智体/被操作目标的快速特征提取、匹配算法研究；（2）以特征线为观测基元，构建矢量视觉方程，研究基于对偶四元数理论的矿井多智体之间的相对导航算法,确定矿井多智体间相对位置和姿态；（3）融合形态学特征提取与矿井多智体间的相对位姿信息，重点研究基于对偶四元数的矿井多智体与被操作目标的立体视觉相对导航融合算法，确定矿井多智体与被操作目标的相对位置和姿态。.该研究将为我国的矿井智能救援、自动化开采/装卸等工作提供理论支持和技术储备。

Muti-agents cooperation can minimize the effect of mine geologic hazard, and it is a valid way for disaster reduction and relief. Autonomous relative navigation, as the key technology of mine muti-agents cooperation, need to be resolved firstly. .On the theories of mathematical morphology and dual quaternion, the vision autonomous relative navigation algorithm for mine muti-agents is proposed. The main contents of this project including: (1) according to the attributes of mine muti-agents and mine environment, extraction algorithm based on mathematical morphology for mine muti-agents or object (such as miner, mine car etc) will be studied; (2) looking at feature line as observation value, vision vector equation will be built, and the relative navigation algorithm based on dual quaternion for mine muti-agents will be studied ulteriorly; (3) combining the information of feature extraction based-on mathematical morphology with the relative position & attitude of mine muti-agents, stereo vision relative navigation algorithm based on dual quaternion for mine muti-agents and object will be studied chiefly. And the relative position & attitude between mine muti-agents and the operated object are determined..The researches of this project will endeavour to provide the theoretical support and technical reserve for the applications of the intellectualized cooperation rescuing or mining in our country.